(Social Cognition and Social Neuroscience) Patrick Haggard, Baruch Eitam-The Sense of Agency-Oxford University Press (2015)

The Sense of Agency
Patrick Haggard and Baruch Eitam
ABSTRACT
This book focuses on the scientific study of the human sense of agency. It discusses the causes and
consequences of the subjective experience of being in control of ones actions, and, through them, of
events in the outside world. The book brings together some of the worlds leading researchers on the topic.
It aims to provide the first structured survey of this nascent but rapidly growing field.
Keywords: sense of agency, psychological agency, bodily self, sense of self, consciousness, action-perception interactions
BIBLIOGRAPHIC INFORMATION
Print publication date: 2015 Print ISBN-13: 9780190267278
Published to Oxford Scholarship Online:

DOI:10.1093/acprof:oso/9780190267278.001.0001
September 2015
AUTHORS
Affiliations are at time of print publication.
Patrick Haggard, editor

Professor in the Institute of Cognitive Neuroscience and the Department of Psychology, University College London
Baruch Eitam, editor
Professor of Psychology, University of Haifa
Front Matter
Title Pages
Contributors
Introduction
Part I Volition
1 Time to Act
Elisabeth Pacherie
2 Deconstructing Voluntary Action
Lara Krisst, Carlos Montemayor, and Ezequiel Morsella
3 Action Control by If-Then Planning
Torsten Martiny-Huenger, Sarah E. Martiny, and Peter M. Gollwitzer
4 Neural Correlates of Intention
Roee Gilron, Shiri Simon, and Roy Mukamel
5 Explicit and Implicit Beliefs, Attitudes, and Intentions
Icek Ajzen and Nilanjana Dasgupta
6 The Neural Basis Underlying the Experience of Control in the Human Brain
Lauren A. Leotti, Catherine Cho, and Mauricio R. Delgado
7 Goals and the Sense of Agency
Tali Kleiman
Part II Determining Authorship

8 Inference Processes Underlying the Human Experience of Agency over Operant Actions
Myrthel Dogge and Henk Aarts
9 Agency and Outcome Prediction
Antje Gentsch and Simone Schtz-Bosbach
10 The Relations Between Agency and Body Ownership
Manos Tsakiris
11 The Innate Experience of Self-Agency
Philippe Rochat
12 Motivation from Control
Noam Karsh and Baruch Eitam
Part III Beyond Authorship

13 Comparators and Weightings
Matthis Synofzik
14 Action Control and the Sense of Agency
Bernhard Hommel
15 Control and Truth Working Together
E. Tory Higgins
Part IV Disturbances
16 Sense of Agency and Its Disruption
Paul Fletcher and Aikaterini Fotopoulou
17 Action Generation, Intention, and Agency in Motor and Body Awareness Deficits
Anna Berti, Francesca Garbarini, and Lorenzo Pia
18 Disorders of Volition from Neurological Disease
James B. Rowe and Noham Wolpe
End Matter
Index
Title Pages
University Press Scholarship Online

Oxford Scholarship Online
The Sense of Agency

Print publication date: 2015

Print ISBN-13: 9780190267278
Published to Oxford Scholarship Online: September 2015
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Title Pages
The Sense of Agency Oxford Series in Social Cognition and Social Neuroscience
The Sense of Agency
Series Editor
Ran R. Hassin
Series Board
Mahzarin Banaji, John A. Bargh, John Gabrieli, David Hamilton,
Elizabeth A. Phelps, and Yaacov Trope
The New Unconscious
Edited by Ran R. Hassin, James S. Uleman, and John A. Bargh
Oxford Handbook of Human Action
Edited by Eziquiel Morsella, John A. Bargh, and Peter M. Gollwitzer
Social Neuroscience: Toward Understanding the Underpinnings of the Social Mind
Edited by Alexander Todorov, Susan T. Fiske, and Deborah Prentice
Self Control in Society, Mind, and Brain
Edited by Ran R. Hassin, Kevin N. Ochsner, and Yaacov Trope
Oxford Handbook of Social Neuroscience
Edited by Jean Decety and John T. Cacioppo
Page 1 of 3
Title Pages
Attention in a Social World
Michael I. Posner
Beyond Pleasure and Pain
E. Tory Higgins
The Sense of Agency
Edited by Patrick Haggard and Baruch Eitam
(p.iv)
Oxford University Press is a department of the University of

Oxford. It furthers the Universitys objective of excellence in research,
scholarship, and education by publishing worldwide.
OxfordNew York
AucklandCape TownDar es SalaamHong KongKarachi
Kuala LumpurMadridMelbourneMexico CityNairobi
New DelhiShanghaiTaipeiToronto
With offices in
ArgentinaAustriaBrazilChileCzech RepublicFranceGreece
GuatemalaHungaryItalyJapanPolandPortugalSingapore
South KoreaSwitzerlandThailandTurkeyUkraineVietnam
Oxford is a registered trademark of Oxford University Press

in the UK and certain other countries.
Published in the United States of America by

Oxford University Press
198 Madison Avenue, New York, NY 10016
Oxford University Press 2015
All rights reserved. No part of this publication may be reproduced, stored in

a retrieval system, or transmitted, in any form or by any means, without the
prior
permission in writing of Oxford University Press, or as expressly permitted by
law,
by license, or under terms agreed with the appropriate reproduction rights
organization.
Inquiries concerning reproduction outside the scope of the above should be
sent to the
Page 2 of 3
Title Pages
Rights Department, Oxford University Press, at the address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

The sense of agency / edited by Patrick Haggard and Baruch Eitam.
pages cm. (Oxford series in social cognition and social neuroscience)
Includes bibliographical references and index.
ISBN 9780190267278
1.Control (Psychology)I.Haggard, Patrick, editor.II.Eitam, Baruch,
editor.
BF611.S427 2015
153.8dc23
2015006192ss
987654321
Printed in the United States of America
on acid-free paper
Page 3 of 3
(p.vii)
Contributors

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
(p.vii) Contributors
Henk Aarts
University Medical Centrum Utrecht
Utrecht University
Utrecht, the Netherlands
Icek Ajzen
Department of Psychology
University of Massachusetts
Amherst, MA
Anna Berti
SAMBA (SpAtial, Motor & Bodily Awareness) Research Group
Psychology Department
University of Turin
Turin, Italy
Catherine Cho
Rutgers University, Newark
Newark, NJ
Nilanjana Dasgupta
University of Massachusetts
Amherst, MA
Mauricio R. Delgado
Rutgers University
Newark, NJ
Myrthel Dogge
University Medical Centrum Utrecht
Utrecht University
Page 1 of 4
(p.vii)
Contributors
Utrecht, the Netherlands

Baruch Eitam
University of Haifa
Haifa, Israel
Paul Fletcher
Department of Psychiatry
CEHP Research Department, Division of Psychology and Language Sciences
University of Cambridge
Cambridge, UK
Aikaterini Fotopoulou
Research Department of Clinical, Educational and Health Psychology
Psychoanalysis Unit
University College London
London, UK
Francesca Garbarini
University of Turin
Turin, Italy
Antje Gentsch
Max Planck Institute for Human Cognitive and Brain Sciences
Leipzig, Germany
Roee Gilron
School of Psychological Sciences and Sagol School of Neuroscience
Tel-Aviv University
Tel-Aviv, Israel (p.viii)
Peter M. Gollwitzer
New York University
New York, NY
E. Tory Higgins
Columbia University
New York, NY
Bernhard Hommel
Department of Cognitive Psychology
Leiden University
Leiden, Germany
Noam Karsh
University of Haifa
Haifa, Israel
Tali Kleiman
Trope Lab
New York University
New York, NY
Page 2 of 4
(p.vii)
Contributors
Lara Krisst
San Francisco State University
San Francisco, CA
Lauren A. Leotti
Rutgers University
Newark, NJ
Sarah E. Martiny
University of Troms
Troms, Norway
Torsten Martiny-Huenger
University of Konstanz
Konstanz, Germany
Carlos Montemayor
Department of Philosophy
San Francisco, CA
Ezequiel Morsella
Department of Neurology
University of California
San Francisco, CA
Roy Mukamel
Tel-Aviv University
Tel-Aviv, Israel
Elisabeth Pacherie
Institut Jean Nicod
Ecole Normale Suprieure
Paris, France
Lorenzo Pia
University of Turin
Turin, Italy
Philippe Rochat
Emory University
Atlanta, GA
James B. Rowe
Department of Clinical Neurosciences
Medical Research Council Cognition and Brain Sciences Unit
Cambridge
Page 3 of 4
(p.vii)
Contributors
Behavioural and Clinical Neuroscience Institute

Cambridge, UK (p.ix)
Simone Schtz-Bosbach
Friedrich Alexander University Erlangen Nuremberg
Max Planck Institute for Human Cognitive and Brain Sciences
Leipzig, Germany
Shiri Simon
Tel-Aviv University
Tel-Aviv, Israel
Matthis Synofzik
Department of Neurodegenerative Diseases, Hertie-Institute for Clinical Brain
Research
University of Tbingen
German Research Center for Neurodegenerative Diseases (DZNE)
University of Tbingen
Tbingen, Germany
Manos Tsakiris
Laboratory of Action and Body
Royal Holloway
University of London
Egham, UK
Noham Wolpe
Department of Clinical Neurosciences
Medical Research Council
Cognition and Brain Sciences Unit
Cambridge, UK
(p.x)
Page 4 of 4
(p.xi)
Introduction

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
(p.xi) Introduction
This book is about the human sense of agency. In psychology, the term agency refers to the
capacity to perform an action, or the event of performing that action. The term is normally used
to refer to specific classes of actionactions that are in line with a persons conscious goals and
intentions, and that the person herself initiates and controls. One can therefore speak about
objective facts of agency. Indeed, facts of agency are the main concern of the law, when trying
to establish criminal responsibility. In addition, agency is also a subjective judgment, state, or
experience. We prefer the term sense of agency to refer to this subjective element, to contrast
with the objective facts of agency. However, the term agency tout court is sometimes used to
refer to the subjective experience as well. The sense of agency thus refers to the experience or
judgment that one initiates and that controls an action, particularly an intentional, goal-directed
action, regardless of whether one objectively initiated, or is responsible for, that action. To
experience a sense of agency with respect to an action and an external event, then, an agent
must make an intentional action or must consciously think that she has made such an action.
Further, the content of her intention must be to produce that specific external event. Therefore,
the normal sense of agency includes the experience that one actually achieves the intended goal,
through ones own action.
The importance of both facts of agency and sense of agency in human life is undeniable. Much of
what you see around you as you read this is probably man-made: buildings, furniture, clothes,
and books. These are all products of human agency. Bergson (1998) used the concept of homo
faber to indicate how much our neurobiological capacity for goal-directed action has marked
our society and our world. More than any other animal, humans construct their physical and
social environment, and cause their own key life events, through their own actions.
The capacity for agency seems central to key steps in human history: agriculture, technology,
and civic society can all be seen as products of agency, and examples of our mental capacity to
change our world. Intentional agency is, moreover, embedded in our social institutions. Society
makes external judgments of agency for allocating both praise and rewards and blame and
(p.xii) punishments. For example, many judicial systems treat premeditated murder very
Page 1 of 6
(p.xi)
Introduction
differently from involuntary manslaughter such as accidental killing, or even from voluntary
manslaughter such as crime passionnel. The outcome is identical in each case, but society
cares deeply about the differences in the degree of perceived agency in these situations.
The underpinnings of humans unique capacity for agency are mainly cognitive. First, the human
mind can swiftly associate actions with outcomes that are remote in space and time. Second, the
human mind can readily use complex mediating, indirect, or branching chains between action
and outcome. For example, if my goal is to dig a hole, I may begin my efforts by the sub-goal of
making a spade. Or, if my goal is to maintain my standard of living when I retire, I may begin
saving years before retirement. Interestingly, some of these cognitive feats are clearly
accompanied by a sense of agency, as well as an objective fact of agency. For example,
technological and political innovations are often accompanied by conscious and deliberate
reflection about reasons, possibilities, and advantages of those innovations. In fact, the
importance of the sense of agency, of knowing or feeling that the self has produced an action
or outcome, can scarcely be underestimated. Agency can serve, for example, as a guide to ones
behavioural repertoire (Bandura, 1977), as a guide to expecting reward or punishment given
specific outcomes, and as a guide to the possibilities or affordances of a given environment. Yet,
because sense of agency is both a ubiquitous feature of human life and has multiple phenomenal
and cognitive manifestations, it can be hard to isolate and define. In particular, everyday agency
can feel phenomenally thin. We control riding our bicycle to work, preparing a meal, and talking
to a friend as intentional actions, but performing them does not produce a strong experience of
each discrete action event being intentionally controlled. Sense of agency more resembles a
phenomenal background than a clear and distinct perceptual experience. However, both the
uniqueness and informative aspect of the phenomenology of agency is clearly shown by recent
studies of disorders of agency. For example, psychotic patients may experience that they do not
cause their own actions, or they may overestimate their own agency, for example in believing
that their thoughts can directly influence external events. As a number of chapters in this
volume show ( Chapter 16 by Fletcher & Fotopoulou; Chapter 17 by Berti, Garbarini, & Pia;
Chapter 18 by Rowe & Wolpe), a range of psychiatric and neuropsychological phenomena, such
as the alien hand and anarchic hand syndromes, have been explained with reference to models
of normal agency.
Causes and Consequences of Sense of Agency

The vast majority of work on sense of agency has focused on its causes. Several chapters in this
book show the importance of the comparator model as a (p.xiii) computational account of
sense of agency (Chapter 1 by Pacherie; Chapter 9 by Gentsch & Schtz-Bosbach; Chapter 12 by
Karsh & Eitam; Chapter 13 by Synofzik; Chapter 16 by Fletcher & Fotopoulou; Chapter 17 by
Berti, Garbarini, & Pia). This model has its basis in the mid-twentieth-century rapprochement
between systems engineering and biological cybernetics (Sperry, 1950; Von Holst &
Mittelstaedt, 1950). For our purposes, the key step is linking sense of agency to unconscious
predictions about the consequences of action. Once a movement is planned, its predicted
consequences can be estimated. If and only if the final motor command is actually sent, and the
predicted sensory event occurs, then one feels a sense of agency with respect to that event. In
essence, comparator models generate an experience of I did that by checking whether that
fits a model of the estimated consequences of action (Synofzik, Vosgerau, & Newen, 2008;
Synofzik, Chapter 13 of this volume).
Page 2 of 6
(p.xi)
Introduction
There is widespread agreement on the importance of prediction as one cause of sense of agency.
However, the comparator model seems unable to account for some features of the subjective
experience of ones own agency. First, the model produces no net signal when there is a match
between an action and its expected consequences, but does produce a net signal when there a
mismatch. This feature of the model could explain why agency is a thin, default, background
experience. At the same time, the comparator model could be viewed as a model of non-agency
rather than a model of any positive experience of agencyand under some circumstances we
are in fact acutely aware of our own agency. Second, the model produces a sense of agency only
after a sensory event occurs, that is, retrospectively. However, one may also feel a sense of
anticipatory agency: someone who orders a meal in a restaurant feels in control of the situation
before the waiter arrives with his food. Accordingly, several of the chapters in this book describe
experiments investigating exactly what is predicted in advance of an action, and how and when
these predictions influence feelings of control.
In contrast, there is far less recent research and even less consensus regarding the
consequences of agency. Perhaps trivially, recognizing oneself as the author of action is a
consequence of the fact of agency. Explicit agency judgments of this kind are rare in everyday
life, but have been important in laboratory studies as criteria for uncovering the parameters of
the process of agency decision-making. Explicit judgments of agency are also strongly linked to
self-recognition and self-awareness, and this scientific tradition accordingly suggests a strong
link between agency and individual psychological identity (see Chapter 10 by Tsakiris; Chapter
11 by Rochat; Chapter 16 by Fletcher & Fotopoulou; Chapter 17 by Berti, Garbarini, & Pia). In
one paradigm, participants act and then make attribution judgments about the actions
consequences. For example, Daprati and colleagues (Daprati et al., (p.xiv) 1997) asked
participants to judge whether a gesture on a visual display did or did not correspond to the
gesture they were making (or to that of a confederate making similar actions in an adjacent
room). These experiments are interestingly close to the classic rouge test (Gallup, 1970) that has
been proposed as a marker of self-awareness in animals. This simple laboratory paradigm has
yielded three important results that seem fundamental to the sense of agency. First, people
routinely overestimate their own agency, claiming authorship of gestures that are not their own.
Second, spatial (Fourneret & Jeannerod, 1998), and even more so temporal (Farrer et al., 2003),
perturbation of the relation between action and visual feedback strongly reduces authorship
judgments. Third, disorganized and excessive acceptance of agency occurs in psychosis (Franck
et al., 2001). These basic processes underlying the sense of agency may in the future be found to
underlay the highest aspects of sense of self, such as self-image (Piryankova et al., 2014).
Agency and Motivation

Sense of agency is also deeply linked to motivation, to our reasons for acting (see Chapter 2 by
Krisst, Montemayor, & Morsella; Chapter 3 by Martiny-Huenger, Martiny, & Gollwitzer; Chapter
5 by Ajzen & Dasgupta; Chapter 7 by Kleiman; Chapter 8 by Dogge & Aarts; Chapter 14 by
Hommel; Chapter 15 by Higgins). First, human actions often aim at goals, often biologically
sensible goals such as food, reproduction, and security. In contrast, most laboratory studies of
sense of agency have considered the direct sensorimotor experience of controlling either ones
own actions, or some arbitrary outcome of ones own action. Therefore, little scientific work has
focused on the relation between the experience of agency and motivational factors such as
outcome valence and reward (but see Dogge & Aarts, Chapter 8 of this volume). Nevertheless,
Page 3 of 6
(p.xi)
Introduction
social psychologists and others working on motivation have recognized two reciprocal cause
links. First, the valence of outcomes may influence ones sense of agency over them. Second, the
sense of agency itself may play an important role in motivating and guiding human behavior.
Regarding the first causal direction, it has been shown that the tendency to overestimate ones
own agency is further amplified when outcomes are positive (Langer & Roth, 1975; Miller &
Ross, 1975; Wohl & Enzle, 2002) and is reduced when they are negative (Morewedge, 2009).
Accordingly, people accept credit for positive outcomes, while denying blame for negative ones,
in a form of self-serving bias (Takahata et al., 2012). Second, motivational constructs such as
intentions and goals strongly contribute to the sense of agency, as several chapters in this book
show (Chapter 2 by Krisst, Montemayor, & Morsella; Chapter 3 by Martiny-Huenger, Martiny, &
Gollwitzer; Chapter 4 by Gilron, (p.xv) Simon, & Mukamel; Chapter 5 by Ajzen & Dasgupta;
Chapter 8 by Dogge & Aarts; Chapter 14 by Hommel).
The relation between agency and motivation is reciprocal, because judgments of agency can also
provide motivation. The obvious link here is through self-efficacy. Classically, this link is made
via peoples conscious beliefs about their capabilities for performing a specific action (Bandura,
1977). However, the relation between self-efficacy and implicit (i.e., unconscious) judgments of
agency remains unclear. A second, even less explored, route linking sense of agency and
motivation may be via action-selection and reward mechanisms (e.g, Samejima, Ueda, Doya, &
Kimura, 2005). As chapters in this volume propose, the fact of control itself, such as choosing, or
self-causing a perceptual event, may be rewarded in the brain (Chapter 6 by Leotti, Cho, &
Delgado; Chapter 12 by Karsh & Eitam). Acquiring and maintaining a sense of agency may
therefore become a desired goal (Chapter 6, Leotti, Cho, & Delgado) or an outcome-independent
driving force (Chapter 12, Karsh & Eitam; Chapter 15, Higgins). Thus, these chapters suggest
that explicit and implicit sense of agency can influence the probability and vigor of further
action (Bednark & Franz, 2014; Eitam, Kennedy, & Higgins, 2013).
Despite the importance of agency in human life, the history of psychology is strangely silent on
the topic. Often, agency was simply considered as an adjunct to the more important question of
volition. This book aims to redress that imbalance, by bringing together some of the experts in
the field to contribute their own unique research perspectives. The astute reader will notice that
the chapters represent two broad traditions of agency research. One is based on computational
brain theory, while the other is based in cognitive and social-cognitive psychology.
Computational brain theory offers the advantage of clear mechanistic models and enables easier
links to neural substrate, while social psychology affords greater (if coarser) generalizations and
far more links to peoples implicit theories of agency with their potential causal roles. We were
motivated to edit this book because the two strands are coming together in current research,
producing some important new insights into this key feature of our psychology. We are grateful
to the contributors for their time in writing these chapters, and we hope that you, the reader,
find the book both useful and enjoyable.
References
Bibliography references:
Page 4 of 6
(p.xi)
Introduction
Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological

Review, 84, 191215.
Bednark, J. G., & Franz, E. A. (2014). Agency attribution: event-related potentials and outcome
monitoring. Experimental Brain Research, 232, 11171126.
Bergson, H. (1998). Creative evolution. Trans. Arthur Mitchell. New York: Dover [1911].
Daprati, E., Franck, N., Georgieff, N., Proust, J., Pacherie, E., Dalery, J., & Jeannerod, M. (1997).
Looking for the agent: an investigation into consciousness of action and self-consciousness in
schizophrenic patients. Cognition, 65, 7186.
Eitam, B., Kennedy, P. M., & Higgins, E. T. (2013). Motivation from control. Experimental Brain
Research, 229, 475484.
Farrer, C., Franck, N., Georgieff, N., Frith, C. D., Decety, J., & Jeannerod, M. (2003). Modulating
the experience of agency: a positron emission tomography study. Neuroimage, 18, 324333.
Franck, N., Farrer, C., Georgieff, N., Marie-Cardine, M., Dalry, J., dAmato, T., & Jeannerod, M.
(2014). Defective recognition of ones own actions in patients with schizophrenia. American
Journal of Psychiatry.
Fourneret, P., & Jeannerod, M. (1998). Limited conscious monitoring of motor performance in
normal subjects. Neuropsychologia, 36, 11331140.
Gallup, G. G. (1970). Chimpanzees: self-recognition. Science, 167, 8687.
Langer, E. J., & Roth, J. (1975). Heads I win, tails its chance: the illusion of control as a function
of the sequence of outcomes in a purely chance task. Journal of Personality and Social
Psychology, 32, 951955.
Miller, D. T., & Ross, L. (1975). Self-serving biases in the attribution of causality: fact or fiction?
Psychological Bulletin, 82, 213225.
Morewedge, C. K. (2009). Negativity bias in attribution of external agency. Journal of

Experimental Psychology: General, 138, 535.
Piryankova, I. V., Wong, H. Y., Linkenauger, S. A., Stinson, C., Longo, M. R., Bulthoff, H. H. J.,
Mohler, B. J. (2014). Owning an overweight or underweight body: distinguishing the physical,
experienced and virtual body. PLoS ONE, 9, e103428.
Samejima, K., Ueda, Y., Doya, K., & Kimura, M. (2005). Representation of action-specic. reward
values in the striatum. Science, 310, 13371340.
Sperry, R. W. (1950). Neural basis of the spontaneous optokinetic response produced by visual
inversion. Journal of Comparative and Physiological Psychology, 43, 482489.
Page 5 of 6
(p.xi)
Introduction
Synofzik, M., Vosgerau, G., & Newen, A. (2008). Beyond the comparator model: a multifactorial
two-step account of agency. Consciousness and Cognition, 17, 219239.
Takahata, K., Takahashi, H., Maeda, T., Umeda, S., Suhara, T., Mimura, M., & Kato, M. (2012).
Its not my fault: postdictive modulation of intentional binding by monetary gains and losses.
PloS One, 7, e53421.
Von Holst, E. and Mittelstaedt, H. (1950). Das Reafferenzprinzip. Die Naturwissenschaften, 20,
464476.
Wohl, M. J. A., & Enzle, M. E. (2002). The deployment of personal luck: illusory control in games
of pure chance. Personality and Social Psychology Bulletin, 28, 13881397.
Page 6 of 6
Time to Act

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Time to Act
The Dynamics of Agentive Experiences
Elisabeth Pacherie
DOI:10.1093/acprof:oso/9780190267278.003.0001
Abstract and Keywords

Actions unfold in time, and so do experiences of agency. Yet, despite the recent surge of interest
in the sense of agency among both philosophers and cognitive scientists, the import of the fact
that agentive experiences unfold in time remains to this day largely underappreciated. This
chapter argues that agentive experiences should be conceptualized as continuants, whose
contents evolve as actions unfold. It attempts to characterize these content shifts, distinguishing
two main dimensions of changechanges in scale, or fine-grainedness, and changes in tenseas
well as the main action control and action specification processes that underlie them. The
chapter further argues that taking into account this temporal dynamics of agentive experiences
can help researchers better appreciate in what sense some of the apparently conflicting
empirical models of the sense of agency proposed in recent years can be seen as
complementary, rather than as rival, thus refining integrative models.
Keywords: sense of agency, agentive experiences, action, temporal dynamics, content shifts
This chapter is dedicated to the memory of Marc Jeannerod.
Introduction
Even the briefest of actions, such as pressing a button, flexing the wrist, or raising an arm, takes
time. Before one even acts, one has to decide what to do and when. When the time to act has
come (which may be immediately following the decision, as when one acts on the spur of the
moment), and provided one recognizes that it has come, the action starts and then progresses
Page 1 of 19
Time to Act
until it either reaches successful completion, or, as the case may be, is aborted or ends in
failure. Obviously then, actions unfold in time, but so do experiences of agency or, as I shall call
them, agentive experiences.
Despite the recent surge of interest in the sense of agency among both philosophers and
cognitive scientists, the import of the fact that agentive experiences unfold in time remains to
this day largely under-appreciated. This is not to say that the temporal characteristics of
agentive experiences have been completely neglected. Indeed, one of the main findings of
experimental investigations into the sense of agency is the phenomenon of intentional binding,
first reported by Patrick Haggard and his colleagues (Haggard, Clark, & Kalogeras, 2002),
whereby a voluntary action and its external sensory consequences are (p.4) compressed
together in subjective time. The effect of voluntary action on subjective time perception is not,
however, what I am primarily interested in. Rather, my main concern is with how the contents
and the strength of agentive experiences evolve with time. While I will suggest a number of
ways in which the contents of agentive experiences can shift as an action unfolds, as well as a
number of factors that can influence these shifts, I offer these as starting points for future
research. My hope is to convince cognitive scientists that these issues are important, that many
questions are yet unanswered, and that they are worth investigating empirically.
I start with a preliminary characterization of agentive experiences in the following section. I

then present a brief survey of recent empirical work on the sense of agency. Next, I discuss the
contents of agentive experiences and try to characterize the two main dimensions along which
these contents shift as the actions unfold, namely changes in scale or level of grain and changes
in tense, as well as the factors that influence these shifts. Finally, I argue that taking into
account the time dynamics of agentive models can also help us better appreciate in what sense
some of the apparently conflicting empirical models of the sense of agency proposed in recent
years can be seen as complementary rather than as rival.1
Agentive Experiences
States of agentive self-awareness comprise both agentive beliefs and agentive experiences
(Bayne & Pacherie, 2007). Agentive experiencethat is, our moment-by-moment sense of
ourselves as performing various actionsconstitutes our basic form of agentive awareness. Our
agentive beliefs, while often based on our agentive experiences, can also be based on more
indirect evidence. For instance, if I find my car keys in the fridge, I may form the belief that I
was the one who put them there, despite having no recollection of having done so. In this case
my agentive belief will be based on other doxastic states, such as my belief that I was alone at
home and perhaps my further belief that I have a pronounced tendency toward absent-
mindedness, rather than on any specific agentive experience or memory thereof.
Here, my focus will be on agentive experiences. The burgeoning literature on the content and
sources of agentive experiences highlights their many facets. A non-exhaustive list of proposed
distinctions includes experience of deliberation, experience of decision, awareness of a goal,
awareness of an intention to act, awareness of initiation of action, awareness of movement,
sense of activity, sense of mental effort, sense of physical effort, sense of control, experience of
authorship, experience of intentionality, experience of purposiveness, experience of freedom,
and experience of mental causation. To avoid (p.5) becoming overwhelmed by this
Page 2 of 19
Time to Act
terminological profusion, it may be useful, following Bayne (2010), to distinguish between core
and non-core elements of agentive experiences. Core elements are elements that must be
possessed by any agentive experiences whatsoever and that are essential to agentive
experiences; non-core elements are elements that may but need not be present within
experiences of agency.
Among non-core elements of agentive experiences, one may include, among other things,
experience of effort, experience of deliberation, or experience of decision. Arguably, not all
actions are effortful or felt as such. Indeed, one finds reports in the clinical literature of people
who after brain damage have lost the capacity to experience actions as effortful but still enjoy
agentive experiences (Naccache et al., 2005). It is also not uncommon for one to act on the spur
of the moment, or for actions, especially routine actions, to be triggered more or less
automatically by cues in the environment. We can have agentive experiences for these actions,
but these will not include experiences of deliberation or of decision. This is often the case for
routine actions that can be triggered more or less automatically. For instance, I found myself a
few minutes ago cleaning my glasses. As soon as I became aware of what I was doing, I had an
agentive experience for the action I was performing, but it did not include among its elements
an experience of deliberating whether or not to clean my glasses or even deciding to do so.
Other elements, however, are essential to agentive experiences, and in what follows I will
concentrate on these core elements. At the very minimum, to have an agentive experience is to
be aware of oneself as acting, that is, to experience a sense of agency, narrowly conceived. Note
that in saying that the sense of agency narrowly conceived is a core component of agentive
experiences, I am not implying that this core component is itself necessarily atomic. Indeed, as
we will see in the next section, it is often proposed that the sense of agency is itself
decomposable into at least a sense of initiation and a sense of control.
It is difficult, however, to conceive of a sense of agency or a feeling of doing as existing

independently of any experience of what it is that one is doingas a feeling simply floating in
the air, so to speak. It seems that an agentive experience also minimally involves some sense of
what it is one is doing, whether it be an experience of moving ones body in a certain way (e.g.,
moving a finger), of performing an action of particular kind (e.g., pressing a key), or of acting
with a certain goal (e.g., starting a new paragraph). Having an agentive experience does not
require one to identify the action the experience is about at a particular level of specification,
but it demands that the action be specified in some way, however vague the specification is. So I
take it that agentive experiences include two core elements: a sense of agency, and some
specification of the action that the sense of agency is for.2
(p.6) It may be objected that in routine actions one typically has at least a minimal awareness
of oneself as acting, but not an experience of what one is doing.3 In answering this objection, it
is first important to note that the category of routine actions is a mixed bag. Actions such as
playing with a hair lock or tapping ones foot while attending a lecture are sometimes classified
as routine actions, but so is driving (at least for experienced drivers) or coffee making if you do
it every morning. Routine actions of the first kind are typically involuntary and also unconscious
in the sense that we lack even a minimal experience of ourselves as acting. Consider, for
instance, the following dialogue:
Page 3 of 19
Time to Act
Could you please stop that?

Stop what?
This constant tapping with your pen on the table.
Oh, Im sorry. I had no idea I was doing that.
One should also note that lacking an experience of acting is not tantamount to having an
experience of passivity. An alternative to both is the lack of any experience. Thus, the objection
does not seem to apply to this first kind of routine actions, if, as I suggest, we are simply
unaware that we are acting when performing them, and thus lack both a sense of agency stricto
sensu and a sense of what it is we are doing.
Now, does the objection hold for routine actions of the second kind, actions that are voluntary
but that one can perform without having to attend to what one is doing? The key word here is
attention. There is a complex ongoing debate on the relationship between attention and
consciousness, with some holding that there is no consciousness without attention (ORegan &
Noe, 2001; Posner, 1994), and others holding that the two processes can be dissociated and thus
that there can be conscious awareness without attention (Koch & Tsuchiya, 2007; Lamme, 2003;
Wyart & Tallon-Baudry, 2008).4 I am not taking a position on this debate here, but what I want
to do is sketch two stories about what might be going on in routine actions that preserve the
two-core-component view of agentive awareness.
The first story is meant to be compatible with the view that there can be some conscious
awareness without attention. This conscious awareness would be marginal, however, and would
lack depth. For instance, when I am in the subway reading the newspaper, I am marginally
aware of people around me, but my awareness of them is not such that I could recognize them if
I were to meet them later elsewhere. Similarly, it may be that when I am performing a routine
action I have some marginal awareness of myself as acting and also some marginal awareness of
what I am doing, but that I consciously access only some general features of the action
representations that drive my action (p.7) and have therefore an agentive experience of what I
am doing that is largely underspecified.
The second story is more radical and is meant to be compatible with the no-consciousness-
without-attention view. Heres how it goes. When one is performing a routine action without any
attentional resources being engaged, one has no agentive experience for that action, not even a
minimal sense of oneself as acting. However, (some of) the action representations that drive the
action are poised for consciousness. As soon as, for whatever reason, attention becomes
engaged, one starts enjoying some awareness of oneself as acting and becomes aware of at least
some aspects of what one is doing. On this story, the idea that during the course of routine
actions we always enjoy some sense of agency is an illusion. It is, so to speak, an instance of the
fridge light illusion. In the same way that we may come to mistakenly believe that the light is
always on in the fridge because it is always on when we open the fridge, we may come to
mistakenly believe that we always enjoy an experience of acting for routine action, because each
time we attend to the action, we have the experience. This illusion would be reinforced by the
fact that it is rarely the case that we deploy no attentional resources whatsoever while
performing a routine action. Even though making coffee is something I do every morning, I still
have to pay attention to how much water I pour in the coffee maker, how I position the cup, and
Page 4 of 19
Time to Act
so on. According to this second story, the question to ask is not at how many (and at which) of
the stages of action an ongoing agentive experience must include an experience of what one is
doing, but rather at how many (and at which) of the stages of action there must be an agentive
experience.
Which of these two stories, if any, is correct remains at present an open and a difficult empirical
question. My point here is simply that both stories can accommodate the claim that an agentive
experience does not reduce to a pure, or naked, sense of agency but also involves some
awareness of what one is doing, however partial and unspecific the content of this conscious
representation. Both stories suggest that when you interrupt a routine action of mine and ask
me what I am doing, I may give an answer by reporting the contents of the action
representations that were active just before you interrupted me and that remain active and thus
can be attended to precisely because you interrupted me and thus prevented their being erased
by corresponding reafferences. In addition, the consciousness-without-attention view allows for
some marginal sense of agency and some marginal awareness of what one is doing, even when
no attentional resources are engaged; on this view, the deployment of attentional resources
simply transforms and enriches our agentive experience, making it more vivid and allowing us to
consciously access richer and more detailed representations of what we are doing. In contrast,
on (p.8) the no-consciousness-without-attention view, unless attentional resources are
deployed, we simply have no agentive experience whatsoever: we experience no sense of agency
and are not aware of what we are doing. However, when attentional resources are deployed we
enjoy both a sense of agency and a sense of what we are doing, and the more attentional
resources are deployed the richer and more detailed the contents of our conscious experiences
become.
The Sources of Agentive Experiences

Empirical research on agency has explored a variety of potential cues to agency, and a number
of different models of how agentive experiences are generated have been proposed. Typically,
these models focus more on core elements of agentive experiences than on non-core elements,
and, within these core elements, also tend to concentrate more on the sense of agency narrowly
conceived than on how actions are represented in agentive experiences. These models all rely to
a greater or lesser extent on a congruence principle: the sense of agency is produced when
there is a match between cues x and y. The points on which they differ concern the nature of the
cues being compared, the nature of the processes involved in the production of the sense of
agency, and how closely these processes are related to action production and control processes.
Two theoretical positions define the two ends of the spectrum of possibilities: the motor
prediction view and the cognitive reconstruction view. On the motor prediction view, the sense
of agency is generated by processes dedicated to action control. On the cognitive reconstruction
view, the sense of agency is generated by general-purpose processes of retrospective causal
inference.
The motor prediction view is inspired by computational theories of motor control. According to
these theories, when the motor system generates a motor command, an efference copy of this
command is sent to forward models whose role is to generate predictions about its sensory
consequences in advance of actual execution. Error signals arising from the comparison of
Page 5 of 19
Time to Act
desired, predicted, and actual states (as estimated from sensory reafferences) are used to make
corrections and adjustments. The motor prediction view holds that the signals used for motor
control also provide cues to agency (Frith et al., 2000). In particular, it holds (1) that awareness
of initiating an action is based on a representation of the predicted consequences of making that
action, rather than its actual consequences, and on the congruence between the predicted state
and the desired state, and (2) that for this experience of agency to continue, the predicted
consequences would also have to remain congruent with the sensory reafferences when they
become available. These two claims are associated with the idea that the sense of agency has
two main components: the sense of initiation linked to the congruence between the predicted
state and the desired (p.9) state, and the sense of control linked to the degree of congruence
between the predicted state and the actual state. Indeed, more recent empirical work within the
motor prediction framework suggests a more complex picture of the determinants of the sense
of agency, highlighting in particular the contribution of action selection fluency, and thus
potentially of metacognitive processes, to the sense of agency (Chambon et al., 2014; Chambon
& Haggard, 2012; Sidarus et al., 2013; Wenke et al., 2010).
Claim (1)and therefore the possibility that the sense of agency can emerge in advance of
actual sensory effect and can be based on premotor processes aloneis supported by evidence
that awareness of initiating a movement in healthy subjects is reported by the agent between 80
and 200 milliseconds before the movement actually occurs (Haggard & Eimer, 1999; Libet,
1985; Libet et al., 1983) or, in the case of patients with anosognosia for hemiplegia, can be
reported despite no movement being actually produced (Berti et al., 2008). Evidence for claim
(2)that the sense of agency also depends on the congruence between predictions and sensory
reafferencescomes from studies where these reafferences are artificially manipulated by
introducing temporal delays and spatial distortions of feedback. These studies demonstrate that
the sense of agency is gradually reduced as these discrepancies increase (Fourneret &
Jeannerod, 1998; Knoblich & Kircher, 2004; Sato & Yasuda, 2005).
In contrast, the cognitive reconstruction view downplays the contribution of the motor system to
the sense of agency and proposes that it is inferred retrospectively from the existence of a
match between a prior thought and an observed action. Thus, on Wegners theory of apparent
mental causation (Wegner, 2002), a general-purpose causal inference process is at play. If an
action is consistent with a prior thought of the agent and other potential causes of the action are
not present or salient, a sense of agency for the action will be induced.
There is also empirical evidence that high-level inferential processes play a role in determining
the sense of agency for an action. Studies of Wegner and colleagues have demonstrated that
cognitive cues can alter the sense of agency for an action independently of changes in
sensorimotor and perceptual cues. For instance, in their I-Spy study (Wegner & Wheatley,
1999), a participant and a confederate of the experimenter had joint control of a computer
mouse that could be moved over any one of a number of pictures on a screen. When participants
had been primed with the name of an item on which the mouse landed, they expressed a
stronger sense of agency for the action of stopping on that object (when in fact the stop had
been forced by the confederate). Further studies also suggest that subliminally priming an
outcome just before the outcome is produced can enhance the sense of agency for that outcome
Page 6 of 19
Time to Act
(Aarts et al., 2005) and that priming an outcome relatively far in advance can (p.10) augment
self-agency as well, but only if the outcome is attached to positive affect (Aarts et al., 2009).
There is now a growing consensus that the motor prediction view and the cognitive
reconstruction view are not mutually exclusive but complementary and that intrinsic cues (cues
provided by the motor system) and extrinsic cues (such as cognitive primes) both contribute to
the sense of agency (Moore et al., 2009; Moore & Fletcher, 2012; Pacherie, 2008, 2010; Sato,
2009; Synofzik et al., 2008). Researchers are now trying to develop integrative frameworks and
to get a better understanding of how all these agency cues interact.
One way to try to combine the motor prediction view and the cognitive reconstruction view is to
appeal to the distinction between pre-reflective agentive experiences and reflective agentive
beliefs or judgments (Bayne & Pacherie, 2007; Gallagher, 2007; Haggard & Tsakiris, 2009) and
to argue that while motor processes contribute mainly to agentive experiences, interpretive and
general inferential processes contribute mainly to agentive beliefs or judgments. Thus, the
example given at the beginning of the previous section of my finding my car keys in the fridge
and judging that I must have put them there myself may constitute an instance where an
agentive judgement is not based at all on an agentive experience but is the result of inferential
processes exploiting general causal knowledge about the world. This, I know that car keys dont
move on their own and that no one else was around, leading me to conclude that I must have
done it. Typically though, we may expect judgments to be based at least in part on agentive
experiences when such experiences are available.
This conceptual distinction between agentive experiences and judgments is echoed

methodologically in the ways agency is measured in experimental studies. While some studies
(Farrer et al., 2003; Metcalfe & Greene, 2007; Sato & Yasuda, 2005) investigate agency by
asking participants to explicitly judge whether they caused a particular sensory event, other
studies use implicit agency measures such as intentional binding and sensory suppression.
Intentional binding is a phenomenon, first reported by Haggard and his colleagues (Haggard et
al., 2002), whereby an action and its external sensory consequences are compressed together in
subjective time. As intentional binding occurs only for voluntary actions (Tsakiris & Haggard,
2003) and is furthermore modulated by the statistical relation between events (Moore &
Haggard, 2008), it is considered to provide an implicit measure of agency. Sensory attenuation
of self-produced action effects has also been used as an implicit measure of agency. When the
internally generated motor predictions about the sensory consequences of ones ongoing actions
and their actual sensory consequences are congruent, the sensory percept is attenuated,
thereby enabling a differentiation between self-generated and externally (p.11) generated
sensory events (Blakemore et al., 2000). However, recent studies showing that prior authorship
beliefs can modulate both sensory attenuation and intentional binding (Desantis et al., 2011;
Desantis et al., 2012) suggest that drawing a sharp distinction between agentive experiences
supported by motor processes and agentive beliefs supported by interpretive processes may be
over-simplistic.
A promising approach is to appeal to a Bayesian integrative framework involving a hierarchy of

prediction and model building. Thus, Moore and Fletcher (2012) propose that the sense of
agency is determined by a Bayesian process of cue integration, where the predictions generated
Page 7 of 19
Time to Act
at higher levels of the hierarchy provide the priors for the lower levels, that is, constrain the
interpretation of cues available at lower levels. In this model, cue integration is itself the
product of both the strength of the priors and the weights attached to the available cues as a
function of their reliability. When priors are weakas, for example, when one is quite unsure
what the effects of pressing a button will beone may still have a strong sense of agency for the
ensuing consequence, provided that perceptual reafferences carrying information about it are
very reliable. Conversely, if my priors are very robust, I may have a strong sense that I produced
a certain effect in the world, even though the feedback I get is weak or ambiguous. When both
priors and reafferent cues are weak, my sense of agency may be correspondingly weakened.
While this Bayesian approach does not allow for a sharp distinction between agentive
experiences and agentive judgments, it can accommodate the idea that high-level priors exert
more influence on agentive judgments than on agentive experiences. As we will see later in this
chapter, another important parameter we must take into account when trying to understand
how cue integration works is time and, in particular, the different times at which different cues
become available.
Agentive Experience: Content Shifts

I proposed earlier that agentive experiences include two core elements: a sense of agency and
some specification of the action that the sense of agency is for. I pointed out, however, that
having an agentive experience does not require one to identify the action that the experience is
about at a particular level of specification. Not imposing such a requirement makes sense if
agentive experiences evolve in time and if their intentional contents change as actions unfold.
My aim in this section is to characterize the kinds of changes the intentional contents of
agentive experiences can undergo, to explore the ways in which these changes relate to the
hierarchical structure of action specification and control processes, and to discuss additional
factors that could also modulate which action representations enter the contents of agentive
experiences.
(p.12) Contents-shifts in agentive experiences can involve both changes in tense and changes
in the level of grain at which the action is identified. With respect to tense, one can have an
experience of being about to A, of initiating A, of being in the middle of A-ing, of nearing the
completion of A-ing, and of having just A-ed. We should expect that, as an action unfolds, there
should be tense-shifts in the contents of the agentive experience that accompanies it.
Additionally, if one considers actions somewhat more complex than simply flexing a wrist or
raising an arm, it becomes clear that multiple tenses can in principle coexist within the same
agentive experience. Suppose that I am preparing a chocolate mousse; my agentive experience
at a given moment may be of having melted the chocolate and being about to whip the egg
whites. A second form of content-shift concerns the level of grain at which the action is
identified. As the chocolate mousse example can also serve to illustrate, I can experience my
action as preparing a chocolate mousse, whipping the egg whites, rotating my wrist with great
energy, and so on. More generally, it seems that we can zoom in and out as we perform an
action, focusing at some points on specific details of action execution and at other points of
more abstract aspects of our action plan.
All of this sounds commonsensical enough, or so I hope. But can we say more about what makes
one identify an action at a given level of grain, or what makes one concentrate on the past, the
Page 8 of 19
Time to Act
present, or the future aspects of the action being performed? As I remarked earlier, empirical
work on agentive experiences tends to be more concerned with investigating the cues and the
processes that generate the sense of agency than with studying how actions are represented in
agentive experiences. Nevertheless, it is quite plausible that the various cues that contribute to
the sense of agency also contribute to an agents awareness of what he or she is doing. As the
discussion in the previous section indicates, a variety of agency cues contribute to the sense of
agency and, as hierarchical cue integration models suggest, different cues may pertain to
different levels in a hierarchy of action representations and predictions. One early proponent of
this idea was Marc Jeannerod (1995, 1997), who captured it in the form of a flow chart model,
reproduced as Figure 1.1. In what follows, I propose to use this model as a guide to address the
following questions: What are the action representations available at given stages of action
preparation and execution? When should we expect an agentive experience to represent the
whole action? When should we expect it to represent a particular component of an action? When
should we expect transitions from one component to another or from one level of representation
to another (zooming in and zooming out) to take place?
In his comments on this flow chart,

Jeannerod (1995, 1997) makes several
important points. First, while the main flow
of information goes through the (p.13)
cognitive steps of action specification (e.g.,
intending, planning, programming, and
executing), Jeannerod notes that the
impression given by this flow-chart
representation that these steps are
organized serially must be corrected. In
other words, while the initial component or
segment of an action is prepared and Figure 1.1 A flow-chart model of the
executed, the preparation of the next ongoing regulation of action.
components of the action is already Reproduced from Jeannerod, 1995.
engaged. As he puts it, the diagram, at
least for what concerns its distal steps,
should be considered as one of many
parallel lines with a similar structure (1997, p. 172). Second, the predictions made are stored in
as many short-term memories (represented as M1, M2, and M3 in Figure 1.1) as there are
levels, and reafferences documenting the current state of the action are later fed into the
memories. If the action has been completed, the reafferences and the content of the memories
will coincide, and the latter will be erased. If the action is incomplete, the memories remain
active, which reactivates the corresponding module and generates corrections. While Jeannerod
himself does not use this terminology, the mechanism he proposes is akin to the comparison
processes appealed to in the motor prediction view discussed in the previous section.
The idea of a set of short-term memories storing predictions and remaining activated until
matching reafferences erase their contents gives us a way of understanding which
representations of an action are active at a given moment and can thus potentially enter the
contents of agentive experiences. This model suggests certain constraints on which action
Page 9 of 19
Time to Act
representations can contribute to (p.14) agentive experience at a given moment, and suggest
more generally that our agentive experiences should tend to be present and future oriented,
given that short-term memories for actions or action components already completed are
promptly erased. Yet, it still implies that many representations of an action can be
simultaneously active, first because, as Jeannerod puts it, in a given line, corresponding to the
preparation of a given component or segment of an action, several memories storing
representations of this segment at different levels of specification can be active, and second,
because this line is only one of possibly many such lines where memories are also currently
active. Yet, it seems unlikely that all the action representations active at a given time contribute,
or at least contribute equally, to agentive experiences.
Suppose action representations are, as the model suggests, arranged in a hierarchy, with the
more abstract representation of the action at the top of the action tree (e.g., making coffee),
representations of component actions (e.g., filling the water reservoir of the coffee machine,
installing a new filter in the filter basket, putting coffee in the filter, turning the machine on,
etc.) at the next level, representations of subcomponents (e.g., fill coffee pot with required
amount of water, open water reservoir, pour the water from the coffee pot into the reservoir,
etc.) one further level down, then representations specifying the exact effectors used (e.g., open
water reservoir with right hand, and pour water with left hand, or vice versa), and so on until
one reaches the level of basic motor actions. One prediction we can make is that while a number
of different, more or less specific, action representations would be active at any point during the
action, the most active at a given point would be those that are most directly relevant in guiding
the transition from the action segment currently being performed to the next. Using tree
terminology, the representations most relevant to guiding the transition from the current
terminal node of the action tree to the next adjacent terminal node are those attached to the
closest common ancestor node of these two terminal nodes. If these two nodes happen to belong
to the same sub-branch of the tree, the representation at this common ancestor note should
remain relatively specific; if the two nodes are on two different sub-branches, they will have as a
common ancestor a more abstract action representation. In other words, one would predict that
the abstract representation preparing coffee would be more readily available when one has
put coffee in the filter and is about to turn the machine on, than when one is opening the water
reservoir and is about to pour water into it. One way to test such predictions, and thus to
explore to what extent our awareness of what we are doing is a reflection of how we navigate a
tree structure of action representations, would be to have participants perform such action
scripts, interrupting them at various junctures and asking them to answer a yes/no question as
quickly as possible (e.g., are you making coffee? are you filling the water (p.15) reservoir?
are you raising your right arm?). If we assume that they would be faster at answering a
question when its content corresponds to the content of their agentive experience at the time of
the interruption, their reaction time would presumably give us an indication of what the
contents of agentive experience are and how they shift at various points during action execution.
Several other factors can further influence which action representations enter the contents of
agentive experiences. First, although, as I just suggested, we normally attend to what we must
do next and navigate our action representation tree accordingly, what we are attending to can
also be modulated by endogenous and exogenous factors. On the one hand, we can voluntarily
allocate attention to some aspects of an action more than to others. For instance, when I play a
Page 10 of 19
Time to Act
serve in tennis, I can choose to attend to distal aspects of the action, (e.g., where in the court
the ball lands), or to more proximal aspects of the action (e.g., how I move my arm and
shoulder). On the other hand, stimulus events can also involuntarily capture attention. The
short-term memories discussed by Jeannerod are short-lived, since they are erased as soon as
matching reafferences become available. The longer they last, the more likely it is that they will
capture attention, since their longevity is a sign of trouble.
One further factor that may modulate the content of our agentive experiences is the agents
preferred level of action identification. According to Action Identification Theory (Vallacher &
Wegner, 1985, 1989), any behavior can be represented at multiple levels, but there are
individual differences in agents preferred level of action identification. Some people show a
general tendency to (consciously) represent their behavior at a low level, that is, in terms of the
concrete, mechanistic aspects of their action (e.g., turning a doorknob), whereas other people
show a general tendency to represent their behavior at a higher level, that is, in terms of the
goals or purposes of their action (e.g., opening the door or going out). While other factors can
also influence the level at which an action is represented, still, ceteris paribus, agents whose
preferred level of action identification is low will tend to experience their action in terms of
producing sensorimotor consequences or outcomes, while agents whose preferred level of action
identification is high will tend to tend to experience their action in terms of the goal it serves.5
Yet another factor is the level of skill of the agent. When an action is overlearned and has
become routinized, the transition between action segments or components can proceed more or
less automatically. If, however, the agent is not performing a routine, he or she may have to
consciously consider what comes next when a given segment is completed. However, it can be
argued that what characterizes expert performance is not so much that it can proceed
automatically, without the agent needing to pay attention to what he is doing or to transitions
between action segments, as the fact that the agent can flexibly (p.16) navigate between levels
of action representations, attending to global aspects of his performance as well as to details of
execution. On the one hand, there is evidence that highly skilled performers are able to attend to
details of motor execution that remain inaccessible to less proficient performers (Knoblich &
Repp, 2009). On the other hand, while less proficient performers are often unable to execute
certain actions without consciously attending to what they are doing, expert performers may be
able to perform them without mobilizing attentional resources, focusing their attention instead
on more global or distal aspects of their performance. Thus, the expert piano player may be able
to concentrate her attention on the phrasing of a passage, while all the attention of the novice
player is concentrated on how to play the next chord. Expertise is thus characterized both by the
range of action representations the expert can consciously access, from low-level sensorimotor
representations to much more global and abstract representations of his or her action, and by
the flexibility with which he or she can navigate between representations at various levels.
While the less proficient agent may be forced to attend to representations at a certain level, the
expert can choose what to attend to.
It is a controversial issue whether low-level sensorimotor representations can contribute

anything to agentive experiences, given that detailed sensorimotor specifications of actions and
sensorimotor reafferences may both remain below the threshold of awareness. This is to a large
degree an empirical question and I therefore limit myself to some tentative remarks. For a start,
Page 11 of 19
Time to Act
the question itself is in need of clarification. Let me throw in three distinctions. First, we can
distinguish between contributions to the core components of agentive experiences and
contributions to their non-core components. Second, we can also distinguish between
contributions to one or the other of the two core components of agentive experiences: sense of
agency and sense of what it is that one is doing. Third, we can distinguish between direct and
indirect contributions to the contents of agentive experiences, where by direct contributions I
mean that an action representation contributes (some of) its contents to the intentional content
of agentive experiences.
Presumably, the version of the question that is most controversial is whether low-level
sensorimotor representations can make a direct contribution to the intentional content of an
agentive experience. For instance, there is experimental evidence that when performing a target
reaching task, subjects can make rapid corrections of their hand movements in response to
biased feedback on their hand trajectory without being aware that they are making these
corrections (Fourneret & Jeannerod, 2008). These results suggest that the motor system
constructs precise sensorimotor representations, since small discrepancies between predicted
and actual sensorimotor reafferences are detected and used to make appropriate corrections of
the hand movement, (p.17) but that these sensorimotor representations are not consciously
accessed, since subjects are not aware of these discrepancies and are not aware that they are
making corrections. However, evidence from expert performance suggests that highly skilled
performers acquire very specific sensorimotor mappings and may be able to attend to details of
execution and to access sensorimotor representations that remain inaccessible to less skilled
agents (Knoblich & Repp, 2009). Whether or not certain sensorimotor representations directly
contribute to our awareness of what we are doing may thus depend on the agents level of skill
or expertise. This is not to say, however, that experts have unlimited access to their
sensorimotor representations. Distinguishing between high-level cognitive representations, mid-
level perceptual representations, and low-level sensorimotor representations is probably a
simplification, as each of these three layers of action representations is likely to contain many
sub-layers. It may thus remain true that even for experts, very low-level sensorimotor
representations remain below the threshold of awareness and thus make no contributions to our
awareness of what we are doing.
With regard to the possible indirect contribution of sensorimotor representations to the contents
of agentive experiences, it seems plausible that prediction errors arising from mismatches
between sensorimotor predictions and sensorimotor reafferences could contribute to attentional
amplification at the next level and thus to an increased awareness of what is going on as it is
represented at the next level. In such a scenario, low-level sensorimotor predictions and
reafferences would have an effect on how rich the contents of an agentive experience would be,
not by contributing their own contents, but because their comparison generates prediction
errors that trigger the deployment of attention resources one level upward. Finally, there are
two further contributions these sensorimotor representations may make to agentive
experiences. First, it may also be that the number and strength of the error signals generated at
the sensorimotor level modulate to some extent the sense of agency for an action. Agents may
not have conscious access to what went wrong at the sensorimotor level, yet may have a sense
that something was not quite right. Second, and perhaps relatedly, the absence or near absence
Page 12 of 19
Time to Act
of error signals at that level may also contribute to the sense of flow that can accompany
flawlessly executed actions.
To recap, hierarchical models of action organization suggest that action is guided and controlled
by multiple action representations, representing different components of the actions and
representing them at various levels of grain, and that many such representations can be
activated in parallel. Following Jeannerod, I suggested that the action representations that
contribute to the contents of agentive experiences should be predominantly forward-looking,
that is, they should be mostly representations of what the agent is currently (p.18) doing or
about to do. I also suggested that the action representations most active at a given point should
be those that are most directly relevant in guiding the transition from the action segment
currently being performed to the next and that their degree of abstractness should depend on
whether these two segments belong or not to the same sub-branch of the action representation
tree. In other words, the changing contents of our agentive experiences should reflect the way
in which we navigate an action representation tree when performing an action, zooming in when
performing a particular component and zooming out at transition points between components.
This idea is in principle empirically testable and I mentioned one way it could be tested.
However, other factors may also modulate the contents of agentive awareness, including
endogenous and exogenous factors affecting what the agent is attending to, the agents
preferred level of action identification, and her degree of expertise.
Sense of Agency: Variations in Sources and Strength

Earlier in this chapter, I distinguished two core components of agentive experience: awareness
of oneself as acting (i.e., the sense of agency narrowly conceived) and awareness of what one is
doing (i.e., the intentional content of agentive experience). In the previous section, I discussed
shifts in the intentional contents of agentive experiences. In what follows, I come back to the
sense of agency.
First, as I mentioned earlier, empirical research on agentive experiences has concentrated more
on the sense of agency stricto sensu than on the content of agentive experiences. This is not to
say that it has completely neglected content. For instance, Haggard and Eimer (1999) asked
participants to choose voluntarily between left and right key-press actions in a modified version
of Libets task and showed that awareness of initiating an action correlated not with the general
readiness potential but covaried with the lateralized readiness potential (the later phase of
preparation, in which brain activity contralateral to the selected hand exceeds ipsilateral
activity), suggesting that the awareness of initiating an action is tied to the preparation of a
specific body movement rather than to a general preparation to move. Experimental studies
have also examined the effects of action identity manipulations on intentional binding and
sensory attenuation (for a review, see Hughes et al., 2013). However, existing experimental
studies typically manipulate action representations in order to assess the effects of these
manipulations on the sense of agency narrowly conceived, rather than to investigate their
impact on the contents of agentive experiences.
(p.19) Second, as the earlier discussion also indicated, one important area of disagreement in
empirical research concerns the nature of the processes underpinning the sense of agency: Are
these processes essentially predictive or postdictive? Are the cues or representations they
Page 13 of 19
Time to Act
exploit rather low-level or high-level? While a consensus seems to be emerging that both kinds
of processes may be at play and that a variety of lower- and higher-level agency cues contribute
to the sense of agency, the integrative frameworks that have been proposed thus far still lack
sufficient specificity. In particular, they do not sufficiently take into account the temporal
dimension of the phenomenon at stake and thus fail to avail themselves of one important source
of constraint. For instance, while Bayesian hierarchical frameworks propose that the sense of
agency is the product of the interaction between sensory or perceptual (bottom-up) evidence,
weighted according to its reliability, and prior (top-down) expectations weighted according to
their strength, they do not explicitly take into account the temporal properties of these different
sources of information and therefore the way in which the sense of agency and its sources may
evolve as the action unfolds.
My main point here is very simple: when thinking of the sense of agency as determined by a
Bayesian process of cue integration, one should not lose sight of the fact that these cues are not
all available at the same time. Both higher- and lower-level priors and predictions based on them
are available from the start, while feedback and signals arising from the comparison of feedback
with predictions only become available at later stages. We should therefore expect our sense of
agency to be mostly based on predictive processes and prospective cues in the early stages of an
action, with a progressive shift toward post-dictive processes and retrospective cues as the
action unfolds. We should also expect our sense of agency to fluctuate as the action unfolds,
according to the congruence or lack thereof between prospective and retrospective cues and to
their respective strength. For instance, our sense of agency may initially be based on weak
priors and may be correspondingly weak, but may increase as reafferences that match our
expectations become available. Conversely, our sense of agency may initially be based on strong
priors and thus be strong, but may decrease when we receive reliable incongruent feedback. As
proposed by Moore and Haggard (2008), retrospective inferential processes may kick in when
predictions are invalidated by reafferences, in order to ensure that our agentive experience is a
correct reflection of the actual relations between our actions and their consequences. What
these retrospective inferential processes do, in such a case, is tie efferent signals to their actual
effects in the world. In so doing, they help us regain a sense of agency over our actions by
changing the contents of our agentive experience and improving their accuracy.
(p.20) Moore and Haggards proposal also highlights one way in which the two core
components of agentive experience, their contents and their strength, can interact. This
interaction could also be manifested in other ways. I argued earlier that the intentional content
of an agentive experience changes as the action unfolds and that such changes may reflect the
way in which we navigate an action representation tree when performing an action. The
changing intentional focus of our agentive experience may also have an effect on our moment-to-
moment sense of agency by modulating the weights attached to the various available agency
cues. For instance, if, as I suggested, we zoom in and out as we perform an action, being
centrally aware at some points of very specific aspects of action execution and at other points of
more abstract aspects of our action plan, we may expect lower-level cues to be given more
weight when zooming in and higher-level cues to be given more weight when zooming out.
Page 14 of 19
Time to Act
Conclusion
In a nutshell, I argued here that when attempting to characterize the core components of
agentive experiences we need to pay closer attention to their contents and temporal dynamics. I
tried to emphasize two dimensions along which the contents of an agentive experience can vary
as the corresponding action unfolds in time. The first has to do with the level of grain or scale at
which the ongoing action is consciously represented. Many representations at various levels of
the action representation hierarchy could in principle contribute to the intentional contents of
agentive experience. It is important that we gain a better understanding of what determines
which of these action representations will contribute to the contents of agentive experience at a
given moment and how these contents may shift as the action unfolds. I considered several
factors that could in principle influence what enters the contents of agentive awareness and
suggested some ways in which these ideas could be empirically tested. I proposed that the
changing contents of our agentive experiences should reflect the way in which we navigate an
action representation hierarchy when performing an action, zooming in when performing a
particular component and zooming out at transition points between components. The second
dimension of variation in the content of an agentive experience concerns its tense, that is,
whether its content is forward looking (I am about to jump), backward looking (I just
jumped), or present centered (I am jumping). Whether these two dimensions of variation are
correlated or orthogonal is yet another issue awaiting empirical scrutiny. My personal hunch is
that there is some degree of correlation. If we take a simple action, we should probably expect
that, on the scale dimension, we will initially zoom (p.21) in and then zoom out and that, on the
tense dimension, experiential contents should initially be forward looking but later backward
looking. However, the picture is likely to become more complicated when we consider complex,
extended-action sequences.
Regarding the other core component of agentive experiences, the sense of agency stricto sensu,
we still lack a clear understanding of how the many agency cues that can in principle contribute
to the sense of agency interact. I suggested that in order to improve our understanding of how
these cues are integrated, we should take into account their temporal properties and their
effects on the evolution of the sense of agency in the course of the action. Finally, I pointed out
that sense of agency and intentional content are not two completely independent components of
agentive experience. Rather, they interact, and changes in one component may induce changes
in the other.
Acknowledgments
The preparation of this chapter was supported by ANR grants ANR-10-LABX-0087 IEC and
ANR-10-IDEX-0001-02 PSL*. Some of the ideas discussed here were presented at a Special
Session on Action Consciousness at the Fourth Online Consciousness Conference, February 17
March 2, 2012 (http://consciousnessonline.com/2012/02/17/special-session-on-action-
consciousness/). I wish to express my heartfelt thanks to Myrto Mylopoulos, the organizer of this
special session, to my two commentators, John Michael and Markus Schlosser, and to all the
participants in this session for their many insightful comments and suggestions.
Notes
References
Page 15 of 19
Time to Act
Aarts, H., Custers, R., & Marien, H. (2009). Priming and authorship ascription: when
nonconscious goals turn into conscious experiences of self-agency. Journal of Personality and
Social Psychology, 96, 967979.
Aarts, H., Custers, R., & Wegner, D. (2005). On the inference of personal authorship: enhancing
experienced agency by priming effect information. Consciousness and Cognition, 14(3), 439458.
Bayne, T. (2010). Agentive experiences as Pushmi-Pullyu representations. In J. Aguilar, A.

Buckareff and K. Frankish (Eds.), New waves in the philosophy of action (pp. 219236). New
York: Palgrave Macmillan.
Bayne, T., & Pacherie, E. (2007). Narrators and comparators: the architecture of agentive self-
awareness. Synthese, 159(3), 475491.
Belayachi, S., & van der Linden, M. (2009). Level of agency in sub-clinical checking.
Consciousness and Cognition, 18, 293299.
Berti, A., Spinazzola, L., Pia, L., Rabuffetti, M., Haggard, P., & Rossetti, Y. (2008). Motor
awareness and motor intention in anosognosia for hemiplegia. In P. Haggard, Y. Rossetti, & M.
Kawato (Eds.), Sensorimotor Foundations of Higher Cognition: Attention and Performance XXII
(pp. 163181). Oxford: Oxford University Press.
Blakemore, S. J., Wolpert, D., & Frith, C. (2000). Why cant you tickle yourself? Neuroreport, 11,
R11R16.
Chambon, V., Filevich, E., & Haggard, P. (2014). What is the human sense of agency, and is it
metacognitive? In S. M. Fleming & C. D. Frith (Eds.), The cognitive neuroscience of
metacognition (pp. 321342). Berlin: Springer.
Chambon, V., & Haggard, P. (2012). Sense of control depends on fluency of action selection, not
motor performance, Cognition, 125, 441451.
Cohen, M. A., & Dennett, D. C. (2011). Consciousness cannot be separated from function. Trends
in Cognitive Science, 15, 358364.
Desantis, A., Roussel, C., & Waszak, F. (2011). On the influence of causal beliefs on the feeling
of agency. Consciousness and Cognition, 20(4), 12111220.
Desantis, A., Weiss, C., Schutz-Bosbach, S., & Waszak, F. (2012). Believing and perceiving:
authorship belief modulates sensory attenuation. PLoS ONE, 7(5), e37959.
the experience of agency: a positron emission tomography study. Neuroimage, 18, 324333.
normal subjects. Neuropsychologia, 36(11), 11331140.
Page 16 of 19
Time to Act
Frith, C., Blakemore, S., & Wolpert, D. (2000). Abnormalities in the awareness and control of
action. Philosophical Transactions of the Royal Society B: Biological Sciences, 355, 1404, 1771
1788.
Gallagher, S. (2007). The natural philosophy of agency. Philosophy Compass, 2(2), 347357.
Haggard, P., & Eimer, M. (1999). On the relation between brain potentials and the awareness of
voluntary movements. Experimental Brain Research, 126, 128133.
Haggard, P., Clark, S., & Kalogeras, J. (2002). Voluntary action and conscious awareness.
Nature Neuroscience, 5(4), 382385.
Haggard, P., & Tsakiris, M. (2009). The experience of agency: feeling, judgment and
responsibility. Current Directions in Psychological Science, 18(4), 242246.
Hughes, G., Desantis, A., & Waszak, F. (2013). Mechanisms of intentional binding and sensory
attenuation: the role of temporal prediction, temporal control, identity prediction, and motor
prediction. Psychological Bulletin, 139(1), 133.
Jeannerod, M. (1995). Mental imagery in the motor context. Neuropsychologia, 33(11), 1419
1432.
Jeannerod, M. (1997). The cognitive neuroscience of action. Oxford: Blackwell Publishers.
Knoblich, G., & Kircher, T. T. J. (2004). Deceiving oneself about being in control: conscious
detection of changes in visuomotor coupling. Journal of Experimental Psychology-Human
Perception and Performance, 30(4), 657666.
Knoblich, G., & Repp, B. H. (2009). Inferring agency from sound. Cognition, 111(2), 248262.
Koch, C., and Tsuchiya, N. (2007). Attention and consciousness: two distinct brain processes.
Trends in Cognitive Science, 11, 1622.
Kouider, S., de Gardelle, V. Sackur, J. & Dupoux, E. (2010). How rich is consciousness? The
partial awareness hypothesis. Trends in Cognitive Science, 14, 301307.
Lamme, V. A. (2003). Why visual attention and awareness are different. Trends in Cognitive
Science, 7, 1218.
Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary
action. Behavioral and Brain Sciences, 8, 529566.
Libet, B., Gleason, C. A., Wright, E. W., & Pearl, D. K. (1983). Time of conscious intention to act
in relation to onset of cerebral activities (readiness potential): the unconscious initiation of a
freely voluntary act. Brain, 106, 623642.
Metcalfe, J., & Greene, M. (2007). Metacognition of agency. Journal of Experimental Psychology:
General, 136(2), 184199.
Page 17 of 19
Time to Act
Moore, J. W., & Fletcher, P. C. (2012). Sense of agency in health and disease: a review of cue
integration approaches. Consciousness and Cognition, 21(1), 5968.
Moore, J. W., & Haggard, P. (2008). Awareness of action: inference and prediction.
Moore, J. W., Wegner, D. M., Haggard, P. (2009). Modulating the sense of agency with external
cues. Consciousness and Cognition, 18, 10561064.
Naccache, L., Dehaene, S., Cohen, L., Habert, M.-O., Guichart-Gomez, E., Galanaud, D. & Willer,
J.-C. (2005). Effortless control: executive attention and conscious feeling of mental effort are
dissociable. Neuropsychologia, 43, 13181328.
ORegan, J. K., and Noe, A. (2001). A sensorimotor account of vision and visual consciousness.
Behavioral and Brain Sciences, 24, 939973.
Pacherie, E. (2008). The phenomenology of action: a conceptual framework. Cognition, 107(1),

179217.
Pacherie, E. (2010). Self-agency. In S. Gallagher (Ed.), The Oxford handbook of the self (pp. 440
462). Oxford: Oxford University Press.
Posner, M. I. (1994). Attention: the mechanisms of consciousness. Proceedings of the National

Academy of Science U.S.A., 91, 73987403.
Sato, A. (2009). Both motor prediction and conceptual congruency between preview and action-
effect contribute to explicit judgment of agency. Cognition, 110(1), 7483.
Sato, A., & Yasuda, A. (2005). Illusion of sense of self-agency: discrepancy between the
predicted and actual sensory consequences of actions modulates the sense of self-agency, but
not the sense of self-ownership. Cognition, 94(3), 241255.
Sidarus, N., Chambon, V., & Haggard, P. (2013). Priming of actions increases sense of control
over unexpected outcomes. Consciousness and Cognition, 22(4), 14031411.
two-step account of agency. Consciousness and Cognition, 17(1), 219239.
Tsakiris, E., & Haggard, P. (2003). Awareness of somatic events associated with a voluntary
action. Experimental Brain Research, 149(4), 439446.
Vallacher, R. R., & Wegner, D. M. (1985). A theory of action identification. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency: individual variation in
action identification. Journal of Personality and Social Psychology, 57, 660671.
Page 18 of 19
Time to Act
van der Weiden, A, Aarts, H., & Ruys, K. I. (2010). Reflecting on the action or its outcome:
behavior representation level modulates outcome-priming effects on self-agency experiences.
Wegner, D. (2002). The illusion of conscious will. Cambridge, MA: MIT Press.
Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation: sources of the experience of
will. American Psychologist, 54(7), 480492.
Wenke, D., Fleming, S. M., & Haggard, P. (2010). Subliminal priming of actions influences sense
of control over effects of action. Cognition, 115(1), 2638.
Wyart, V., & Tallon-Baudry, C. (2008). Neural dissociation between visual awareness and spatial
attention. The Journal of Neuroscience, 28(10), 26672679.
Notes:
(1) Note that my concern is with agentive experiences for physical actions. Whether and how the
account proposed here could be extended to cover mental actions as well are difficult issues on
which I shall not pronounce myself here.
(2) Bayne (2010) leaves it open whether the agentive core should take the richer form I suggest
or can reduce to a bare experience of oneself as acting.
(3) Thanks to Markus Schlosser for raising this objection.
(4) See also Kouider et al. (2010) and Cohen and Dennett (2011) for discussions of the
methodological difficulties involved in attempting to empirically confirm or falsify the claim that
there can be consciousness without attention.
(5) See, for instance, van der Weiden et al. (2010) for a study of how the effects of priming on
agentive experiences crucially depend on action identification level, and Belayachi and van der
Linden (2009) for a discussion of the role of low-level action identification in checking behavior
in subjects with obsessive-compulsive disorder.
Page 19 of 19
Deconstructing Voluntary Action

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Unconscious and Conscious Component Processes
Lara Krisst
Carlos Montemayor
Ezequiel Morsella
DOI:10.1093/acprof:oso/9780190267278.003.0002

Investigations of human action have led to the conclusion that much action production can occur
unconsciously. Behaviors such as reflexes (e.g., the pupillary reflex, peristalsis) and automatisms
in neurological conditions reveal that action production can be mediated unconsciously. Less
obvious, however, are the unconscious mechanisms associated with everyday voluntary actions.
Voluntary action is a complex form of action that involves both unconscious and conscious
component processes. This chapter reviews the unconscious components of voluntary action and
then examines how these components interact with consciousness. The analysis includes
treatments of Skinners influential three-term contingency and the phenomenon of effortless
control. The chapter discusses the unresolved issues (e.g. the necessity of a go signal during
action selection) and paradoxes (e.g. the self-report paradox) that are encountered when
investigating action from a reductionistic standpoint in which there is no homunculus-like agent
involved in voluntary action.
Keywords: unconscious processing, consciousness, voluntary action, Skinner, automatism
What in everyday life is referred to as voluntary actiondeciding to make some tea, flick a
light switch, or telephone a friendis a multifaceted event in the nervous system, one involving
a complex interplay of unconscious and conscious1 component processes. To begin to unravel
Page 1 of 37
the nature of this complex interplay, it is helpful to first examine the nature of relatively more
simple actions, such as those that are mediated unconsciously.
Unconscious Components of Voluntary Action

Unconsciously Mediated Actions
An analysis of unconsciously mediated actions, in which processing is unconscious throughout
the entire stimulus-response arc (i.e., from stimulus input processing to overt behavior), reveals
which kinds of component processes can transpire unconsciously and, in voluntary action, occur
unconsciously.
Automatisms occurring during a state of unconsciousness are one kind of unconsciously

mediated action. During these states, actions can be performed when the actor appears to be in
a state of unconsciousness (or, at the least, in a state in which he or she is unaware of the
environment and ongoing action). These automatisms are usually caused by neurological
conditions (e.g., epilepsy). (p.26) Case studies in the literature reveal a wide variety of
automatisms, including humming (Bartolomei et al., 2002), spitting (Carmant et al., 1994),
whistling (Raghavendra, Mirsattari, & McLachlan, 2010), oroalimentary automatisms (Maestro
et al., 2008), laughing (Enatsu, Hantus, Gonzalez-Martinez, & So, 2011), repetitive kissing
(Mikati, Comair, & Shamseddine, 2005), singing (Doherty et al., 2002; Enatsu et al., 2011),
sexual behaviors (Spencer, Spencer, Williamson, & Mattson, 1983), written and spoken
(nonsense) utterances (Blanken, Wallesch, & Papagno, 1990; Kececi, Degirmenci, & Gumus,
2013), the removal of clothing (Wortzel, Strom, Anderson, Maa, & Spitz, 2011), and rolling,
pedaling, and jumping (Kaido et al., 2006). Convergent evidence that actions of such complexity
can transpire unconsciously is found in the literatures on narcolepsy (Zorick, Salis, Roth, &
Kramer, 1979), somnambulism (Plazzi, Vetrugno, Provini, & Montagna, 2005; Schenk &
Mahowald, 1995), and forms of coma and persistent vegetative states (Klein, 1984; Laureys,
2005; Pilon & Sullivan, 1996), in which actions such as automatic ocular pursuit and some
reflexes (e.g., the pupillary reflex) are elicitable. It has been observed that licking, chewing,
swallowing, and other behaviors can occur unconsciously once the incentive stimuli activate the
appropriate receptors (Bindra, 1974; Kern, Jaradeh, Arndorfer, & Shaker, 2001).
In other neurological conditions, the patient can retain consciousness of a general kind (e.g.,
awareness of the environment), but expressed actions are systematically decoupled from
consciousness. This occurs in neurological conditions such as blindsight (Weiskrantz, 1992,
1997), anarchic hand syndrome (Marchetti & Della Sala, 1998), alien hand syndrome (Bryon &
Jedynak, 1972; Chan & Ross, 1997), ambient echolalia (Suzuki et al., 2012), visual form agnosia
(Goodale & Milner, 2004; Milner & Goodale, 1995), and utilization behavior syndrome
(Lhermitte, 1983). In utilization behavior syndrome, for example, patients cannot suppress
stimulus-elicited action. For example, a patient, when confronted with a pair of eyeglasses, will
put the eyeglasses on even if she is already wearing another pair of eyeglasses. Similarly, these
patients cannot, for instance, suppress the act of drinking a glass of water when the glass is
presented to them. During and after such actions, the patients report that the action was against
their will. Similar unintentional behavior is found in ambient echolalia, in which the patient
unintentionally repeats what someone else uttered. The unconsciously mediated actions
displayed in these conditions are quite sophisticated, including the manipulation of objects (e.g.,
in anarchic hand syndrome; Yamadori, 1997) and the walking around of obstacles (e.g., in
Page 2 of 37
blindsight), all while the patient reports that the actions are not mediated consciously (see also
Berti, Garbarini, and Pia, Chapter 17 of this volume; Rowe and Wolpe, Chapter 18 of this
volume). For these kinds of action, not only is motor control mediated unconsciously, but low
levels of perceptuosemantic (p.27) processing (e.g., feature analysis and stimulus
identification) must transpire unconsciously (Zeki & Bartels, 1999). Unconsciously mediated
actions are evident also in people who are neurologically intact, as in the case of reflexive action
(e.g., the pupillary reflex, reflexive blinking, and the pain withdrawal reflex; Bindra, 1974).
One can also observe the unintended activation of action plans in action slips (Botvinick &
Bylsma, 2005; Heckhausen & Beckmann, 1990) and in laboratory response interference
paradigms, in which action-related distractor stimuli interfere with intended actions. Such
interference stems from the activation of distractor-related action plans that conflict with the
intended action toward a target. In the classic flanker task (Eriksen & Eriksen, 1974; Eriksen
& Schultz, 1979), for example, participants are first trained to press one button with one finger
when presented with the letter S or M and to press another button with another finger when
presented with the letters P or H. After training, participants are instructed to respond to
targets that are flanked by distractors. For example, they are instructed to respond to the letter
presented in the center of an array (e.g., SSPSS) and to disregard the flanking distractor letters.
In terms of response latencies and error rates, the least interference is found when the
distractors are identical to the target (e.g., SSSSS, target underscored). Interference is stronger
when the distractors are associated with responses that are different from those associated with
targets (e.g., SSPSS; response interference) than when distractors are different in appearance
(e.g., SSMSS) but are associated with the same response (perceptual interference; Eriksen &
Schultz, 1979). Behavioral and psychophysiological evidence reveals that, during response
interference, competition involves simultaneous activation of the brain processes associated
with the target- and distractor-related responses (Coles, Gratton, Bashore, Eriksen, & Donchin,
1985; DeSoto, Fabiani, Geary, & Gratton, 2001; Eriksen & Schultz, 1979; Georgopoulos,
Caminiti, Kalaska, & Massey, 1983; Goldrick, & Blumstein, 2006; Mattler, 2005; McClelland,
1979; van Veen, Cohen, Botvinick, Stenger, & Carter, 2001; Starreveld, Theeuwes, & Mortier,
2004). (See review of additional evidence for the unintentional activation of action plans by
distractors and other kinds of incidental stimuli in Ellis, 2009.)
That activation from the unintentional processing of visual distractors directly influences
actional processes has been explained by the notions of continuous flow (Eriksen & Schultz,
1979) and cascade processing (McClelland, 1979; Morsella & Miozzo, 2002; Navarrete & Costa,
2004). Both continuous flow and cascade processing are based on the idea that, in the nervous
system, activation cannot help but flow from one stage of processing (e.g., perceptual) to the
subsequent stage (e.g., action-related; Ganz, 1975). This linking of perceptual processing to
action/motor processing has been referred to as (p.28) efference binding (Haggard,
Aschersleben, Gehrke, & Prinz, 2002). This kind of stimulus-response (S R) binding allows one
to press a button when presented with a cue (e.g., the letter S in the flanker task). Research
has shown that responding on the basis of efference binding can occur unconsciously, as in
laboratory paradigms employing the presentation of subliminal, masked stimuli (Fehrer &
Biederman, 1962; Fehrer & Raab, 1962; Hallett, 2007; Taylor & McCloskey, 1990, 1996). (See
review of evidence of unconscious efference binding in Hallett, 2007.) For example, Taylor and
Page 3 of 37
McCloskey (1990, 1996) demonstrated that, in a choicereaction time task, participants could
select the correct motor response (one of two button presses) when confronted with subliminal
stimuli (e.g., a subliminal S in the flanker task). The investigators concluded that appropriate
programs for two separate movements can be simultaneously held ready for use, and that either
one can be executed when triggered by specific stimuli without subjective awareness (Taylor &
McCloskey, 1996, p. 62).
We now turn to an everyday phenomenon that involves unconscious, action-related processes:

motor control.
Unconscious Processes in Motor Control

Motor control, which requires that the correct pattern of efference be transmitted to the
appropriate muscle fibers at the appropriate time, too, is held to transpire unconsciously. Motor
control is unconscious, not only in reflexes, automatisms, and other unconsciously mediated
actions, but in voluntary actions, or actions that are usually mediated consciously. There is a
consensus that one is unconscious of these efferences, which control (in highly complex ways)
the activities of the muscle fibers during action production and control (Decety & Grzes, 1999;
Fecteau, Chua, Franks, & Enns, 2001; Fourneret & Jeannerod, 1998; Grossberg, 1999; Heath,
Neely, Yakimishyn, & Binsted, 2008; Jeannerod, 2003; Liu, Chua, & Enns, 2008; Rosenbaum,
2002; Rossetti, 2001).
Thus, motor control and procedural learning can be implicit (i.e., unconscious and not self-
reportable; Squire, 1987; J. A. Taylor & Ivry, 2013). It has been theorized that this well-
established observation results from the fact that motor programs are computed on-line but
then must be scrapped, because these programs are highly context-sensitive and cannot be
stored for later use (Grossberg, 1999). For example, the motor programs used to grasp a cup on
the left would not be effective at grasping a cup on the right (Rosenbaum, 2002). Similarly, the
motor programs used by a child to grasp a cup would not be effective if used for the body of
adult, because each motor program is tailored to only one body size (Grossberg, 1999). Thus, it
has been proposed that, unlike semantic knowledge about the world, which can be accumulated
(p.29) over the course of years, there should be no long-term memories formed for motor
programs (Grossberg, 1999). Thus, the knowledge acquired for successful motor control is
fundamentally different from the kind of knowledge that people reflect on when deciding what
to do (Goodale & Milner, 2004).
Although motor control is unconscious, it is sophisticated, anticipatory, flexible, and highly

context sensitive. For example, motor control can realize an action goal in various ways, a
feature referred to as motor equivalence (Lashley, 1942). Because of motor equivalence, one
may flick a switch with ones index finger, palm, or even ones elbow. Research has revealed
that the motor programs used to realize these action goals are quite sophisticated and are
influenced not only by current action goals, but also by subsequent action goals (see end-state
comfort effect; Zhang & Rosenbaum, 2008). For example, in the case of co-articulation, the
way one articulates a speech sound depends in part on the nature of the articulation of
subsequent speech sounds (Levelt, 1989; Rosenbaum, 2002).
Page 4 of 37
In summary, examination of unconsciously mediated actions reveals that processes as

sophisticated as (a) the activation of action plans, (b) action selection, (c) certain
perceptuosemantic mechanisms (e.g., feature analysis and stimulus detection), and (d) motor
control can transpire unconsciously. Some of these processes (e.g., motor control) occur
unconsciously even during voluntary action, the next topic of discussion.
Conscious Components of Voluntary Action

Voluntary Action Selection
The foregoing discussion reveals that motor control occurs unconsciously, and that, under
certain circumstances, even the activation and selection of action plans can transpire
unconsciously. In the phenomenon of voluntary action, however, aspects of action selection are
often experienced consciously. (It is well established that action selection, as when one presses
one button instead of another, is distinct from motor control; Proctor & Vu, 2010.) For example,
in voluntary action, action selection involves the selection of what is best characterized as an
action effecta goal to be realized in the world (e.g., flicking a switch, bringing a cup toward
ones mouth) or in the body (e.g., whistling; Hommel & Elsner, 2009). For this form of action,
one is conscious of the representations of action effects. These representations are perceptual-
like (Godwin, Gazzaley, & Morsella, 2013; Gray, 2004). (As discussed earlier, the motor
programs used to implement these goals are unconscious; as discussed later in this chapter,
some of the executive processes involved in the selection process, too, may be unconscious.) As
representations in the nervous (p.30) system, the conscious representations of action effects
occupy an intermediate level of processing in nervous function (Jackendoff, 1990; J. Prinz,
2007). These representations are not associated with early, low stages of sensorial processing
(Zeki & Bartels, 1999), or with processes associated with the later stage of motor control.
Rather, action effect representations belong to a stage of processing in between the analysis of
sensory afference and the calculation of efference to the motor centers. Neither purely sensory
nor purely motor (Sperry, 1952), conscious representations have been construed as a kind of
translation function that, in voluntary action, links perceptual processes to actional processes
(Morsella, 2005; W. Prinz, 2003).
To understand the nuts and bolts of action selection in voluntary action, theorists have appealed
to ideomotor theory, which was introduced to the literature in the nineteenth century (Harless,
1861; Lotze, 1852) and popularized in America by William James (1890). (See contemporary
interpretations of ideomotor theory in Hommel, 2009, Chapter 14 of this volume; Hommel,
Msseler, Aschersleben, & Prinz, 2001). According to ideomotor theory, the activation of the
representation of an action effect leads directly to the (unconscious) motor programs that
realize that action effect. According to James (1890), these representations are consciously
experienced. From the Jamesian standpoint, during action execution, one is unconscious of
motor control, but is conscious of the perceptual representations arising from the action (Gray,
2004). Thus, these representations can be based on afference from the world or from the body,
such as visual afference and proprioception, respectively. After one exhibits an action
voluntarily or involuntarily, a memory of these afference-based representations is formed.
According to ideomotor theory, these memories can then be activated intentionally (e.g., by top-
down processing) to guide future voluntary actions (Hommel & Elsner, 2009; James, 1890).
From this standpoint, all voluntary actions require memory processing, for the representation of
Page 5 of 37
any to-be-produced action effect is based to some extent on prior experience. From this
standpoint (Hommel & Elsner, 2009), initial actions stem from innate action plans or from action
effects that were learned from accidentally performing a novel action. Novel actions can also
arise from simulations based on memories of previous actions (Schacter & Addis, 2007).
Importantly, in ideomotor theory, the activation of the action effect automatically leads to the
realization of that action effect in the world or in the body. For example, the image of flexing
ones finger automatically leads to the flexing of the finger. According to James, this occurs
unless one simultaneously sustains an incompatible idea (e.g., that of not flexing the finger;
James, 1890). Thus, in this framework, there is no homuncular-like, decider operating above
these representations, deciding which representation should influence overt behavior. Instead,
what curbs the influence of one representation on motor behavior is (p.31) only the influence of
another, incompatible representationone that happens to be activated.
Consistent with this view of action conflicts, Curtis and DEsposito (2009) propose that, at one
moment, a conflict may involve representations A and B (associated with neural correlates ANC
and BNC), and then, at a later time, a conflict may involve representations C and D (associated
with neural correlates CNC and DNC). Importantly, the two conflicts involve separate cognitive
and neural processes, which suggests that no single area of the brain is specialized for
inhibiting all unwanted actions (Curtis & DEsposito, 2009, p. 72). This conclusion is consistent
with the non-homuncular, ideomotor theory, in which there is (a) no omnipresent, self-like
decider that adjudicates between possible action plans, and (b) no general braking system,
operated by a decider, that stops all unwanted actions. Instead, action suppression is the
outcome of interactions between two action-related representations (Curtis & DEsposito, 2009).
The nature of the outcome depends, not on general algorithms associated with conscious
processing, but rather on the properties of the action systems involved, including the relative
strengths of the actional dispositions (Campbell & Misanin, 1969; Gold & Shadlen, 2007;
Morsella, 2005; Skinner, 1953).2 Thus, during conscious action selection, one action goal is
selected over another, but not by some (or the same) homunculus. Instead, mental
representations, including those of action sets (Grafman & Krueger, 2009) and of rules (Miller,
2000), compete for the control of action by biasing the activations of representations of action
effects. Often, the competition involves top-down control.
Such competition between action-related representations is evident in the classic Stroop task
(1935). In the Stroop task, experimental subjects are instructed to name the color in which
words are written. Conflict between action plans arises when an incongruent color is presented
on a color name (e.g., red written in blue font). In such a circumstance (the incongruent
condition), the strong and automatic word-reading plan conflicts with the relatively weaker
color-naming plan (Cohen, Dunbar, & McClelland, 1990), leading to response interference
(increased error rates and response latencies). Each of these plans conflicts with the other plan
regarding how overt behavior should be guided.
We will now interpret Stroop conflict from the perspective of ideomotor theory. The theory
proposes that voluntary action control occurs through the intentional activation of perceptual-
like representations. Consistent with this view, during Stroop conflict, top-down activation from
the frontal cortex activates areas in the posterior brain (e.g., visual association cortex) that are
Page 6 of 37
associated with task-relevant dimensions (e.g., color; Enger & Hirsch, 2005; Gazzaley, Cooney,
Rissman, & DEsposito, 2005). Importantly, this intentional (p.32) aspect of Stroop
performance is associated with activation in perceptual regions (Enger & Hirsch, 2005). This is
consistent with the more general observation that, to influence behavior, top-down activation
from action sets held in working memory or long-term memory increases or decreases the
strength of perceptuosemantic information (Gazzaley et al., 2005), along with, most likely, other
kinds of information (e.g., motor priming). (See additional evidence for the role of perceptual
regions and activations in action control in Desmurget et al., 2009; Desmurget & Sirigu, 2010;
Iacoboni & Dapretto, 2006; Miall, 2003.)
In an incongruent trial, the intentional top-down control is challenged by the automatic

activation of the word-reading plan (Cohen et al., 1990), which not only activates incompatible
response codes but, because of the activation of incompatible skeletomotor inclinations, creates
strong urges to read the word aloud (Morsella, Gray, Krieger, & Bargh, 2009a). During a
successful incongruent Stroop trial, the strength of the intentional action plan is stronger than
that of the competing, automatic plan, resulting in the utterance of the name of the color in
which the word is presented. However, this is not to say that the competing action plan, because
of cascade and continuous flow mechanisms, failed to influence overt behavior. Response times
and other measures (including urges to read the word aloud) reveal the subtle influence of the
competing but unselected word-reading plan. Such observations about the multidetermination of
overt behavior led Skinner (1953) to conclude that conflicted behavior is different from non-
conflicted behavior: The former is of weaker strength and more perturbable than the latter.
On the basis of Skinners insight and notions such as continuous flow and cascade processing,
one can conclude that, during action selection, it is not the case that the action plan of an
intended action is the only action plan that influences behavior. Action selection is not a winner
take all process. Instead, distractor stimuli and their associated action plans influence actional
processes all the way down to the muscle fiber (Coles, et al., 1985; DeSoto et al., 2001;
Georgopoulos et al., 1983; Goldrick & Blumstein, 2006). Thus, it is inaccurate to characterize
conflicts in action control as having a clean resolution, leading to adaptive behavior. Instead, it
is more conservative to state that, during conflict, one action plan happens to have a stronger
and more notable influence over behavior than some other action plan.
The Generation of Action Options in Consciousness

The representations that can compete for the control of voluntary action can be triggered by
external stimuli, as in the case of the Stroop task (in which orthographic stimuli trigger word-
reading plans), or by representations (p.33) dependent entirely on memory, as when one
delivers a rehearsed speech. During action selection, these representations of action effects can
be thought of as action options (Angell, 1907; Gibson, 1979), which are the tokens of the
selection process in voluntary action. Hence, it is not surprising that, for every voluntary action,
the actor can self-report about conscious content responsible for that action (Poehlman, Jantz, &
Morsella, 2012). The implications of this often overlooked but reliable observation (that
voluntary actions are connected to conscious content) are important, even if self-reports on the
causes of these actions by actors are often inaccurate (Nisbett & Wilson, 1977; Wegner, 2002).
Page 7 of 37
Although representations of action options are experienced consciously, they often enter
consciousness in a manner that resembles, not voluntary processing, but reflexes: the entry of
these high-level representations is often unintentional and involuntary. (It has been argued a
priori that the conscious contents that influence action selection [e.g., action-related urges] are
seldom generated by intentional processes [see Vierkant, 2013].) Helmholtz (1856)
referred to the sophisticated unconscious and unintentional mechanisms giving rise to such
conscious contents as unconscious inferences. When speaking about such inferences, Helmholtz
referred to automatic word reading, a higher-level process that, though effortless and
unintentional, depends on years of intensive instruction. Helmholtz realized that, when
presented with an orthographic stimulus, one cannot help but read the word subvocally, which is
a complicated process that takes a visual input and engenders a representation (the
phonological representation) that is associated, not with the stimulus input modality (vision), but
with audition. Such unconscious inference can give rise to urges, which then function as options
for voluntary action selection. For example, automatic word-reading process is at the heart of
interference caused by the Stroop incongruent condition (Cohen et al., 1990). An urge may not
be selected to guide behavior, but, as discussed above, its activation influences behavior
nonetheless, albeit in a subtle fashion.
Action-Related Urges as a Form of Action Option

The foregoing discussion reveals that one difference between voluntary actions and
unconsciously mediated actions pertains to the nature of the selection process. In voluntary
action, the tokens of action selection are conscious representations of action effects (Hommel &
Elsner, 2009; Morsella, 2005). There are many kinds of (consciously represented) information
that can influence the selection process in voluntary action (see Morsella & Bargh, 2010). For
example, knowledge of the present physical environment is critical for adaptive action selection
(Ellis, 2009; Gibson, 1979; Merker, 2013). However, at this (p.34) stage of understanding, we
will focus only on one of the most basic forms of action optionsaction-related urges.
Action-related urges enter consciousness when one holds ones breath while underwater,
suppresses dropping a hot dish, or, in the Stroop incongruent condition, suppresses the
inclination to read the word aloud while carrying out the color-naming action plan (Morsella et
al., 2009a; Morsella et al., 2009b). When feeling action-related urges, one experiences the urge
to act in a certain way. Introspections regarding the nature of the urge reveal that the urge is
very action-specific: for example, it is the urge to inhale and not to perform some other action
(Pacherie, 2008). It is important to note that these conscious representations are isomorphic in
some way to the actions associated with the representations. In automatic word reading, for
example, the conscious imagery of the phonological form is isomorphic in many ways to what
would be perceived if the associated action (speaking) were produced overtly (Morsella &
Bargh, 2010; Sperry, 1952). (Interestingly, the phonological form is that which one experiences
when one says the word aloud, subvocalizes the word [i.e., speaks in ones mind], or hears
another person utter the word.)3 Action-related urges are reliably observed also in interference
paradigms such as the flanker task (Morsella et al., 2009c), Stroop task (Morsella et al., 2009a),
and the anti-saccade task (Morsella, Zarolia, & Gazzaley, 2012). For example, in these tasks,
conditions eliciting response interference (e.g., when the targets and distractors are associated
with different responses) lead to the strongest trial-by-trial urges to err. (See quantitative
Page 8 of 37
review of urge-related data from various paradigms in Morsella, Berger, & Krieger, 2011.) Such
subjective effects arise also when participants sustain incompatible intentions (e.g., to point left
and right with the same finger) during a motionless state (Gray, Bargh, & Morsella, 2013;
Morsella et al., 2009a). Consistent with this finding, research reveals that these urges do not
depend on the actual execution of the action plans (Chambon, Wenke, Fleming, W. Prinz, &
Haggard, 2013; Morsella et al., 2009b). In the next section, we discuss how urges, which can be
construed as action options for selection in voluntary action, come to be associated with external
stimuli.
The Three-Term Contingency as One Source of Action-Related Urges

It seems that voluntary action selection operates upon conscious contents. Conscious percepts
of external stimuli form an important class of contents that constrain and guide the selection
process (Ellis, 2009; Gibson, 1979; Merker, 2013). Interestingly, this insight was arrived at, not
only by the ideomotor theorists, but by the behaviorists. To appreciate this insight, it is
important to first (p.35) review the notion of the three-term contingency (Skinner, 1953; see
discussion in Corr & Morsella, in press).
During the first half of the twentieth century, the behaviorists attempted to explain voluntary
action in mechanistic terms. From this standpoint, most voluntary actions could be understood
in terms of operant (or instrumental) conditioning. One influential model was that of the three-
term contingency (Skinner, 1953). In this model, a discriminative stimulus (SD) in the
environment (or in the body, as in the case of proprioception) leads to a response (R), which
then leads to an outcome (e.g., a favorable or unfavorable consequence). The three-term
contingency is thus SD R O. Skinner (1953) proposed that the majority of human behaviors
could be described in terms of this three-term contingency.
The SD could be a green traffic light, a door knob, or a light switch in the upward position. It
could also be based on proprioceptive stimuli, or, one could argue, the activation of a memory.
For a theorist such as William James, the SD is intimately related to consciousness. Thus, from
the perspective of ideomotor theory, the SD can be considered as the stimulus that triggers
representations about potential action effects (referred to above as action options). These
action options include action-related urges in consciousness. According to the behaviorists,
whether one option is selected or not depends upon the O term of the three-term contingency. If
the outcome term is positive, the connection between the SD and R is strengthened in the future,
making the response more likely; if it is negative, the link between the first two terms is
weakened, making the response less likely. (Some contemporary theorists claim that todays
accounts of incentive-based learning have added little to this account of motivated behavior;
Loewenstein, 1996.) During operant conditioning, the O term is often associated with
consciousness, as when one subjectively experiences a positive or negative outcome. After
learning, it is unclear whether or how the O term influences consciousness when influencing
selection. Regarding the R term, one can be aware of the response (through reafference from
proprioception), or one can be aware of the corollary discharges associated with potential action
production. (As mentioned above, one is unaware of the motor programming.) In each of the
instances in which consciousness is involved with an aspect of the three-term contingency,
something resembling the notion of afference, instead of efference, is involved. This is
Page 9 of 37
consistent with the view that consciousness, though essential, is passive in the sense that it is
not about doing but about sensing (Tallon-Baudry, 2012). In other words, consciousness is more
of a talker than a doer (Morsella & Bargh, 2010). The observation is also consistent with
accounts in which consciousness is associated primarily with perceptual-like processes (Gray,
2004; Mller, 1843).
(p.36) Why Consciousness Is Associated with the Action Options in Voluntary Action
At this point in the discussion, it is important to ask, why is consciousness associated with,
specifically, the action options in the selection process of voluntary action? To answer this
question, it is important to appreciate that these action options arise from a heterogeneous
group of systems in the nervous system, some of which are phylogenetically ancient (e.g., the
urge to inhale) and some of which are phylogenetically less ancient (e.g., the urge to utter a
word). These systems not only have different phylogenetic and neuroanatomical sources, but
they also have distinct operating principles (Morsella, 2005). As mentioned earlier, often these
action options are triggered into consciousness in a reflex-like manner. For instance, in the
Stroop task and when holding ones breath while underwater, the incompatible inclinations at
play during each conflict arise unintentionally and are qualitatively distinct from each other,
arising from very different kinds of processes.
In order for the action selection component of voluntary action to yield adaptive action (e.g.,
holding ones breath while underwater), the appropriate options must be included as tokens in
the selection process (Morsella, 2005; Merker, 2013). Thus, speaking figuratively, something
must bring these options together so that they can influence overt action collectively. It has
been proposed that, in humans, conscious states serve this essential, integrative role (Morsella,
2005; Morsella & Bargh, 2010), because conscious states can bring together (or bind)
information and processes that would otherwise be non-integrated (Baars, 1988, 1998; Boly et
al., 2011; Clark, 2002; Damasio, 1989; Dehaene & Naccache, 2001; Del Cul, Baillet, & Dehaene,
2007; Doesburg et al., 2009; Edelman & Tononi, 2000; Freeman, 1991; Koch, 2012; Kriegel,
2007; Llins & Ribary, 2001; Ortinski & Meador, 2004; Sergent & Dehaene, 2004; Srinivasan,
Russel, Edelman, & Tononi, 1999; Tallon-Baudry, 2012; Tononi, 2012; Tononi & Edelman, 1988;
Uhlhaas et al., 2009; Varela, Lachaux, Rodriguez, & Martinerie, 2001; Zeki & Bartels, 1999).
It is important to state the qualification that only certain kinds of integrations in the nervous
system require conscious integration (Morsella, 2005). For example, intra- and inter-sensory
integrations (afference binding; Morsella & Bargh, 2011), simple stimulus-response associations
(i.e., efference binding; Haggard et al., 2002), and integrations involving smooth muscle
(Morsella et al., 2009a) can occur unconsciously (Morsella & Bargh, 2011). As discussed earlier,
unconscious forms of efference binding include reflexive pain withdrawal, or, less commonly,
correct choice responses to subliminal, masked stimuli (Hallett, 2007; Taylor & McCloskey,
1990, 1996). It seems that the integration in the nervous system that requires consciousness is
associated with (p.37) voluntary action and the skeletal muscle output system, which is the only
effector system in the body that can be consciously controlled. (See Morsella, 2005, for a
treatise regarding the special relationship between consciousness and skeletal [voluntary]
muscle.)
Page 10 of 37
From this standpoint, consciousness permits a form of integration,4 one that is essential for
furnishing the action options that voluntary action requires. Consciousness allows for multiple
action options to influence behavior collectively (Morsella, 2005). For example, when
suppressing the urge to drop a hot dish of food, information about potential tissue damage and
about the value of the dish may influence selection. Hence, during such a voluntary action, a
person may feel the subjective tuggings and pullings of action conflict caused by conflicting
action options (Morsella et al., 2009a). Conflicted behavior is a dramatic form of integrated
action. Integrated action occurs when two (or more) action plans that could normally influence
behavior on their own (when existing at that level of activation) are simultaneously co-activated
and are trying to influence the same skeletal muscle effector (Morsella & Bargh, 2011). Thus,
integrated action occurs when one holds ones breath, refrains from dropping a hot dish,
suppresses the urge to scratch an itch, suppresses a pre-potent response in a laboratory
paradigm, or makes oneself breathe faster (Morsella, 2005; Morsella, Gray, Krieger, & Bargh,
2009a). (Importantly, the last example does not involve suppression.) Suppressing (or over-
expressing) a saccade, cough, blink, or some other reflexive behavior is also a case of integrated
action. Integrated action involves the activation of more neural processes than non-integrated
action (DeSoto et al., 2001; Ortinski & Meador, 2004). It is important to note that the level of
activation of the plans involved in integrated action is far beyond that of sub-threshold
activations, which would not influence the entry of action options into consciousness (Morsella
& Bargh, 2011). For example, in psycholinguistic research, there is substantial evidence that
naming dog primes the action plan for naming a member of the same category (e.g., horse;
Levelt, 1989). The level of activation that we are speaking of in our definition of integrated
action is far above this thresholdit is at the level of activation at which action plans would not
only influence overt action but could trigger action. In addition, this level of activation
introduces action-related representations (e.g., action options or urges) into consciousness. In
the case of urges or even in the case of imagery about words to be uttered, the source of the
options in consciousness could stem in part from corollary discharges from unconscious motor
centers to perceptual regions (which are associated with consciousness; Buchsbaum, 2013). (See
Bridgeman, 2007, for a treatment of the limited role of corollary discharges and efference copies
in perception.)
(p.38) Conversely, when actions are mediated unconsciously, the actions are non-integrated
and reflect a certain lack of integration. Non-integrated actions are not influenced by all the
information (e.g., the action options) by which they should be influenced (Morsella & Bargh,
2011). This is obvious when reflexively inhaling while underwater or in the case of neurological
conditions (e.g., automatisms in epilepsy and anarchic hand syndrome).
From this standpoint, the kind of integrated action required to successfully perform an
incongruent Stroop trial, in which two skeletomotor plans are competing to influence overt
action, requires the integrative properties of conscious states (Morsella et al., 2009a).
Otherwise, only one action plan (the stronger one) is reflected in overt behavior. It seems that
the activation of incompatible skeletomotor plans is the essence of conscious conflict.
Accordingly, experiments have revealed that incompatible skeletomotor intentions (e.g., to point
right and left, to inhale and not inhale) do produce strong, systematic intrusions into
consciousness (Gray, Bargh, & Morsella, 2013; Molapour, Berger, & Morsella, 2011; Morsella et
Page 11 of 37
al., 2009c), but no such changes accompany smooth muscle conflicts (Morsella et al., 2009a) or
conflicts occurring at perceptual stages of processing (e.g., intersensory processing; see
quantitative review of evidence from multiple paradigms in Morsella et al., 2011). Such conflict-
related conscious states have been shown to be essential for cognitive control during conflict.
For instance, Desender, van Opstal, & van den Bussche (2014) provide evidence that the
conscious experience of conflict is required for subsequent cognitive adaptation effects (e.g.,
better performance on a Stroop incongruent trial following an incongruent, but not a congruent,
trial). According to Desender et al. (2014) argue that adaptation effects to both supraliminal and
subliminal stimuli require, perhaps not consciousness of the triggering stimuli (adaptation
effects have been observed in response to subliminal stimuli; Desender, van Lierde, & van den
Bussche, 2013; Hommel, 2013; van Gaal & Lamme, 2012; van Gaal, Lamme, & Ridderinkhof,
2010), but consciousness of the conflict itself (Morsella, 2005). These conclusions are in line
with the more general view that one can be conscious of urges and action inclinations but not
necessarily of the sources of such inclinations (Morsella, 2005; Nisbett & Wilson, 1977).
Suppression as an Index of Voluntary Action and the Encapsulation of Action Options

In line with the view that voluntary action involves action options (which, in turn, require
conscious processing for the options to influence actional mechanisms collectively), it has been
proposed by Passingham that, unlike involuntary actions (e.g., reflexes, automatisms during
seizures, and unconscious (p.39) actions during pathological states), voluntary actions are
special in that they are suppressible, as when one refrains from inhaling while underwater or
from dropping a hot dish. Hence, suppressibility serves as a useful behavioral index of voluntary
action in humans and other species. We propose that, based on the foregoing conclusions,
suppression is an instance of the more general condition of interaction across systems, a form of
interaction that requires consciousness. More specifically, the difference between unconscious
action and conscious action is that the former is always a case of non-integrated action, and the
latter a case of integrated action.
During the conflicts in integrated action, the expression of undesired action plans can be
suppressed, whereas action-related inclinations (e.g., conscious urges) cannot be suppressed
(Bargh & Morsella, 2008). For instance, a person can suppress dropping a painfully hot but
expensive dish, but cannot suppress the subjective urges to drop the dish. Thus, inclinations, as
action options, can be behaviorally suppressed, but most often are not phenomenally suppressed
(Bargh & Morsella, 2008). This scenario leads one to Chomskys (1988) conclusion that, unlike
machines, humans can not only be compelled to act in certain ways, but they can be inclined to
act in certain ways. From this standpoint, in response to a stimulus, one may not only act in a
certain way but may be inclined to act in a certain way, a state of affairs unlike anything
instantiated in human artifacts (as far as we know). We believe that Chomskys insight reflects
that (a) consciousness is intimately associated with action options; (b) action-related urges (i.e.,
inclinations) are one major source of such options; and (c) for such options to be adaptive, they
must be encapsulated (Firestone & Scholl, 2014; Fodor, 1983), meaning that the introduction
and nature of an action option (e.g., the urge to inhale) should be independent of general beliefs
and the simultaneous activation of other action options (e.g., to not inhale; see discussion in
Morsella, 2005; Morsella & Bargh, 2010).
Page 12 of 37
In summary, the main difference between voluntary action and other forms of action (e.g.,
automatisms and action slips) seems to pertain to the nature of action selection. For voluntary
action, the selection process involves tokens that can be construed as action options. Of all the
components associated with voluntary action, action options are the most intimately associated
with consciousness. In contrast, the degree to which consciousness is associated with the
executive processes associated with the actual action selection (that which guides integrated,
overt behavior) and with conflict resolution (if such a resolution exists; Morsella, 2005) remains,
at present, less obvious (Crick, 1995; Suhler & Churchland, 2009; Tallon-Baudry, 2012). Because
these tokens stem from various sources but must influence the same final common
path (McFarland & Sibly, 1975), which is the skeletal muscle output system (Morsella,
2005), these tokens must be brought together to influence (p.40) collectively the selection
process that leads to adaptive overt action (Morsella & Bargh, 2010). In ways that remain
mysterious, consciousness furnishes the kind of integration that this process requires. Under
normal circumstances, the entire process leads to adaptive, integrated actionsa successful
Stroop task or holding ones breath while underwater.
We now turn to a higher-level conscious content associated with voluntary action: the sense of
agency, that is, the sense that one is causing a physical or mental act (Engbert, Wohlschlger, &
Haggard, 2008; Sato, 2009; Synofzik, Vosgerau, & Newen, 2008).
Non-Urge Conscious Contents Associated with Voluntary Action: The Sense of Agency
The sense of agency (or of authorship of ongoing action; Wegner, 2003) results from the
perception of the lawful correspondence between action intentions and action outcomes
(Haggard & Clark, 2003; Hommel, 2009; Wegner 2003). For example, if one has the intention of
flexing ones finger and then the finger happens to flex, one is likely to sense that one caused
the action. It has been proposed that this attribution is the outcome of a conceptual process
(Jeannerod, 2009; Synofzik et al., 2008b) that takes into account information from various
contextual factors (Moore, Wegner, & Haggard, 2009; Wegner & Wheatley, 1999), including
motor efference (Cole, 2007; Engbert et al., 2008; Sato, 2009; Tsakiris, Schtz-Bosbach, &
Gallagher, 2007), proprioception (Balslev, Cole, Miall, 2007; Knoblich & Repp, 2009), and the
perception of the real-world consequences of ones intentions (Synofzik, Vosgerau, & Lindner,
2009).5 Such a comparison process is a component of many comparator models of the sense of
agency (Gray, 1995; Morsella et al., 2011). Several comparator models illuminate how intention-
outcome mismatches are detected and influence various levels of agency. Different theorists link
the sense of agency and urges to different phases of the process (cf., Berti & Pia, 2006; David et
al., 2008; Haggard, 2005, 2008).
The sense of agency arises not only for physical actions but also for mental actions (Bortolotti &
Broome, 2009): if one intends to imagine a triangle and then experiences the relevant imagery
(e.g., of a triangle), then one may believe that one caused the imagery, even when the percept
may have been caused by an experimental trick, as in the Perky effect (Perky, 1910). (In the
Perky effect, subjects are fooled into believing that they are imagining an image that is actually
presented physically on a screen.) By manipulating contextual factors (e.g., the perceived
feedback following action executions), experiments have demonstrated that subjects can be
fooled into believing that they caused (p.41) actions that were caused by something else
Page 13 of 37
(Wegner, 2002). For example, in one experiment, participants controlled a manual computer-
drawing device behind a screen. The subject could not see his or her hand in motion. False
feedback about the action was presented on the computer display. Because of this deceptive
feedback, subjects were fooled into thinking that their hands intentionally moved in one
direction when they actually moved in a slightly different direction (Fourneret & Jeannerod,
1998). With similar techniques, subjects in another study were tricked into believing that they
could control the movements of stimuli on a computer screen through a phony brain-computer
interface (Lynn, Berger, Riddle, & Morsella, 2010). The opposite effectthe sense that I did not
intend thathas also been induced experimentally: when intentions and outcomes mismatch,
people are less likely to perceive actions as originating from the self (Wegner, 2002). Most
research has examined how agency is influenced by intention-outcome mismatches or illusory
intention-outcome matches.
Thus, many of the subcomponents of the conceptual process are likely to be shared by other
rational processes, such as those used for inferring physical cause-and-effect relationships
(Jeannerod, 2009; Morsella et al., 2011; Synofzik et al., 2008b). In contrast, lower-level
components of the sense of agency are associated with the actual intending itselfthe subjective
feeling of intending that accompanies the control of ongoing physical and mental action
(Pacherie, 2008). This feeling is closer to the phenomenology of agency than to the concept of
agency discussed earlier. This phenomenology of agency requires the components of an
inclination (or urge) and basic consciousness, components that were discussed previously. These
conscious contents are experienced when one holds ones breath or refrains from dropping a hot
dish. One could propose that, in simple cases, such subjective states can occur independent of
the aforementioned conceptual processes that are necessary to ascribe actions to the self, as in
I did it or it is I who am observing this (Crick & Koch, 2000; James, 1890; Jeannerod, 2009;
Merker, 2007; Synofzik et al., 2008b). From this standpoint, agency can be explained without
invoking the actions of a supervisory system (Angell, 1907; Norman & Shallice, 1980), central
executive (Baddeley, 1986), or other, homuncular-like agent in the brain whose presiding over
action is a necessary ingredient of agency. As in ideomotor theory, there is no decider selecting
one action-effect representation over another. In addition, from this standpoint, there is no
doer in the mind implementing what the organism does: although it is tempting to say that an
action is voluntary only when one intends to do it, there are strong a priori considerations
(the fallacy of ad infinitum) and empirically based considerations (e.g., Libet, 2004) that render
such a position unscientific.
(p.42) The Problem of the Homunculus in the Study of the Sense of Agency
Research on voluntary action has been plagued by the homunculus problem more than by any
other problem. Such a fallacy is obvious in descriptions in which action options are selected,
after some deliberation, by some intelligent internal agent. The fallacy is also obvious in
descriptions in which an action is deemed voluntary when the action is intended by such an
internal agent, as when one states, My actions are voluntary when I intend to do them. In
science, agentic or organismic functions must be explained at the sub-agent or sub-organismic
level of description. It answers nothing to invoke a little person inside the mind that can carry
out all the functions (e.g., sensing, deliberating, deciding, and intending) that one seeks to
Page 14 of 37
explain in a reductionistic, mechanistic manner that is not based on the workings of a sub-
organismic agent.
Theorists have been trying to explain voluntary processing without the need of such an agent, as
is clear in the titles, Whats at the Top in the Top-Down Control of Action? (Roepstorff & Frith,
2004), In Search of the Wild Homunculus (Logan, 2003), and Banishing the Homunculus (Hazy,
Frank, & OReilly, 2006). Complementing the homunculus fallacy and the conclusion that it is
theoretically unnecessary to propose that conscious thought must be the object of some internal
observer, James (1890) proposed that, when introspecting, one is unable to find any evidence
of there being such an observer: Through his minds eye, he encountered nothing but
sensations, inclinations, and other ideas, that is, only the objects of the observer with no
observer to be found, a view that is in line with the contemporary theorizing (Crick & Koch,
2000; Merker, 2013). That James sensed only sensations, inclinations, and other ideas is
consistent with the conclusions of Hume (1888), who proposed that the self is inferred but
cannot be introspected about directly. Thomas Reid ([1785] 1855, p. 119) criticized Humes view
of the self as nothing more than a bundle of sensations: What we call a body is only a bundle of
sensations; and what we call the mind is only a bundle of thoughts, passions, and emotions,
without any subject (p. 119).
Effortless Control
Such a lack of awareness of the doer is evident in cases of effortless control, a topic that has
recently generated interest. The intuitive account of cognitive effort is the following. The higher
the demands of a task, the greater the effort associated with executing the task (Kahneman,
1973). However, in flow experiences, one finds the opposite correlation: if fully engaged (as a
trained athlete in a competition), task-demands may increase while experienced effort may
decrease or remain constant as cognitive demands increase (p.43) (Csikszentmihalyi &
Csikszentmihalyi, 1988). Although the notion of effortless control for highly demanding tasks is
counterintuitive, the experience of flow is a familiar one: one forgets how difficult it is to
perform a highly skilled action if one is very well trained to do it. Actually, the experience of
executing a task that one is trained to perform with a high level of precision is one of enjoyment
and relaxation, as research on these experiences documents (Csikszentmihalyi &
Csikszentmihalyi, 1988). Interestingly, during demanding tasks, one may also become less
aware of the self (see neural evidence in Goldberg, Harel, & Malach, 2006).6 Cases of
effortless control present the theoretical possibility of a spectrum of voluntary action, with
agency without awareness (motor control) at one end of the spectrum and effortless conscious
control at the other end, with different forms of voluntary awareness for action (e.g., sustained
attention, monitoring, and introspection) in between.
In contrast, the kinds of bindings/integrations associated with consciousness, such as those

associated with holding ones breath, suppressing dropping a hot dish of food, or performing an
incongruent trial of the Stroop task, are very much associated with the aforementioned lower-
order form of the sense of agency. Such binding is accompanied by the sense of something
countering the will of the self (Morsella, Berger, & Krieger, 2011). The effect is contextual: in
one context, action plan A may be linked to agency; in another context, the plan may be
Page 15 of 37
perceived as countering the self, as in the case of suppressed visceral urges (i.e., the monkey
on ones back; Riddle & Morsella, 2009).7
When Invoking a Homunculus Is Not Fallacious

It should be noted that homuncular explanations of voluntary action are not fallacious if the
functions carried out by the homunculus do not duplicate the functions that one intends to
explain at the sub-organismic level (Dennett, 1991). Thus, if the workings of the homunculus are
less complex than, and not identical to, those of the system in which it operates, then one could
propose that, at some stage of processing, a subcomponent of the system may function in a
homuncular manner. Perhaps this is best illustrated through an example. Consider a music
producer who can play no instruments but has a good ear for music and always prefers a simple
song structure over a complex one. Whenever a musician performs musical pieces for this
producer, the producer indicates that the simplest piece is the best. The musicians respect the
judgment of the producer and use it to guide their songwriting. In this example, the producer
cannot play any of the music herself. The actions of the performers are very different from what
the producer is capable of doing. In this way, the actions of the producer do not duplicate the
actions of the performers.
(p.44) Therefore, it is not fallacious to propose that, during action selection (based on some
criterion), a sub-organismic agent (e.g., belonging to action systems) can select action options
(generated by unconscious inferences), if this agent does not also duplicate the workings of the
system as a whole. As in our music producer example, the selecting agent (e.g., in unconscious,
action-related systems in Morsella, 2005) may not be capable of creating the options that are the
tokens of the selection process. Moreover, the mechanisms generating the options seem to be
incapable of selecting by themselves the best course of action, because they lack direct access
to the effector system (Morsella & Bargh, 2010). In this way, one could hypothesize that
conscious action options (generated by unconscious inferences) are apprehended by intelligent
but unconscious, homunculus-like mechanisms of the action output system (Morsella & Bargh,
2010; Morsella, Hoover, & Bargh, 2013). In this arrangement, each homunculus-like mechanism
is concerned with the execution of a certain kind of action (e.g., blinking versus reaching) in a
manner such that no single mechanism duplicates the functionality of the system as whole. From
this standpoint, a homuncular description of action selection and voluntary action is not
fallacious.
In summary, there are lower- and higher-level forms of the sense of agency. Higher forms of this
conscious content (e.g., the sense of action authorship) stem from mechanisms that resemble
the conceptual processes used to infer the relationships between causes and effects, whereas
the lower forms of this sense arise during action conflicts (e.g., holding ones breath) when
something is sensed to be countering the will of the self. Everyday, homuncular descriptions of
voluntary action are fallacious (e.g., because of the fallacy of ad infinitum) and are also
inconsistent with evidence (e.g., research on effortless control). However, certain kinds of
homuncular descriptions (e.g., when each homunculus does not duplicate the functionality of the
entire system in which it is a part) are not only logically sound but may be helpful in
conceptually understanding the perception-to-action apparatus.
Outstanding Questions
Page 16 of 37
We now turn to topics that are even more uncharted, which we believe will be relevant to future
investigations on voluntary action. First, we bring to the readers attention the notion of a go
signal.
The Go Signal
As mentioned earlier, according to ideomotor theory, action selection is not driven by a
homunculus selecting one action goal representation over another. (p.45) Instead, according to
James (1890), when one imagines moving ones figure, ones finger would move, unless there is
also the simultaneous activation of an incompatible idea (e.g., that of the finger not moving).
From this standpoint, without the incompatible idea, the action would ensue automatically. This
view is consistent with many of todays models of decision-making, in which the selection of
outcomes is based primarily on the relative level of activation of each outcome (as in
accumulator models; Curtis & DEsposito, 2009; Gold & Shadlen, 2007). Again, though action
options are conscious, the processes mediating the outcome of such conflicts (should such
outcomes exist; Morsella, 2005) may themselves be unconscious (Crick, 1995; Suhler &
Churchland, 2009; Tallon-Baudry, 2012).
According to some theorists, more than just the selection of the action effect is required for
action production. There must also be a go signal, issued after the action option is selected
(see evidence in Bullock & Grossberg, 1988). Interestingly, the activation level of the go signal
for one action can carry over to a subsequent action (Bullock & Grossberg, 1988). The notion of
a go signal is interesting and worthy of further investigation. However, until more is known
about this signal, how it interacts with conscious/unconscious processes, and what it would add
to the action selection process that is not already present in a system in which selection is
driven by the relative weighting of action options, one can, for theory-building purposes, use a
simple Jamesian framework to capture much of how voluntary action involves both conscious
and unconscious processing. Many of the quasi-rational processes proposed to operate over the
conscious representations of action options (e.g., the go signal) may be unconscious (Vierkant,
2013) or difficult to introspect about, as suggested by research on imageless thought
(Woodworth, 1915).
The Self-Report Paradox

Another challenge in studying voluntary action is the self-report paradox (first discussed in
Morsella and Bargh, 2011). To appreciate this paradox, it is important to appreciate the nature
of the phenomenon of binocular rivalry in non-human animals. In binocular rivalry, one visual
stimulus (e.g., the picture of a face) is presented to one eye of the experimental subject, and
another visual stimulus (e.g., the picture of a house) is presented to the other eye (Alais & Blake,
2005). Although one would predict that the subject should experience a composite image (e.g., a
face overlapping a house), in this curious phenomenon, the subject experiences only one object
at a time, for reasons that remain unclear. Interestingly, both humans and monkeys show this
effect. For the latter, the monkeys are trained to press one button when seeing one stimulus (p.
46) and to press another button when perceiving the other stimulus (Leopold & Logothetis,
1996). These responses are considered forms of self-report.
The paradox is as follows. If humans can perform accurate choice-responses to stimuli of which
they are unconscious (e.g., as in subliminal, backward masking), perhaps the monkeys in these
Page 17 of 37
rivalry experiments, too, are unaware of the stimulus to which they respond motorically. When
humans button-press as a function of unconscious efference binding, it is regarded as
unconscious action. However, when the monkeys perform such stimulus-elicited actions, the
actions are regarded as a form of conscious self-report. Yet, both tasks are identical in that
each involves a perceptual discrimination that is reflected in overt action. (See a sophisticated
discussion of the possibility of unconscious self-report in Bayne, 2010.) Hence, given the findings
in which efference binding occurs unconsciously, one can no longer be sure that the animals
were conscious when responding accurately (and as humans do) to the rivalrous stimuli (though
there are good arguments, by analogy, that they were conscious; Gray, 2004). We refer to this as
the self-report paradox: In terms of overt behavior and from an objective standpoint, when
does a discrimination constitute self-report? A second question that emerges from this
paradox pertains specifically to voluntary action: If the correct choice response is issued to a
subliminal stimulus, does such an action qualify as a voluntary action, even if the processes
mediating the action are unconscious?
The Problem of Subjectivity

The most challenging outstanding questions are associated with the subjectivity (i.e., basic
consciousness) associated with voluntary action. As mentioned, Chomsky (1988) notes that
humans, unlike machines, not only can be compelled to act a certain way (e.g., drop a hot dish)
but also can be inclined to act a certain way (e.g., have the urge to drop a hot dish, while not
dropping it). As mentioned earlier, we construed such inclinations as action options. In the
classic models of artificial intelligence, it is these subjective inclinations that are absent from
theorizing. This omission may reflect that it is not only the case that we do not understand how
volition emerges from brain function, but that we do not understand how anything like
subjectively experienced volition could stem from any set of real or hypothetical circumstances.
More generally, there is no theoretical model that explains how something conscious (e.g., the
human organism) could arise from something unconscious (e.g., presumably a collection of
neurons; Godwin et al., 2013). There is nothing included in these mechanistic, neural networks
of intelligent behavior that would give rise to subjectivity (as far as we know). One may well
envision how a neural network lacking consciousness could instantiate many of the actions that
consciousness happens to accompany (e.g., motivated behavior such as for air, hunger), (p.47)
but such speculations reflect more ones powers of imagination than what may have arisen in
the course of natural history. (It should be mentioned that the subfield of synthetic
phenomenology endeavors to address the problem of phenomenology by situating the first-
person and third-person perspectives in a computational architecture; Aleksander & Morton,
2006.)
Thus, the conceptual challenge in reverse engineering consciousness in the brain is

compounded by another challenge: whatever consciousness is, it is something that has been
crafted by the happenstance and tinkering hands of evolution, which tend to produce things in a
manner that is different from our human way of producing things. For instance, artifacts created
to circulate the blood (the artificial heart), to locomote (e.g., wheels), or to lift (e.g., the pulley)
are very different from their biological counterparts (i.e., hearts, legs, and arms). In the course
of phylogeny, suboptimal and counterintuitive strategies are often selected to solve
computational problems (Gould, 1977; Marcus, 2008, 2012; Roe & Simpson, 1958; Simpson,
Page 18 of 37
1949). As natural scientists, neuroscientists are concerned with how a system works (a
descriptive approach), and not necessarily with how a system should work (a normative
approach), which is the concern of subfields of artificial intelligence (Arkin, 1998). Thus, in
natural science, intuitions regarding how the nervous system should work take a back seat to
actual data revealing the manner in which it actually works, regardless of whether such
functioning is optimal or suboptimal. Regarding suboptimal products of evolution, one must
consider that something like the human backbone is designed, not for bipeds, but for
quadrupeds, that is, creatures that support their weight with four legs on the ground. Engineers
have argued that, for humans, a more optimal design for the weight of the head and upper body
requires, not one, but three backbones (Marcus, 2008). Some scientists argue that the brain
contains many Kluges of this kind (Marcus, 2008). (Kluge is an engineers term for a clumsy
or inelegant solution to an engineering puzzle, which results in a functional but inefficient
design.)
It is important to keep in mind that, although our artificial models of nervous function may
actually be more optimal than what nature has provided, one has to consider that, given the
materials that the process of evolution had to work with when crafting intelligent systems, the
strategies picked through natural selection may have actually been quite good, at least given the
context in which the strategies were applied (Allman, 2000). For instance, it may be inefficient
to design a serial processing computer that, like humans, activates conflicting plans
simultaneously. In such conflicts, the plans to perform X and to not perform X are simultaneously
active, which is energy inefficient. In humans, this scenario arises in classic conflicts (e.g.,
approach-approach and avoidance-avoidance conflicts; Lewin, 1935; Miller, 1959). Nevertheless,
it has been argued that, in a system with relatively slow units working in parallel (as (p.48) in
the brain, but not in a fast computer circuit board), intrapsychic conflict is actually optimal
(Livnat & Pippenger, 2006). In this way, what appears to be a suboptimal strategy may not be so
suboptimal when considering the conditions under which the strategy has evolved or is applied.
In the attempt to explain the nature of consciousness by proposing that it resembles something
already known to us, perhaps it could be stated that the evidence suggests that consciousness is
a peculiar form of information gathering, a kind of broadcasting of information (e.g., action
options) that is different in nature from any form of human-made communication. This
broadcasting resembles that of a lighthouse, in which a single signal (i.e., the conscious field)
can be processed by various receivers (perhaps unconscious motor systems; Morsella & Bargh,
2010). With this in mind, one may dare say that, just as the nephron is like a filter and the heart
is like a pump, consciousness is like an information broadcast system, one that we do not yet
understand. There are artificial models of consciousness (Shallice, 1972) that are consonant
with this broadcast notion, including the LIDA (Learning IDA) model, in which the broadcasting
of information during action control is intimately related to conscious processing (Franklin &
Baars, 2009).
The Problem of Affinity-Based Systems

Action-related inclinations are reliably elicited events, whether instantiated in the laboratory or
in everyday life, as when holding ones breath or gargling mouthwash for 30 seconds
conditions in which one experiences incompatible intentions. Like urges, mental imagery, and
conscious thought, these covert processes fall outside the explanatory scope of contemporary
Page 19 of 37
conceptualizations about intelligent processes in the nervous system. As mentioned, no neural

network model has provided an explanation of how an inclination could arise from the activity of
mechanistic units such as neurons. For something to have an inclination toward objects in the
world, it must be capable of representing not only a world within which it resides, but also its
preferences (see relevant neural network approach in Holland & Goodman, 2003). How a
physical thing can prefer to be in certain states versus others remains a mystery for which we
have no conceptual clues (Bargh & Morsella, 2008). Thus, another outstanding question is the
following. How could any physical system be an affinity-based system, preferring some states
over others? When describing the brain, the natural scientist must explain the physical nature of
subjective inclinations, whether the events be counterintuitive or Kluge-like.
One might argue that these challenges could be swept away easily by proposing that subjectivity
is an epiphenomenon that serves no function in behavior (see Hommel, 2013). Indeed, it has
been claimed that, much as the sound of a (p.49) steam whistle is epiphenomenal to the
function of the locomotive, consciousness is epiphenomenal to brain function. (This has been
attributed to Huxley [1874] in his debate with William James about the functions of
consciousness.) Unfortunately, to make such a claim, one must first understand the place in
nature of the phenomenon at hand, which is the responsibility of the reverse engineer. Science
knows why the steam whistle makes a sound, but we have no clue regarding what subjectivity is.
Conclusion
The ancient Egyptians and Greeks, including Hippocrates (c. 460370 B.C.), believed that
volition and consciousness arise from brain function, but such a sensible view was uncommon
for vast stretches of time in intellectual history, as when Cartesian dualism dominated thinking,
or in medieval times, when the mind and will were held to be a function of the heart. Only
during the last four decades have mainstream researchers in psychology and neuroscience
begun to amass a catalog of facts regarding which activities in brain structures are unconscious
and which influence voluntary action. These empirical developments have been of interest to
theoreticians in the fields of psychology, neuroscience, linguistics, and artificial intelligence.
Many theoreticians strive to reverse engineer voluntary action in the brain (Aleksander, 1996;
Franklin, Ramamurthy, DMello, McCauley, Negatu, Silva, & Datla, 2007; Hawkins & Blakeslee,
2005; Holland & Goodman, 2003; Ito, Miyashita, & Rolls, 2007; McDermott, 2007; Minsky, 2006;
Shallice, 1972). As noted by one of the major contributors to artificial intelligence (Kurzweil,
2012, cf. Marcus, 2012), the reverse engineering of the kinds of conscious processes associated
with voluntary action has proven to be one of the most daunting puzzles in science (Levine,
1983; Roach, 2005).
Our review of the literature reveals that unconscious processes can guide action control, from
stimulus input to motor control, as in the case of reflexes, automatisms, and laboratory-based
forms of unconscious efference binding (e.g., choice responses to subliminal stimuli; Hallett,
2007). In addition, during voluntary actionwhich we have construed as a complex and
multifaceted form of actionseveral aspects of processing can be unconscious, such as (a)
motor control, (b) the unconscious mechanisms introducing action options and urges into
consciousness, and (c) the inferred unconscious homunculus (responsible for action selection
and action guidance). Aspects of voluntary action that are most intimately linked to
Page 20 of 37
consciousness are the representations of the action options (e.g., urges) and the sense of
agency, which includes at least two variants. Together, the literature reveals that, even when
action is voluntarythat which is most intimately related to ones conscious willunconscious
mechanisms interact with conscious mechanisms in an influential manner, at all stages of
processing.
Notes
References
Alais, D., & Blake, R. (2005). Binocular rivalry. Cambridge, MA: MIT Press.
Aleksander, I. (1996). Impossible minds: my neurons, my consciousness. London:Imperial

College Press.
Aleksander, I., & Morton, H. (2006). On architectures for synthetic phenomenology. Integrative
Approaches to Machine Intelligence Conference, 1, 108115.
Allman, J. M. (2000). Evolving brains. New York: Scientific American Library.
Angell, J. R. (1907). The province of functional psychology. Psychological Review, 14, 6191.
Arkin, R. C. (1998). Behavior-based robotics. Cambridge, MA: MIT Press.
Baars, B. J. (1988). A cognitive theory of consciousness. Cambridge: Cambridge University Press.
Baars, B. J. (1998). The function of consciousness: reply. Trends in Neurosciences, 21, 201.
Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press.
Balslev, D., Cole, J., & Miall, R. C. (2007). Proprioception contributes to the sense of agency
during visual observation of hand movements: evidence from temporal judgments of action.
Journal of Cognitive Neuroscience, 19, 15351541.
Bargh, J. A., & Morsella, E. (2008). The unconscious mind. Perspectives on Psychological
Science, 3, 7379.
Bartolomei, F., Wendling, F., Vignal, J. P., Chauvel, P., & Liegeois-Chauvel, C. (2002). Neural
networks underlying epileptic humming. Epilepsia, 43, 10011012.
Bayne, T. (2010). The unity of consciousness. New York: Oxford University Press.
Bernstein, N. A. (1996). On dexterity and its development. In M. L. Latash & M. T. Turvey (Eds.),
Dexterity and its development (pp. 1244). Mahwah, NJ: Erlbaum Associates.
Berti, A., & Pia, L. (2006). Understanding motor awareness through normal and pathological
behavior. Current Directions in Psychological Science, 15, 245250.
Page 21 of 37
Bindra, D. (1974). A motivational view of learning, performance, and behavior modification.

Psychological Review, 81, 199213.
Blakemore, S. J., Wolpert, D., & Frith, C. (2000). Why cant you tickle yourself? Neuroreport, 3,
R11R16.
Blanken, G., Wallesch, C., W., & Papagno, C. (1990). Dissociations of language functions in
aphasics with speech automatisms (recurring utterances). Cortex, 26, 4163.
Boly, M., Garrido, M. I., Gosseries, O., Bruno, M. A., Boveroux, P., Schnakers, C., & Friston, K.
(2011). Preserved feedforward but impaired top-down processes in the vegetative state. Science,
332, 858862.
Bortolotti, L., & Broome, M. (2009). A role for ownership and authorship in the analysis of
thought insertion. Phenomenology and the Cognitive Sciences, 8, 205224.
Botvinick, M. M., & Bylsma, L. M. (2005). Distraction and action slips in an everyday task:
evidence for a dynamic representation of task context. Psychonomic Bulletin and Review, 12,
10111017.
Buchsbaum, B. R. (2013). The role of consciousness in the phonological loop: hidden in plain
sight. Frontiers in Psychology, 4, doi: 10.3389/fpsyg.2013.00496
Bullock, D., & Grossberg, S. (1988). Neural dynamics of planned arm movements: emergent
invariants and speed-accuracy properties during trajectory formation. Psychological Review, 95,
4990.
Bridgeman, B. (2007). Efference copy and its limitations. Computers in Biology and Medicine,
37, 924929.
Bryon, S., & Jedynak, C. P. (1972). Troubles du transfert interhemispherique: a propos de trois
observations de tumeurs du corps calleux. Le signe de la main etrangre. Revue Neurologique,
126, 257266.
Campbell, B. A., & Misanin, J. R. (1969). Basic drives. Annual Review of Psychology, 20, 5784.
Carmant, L., Riviello, J. J., Thiele, E. A., Kramer, U., et al. (1994). Compulsory spitting: an usual
manifestation of temporal lobe epilepsy. Journal of Epilepsy, 7, 167170.
Chambon, V., Wenke, D., Fleming, S. M., Prinz, W., & Haggard, P. (2013). An online neural
substrate for sense of agency. Cerebral Cortex, 23, 10311037.
Chan, J.-L., & Ross, E. D. (1997). Alien hand syndrome: influence of neglect on the clinical
presentation of frontal and callosal variants. Cortex, 33, 287299.
Chomsky, N. (1988). Language and problems of knowledge: the Managua lectures. Cambridge,
MA: MIT Press.
Page 22 of 37
Clark, A. (2002). Is seeing all it seems? Action, reason and the grand illusion. Journal of
Consciousness Studies, 9, 181202.
Cohen, J. D., Dunbar, K. & McClelland, J. L. (1990). On the control of automatic processes: a
parallel distributed processing account of the Stroop effect. Psychological Review, 97, 332361.
Cole, J. (2007). The phenomenology of agency and intention in the face of paralysis and
insentience. Phenomenology and the Cognitive Sciences, 6, 309325.
Coles, M. G. H., Gratton, G., Bashore, T. R., Eriksen, C. W., & Donchin, E. (1985). A
psychophysiological investigation of the continuous flow model of human information
processing. Journal of Experimental Psychology: Human Perception and Performance, 11, 529
553.
Corr, P. J., & Morsella, E. (in press). The conscious control of behavior: revisiting Grays
comparator model. In P. J. Corr, M. Fajkowska, & Eysenck, M. (Eds.), Personality and control,
Vol. 4. New York: Eliot Werner Publications.
Crick, F. (1995). The astonishing hypothesis: the scientific search for the soul. New York:
Touchstone.
Crick, F., & Koch, C. (2000). The unconscious homunculus. In T. Metzinger (Ed.), Neural
correlates of consciousness (pp. 103110). Cambridge, MA: MIT Press.
Csikszentmihalyi, M., & Csikszentmihalyi, I. S. (1988). Optimal experience: psychological studies

of flow in consciousness. Cambridge: Cambridge University Press.
Curtis, C. E., & DEsposito, M. (2009). The inhibition of unwanted actions. In E. Morsella, J. A.
Bargh, & P. M. Gollwitzer (Eds.), Oxford handbook of human action (pp. 7297). New York:
Oxford University Press.
Damasio, A. R. (1989). Time-locked multiregional retroactivation: a systems-level proposal for

the neural substrates of recall and recognition. Cognition, 33, 2562.
David, N., Newen, A., & Vogeley, K. (2008). The sense of agency and its underlying cognitive
and neural mechanisms. Consciousness and Cognition, 17, 523534.
Decety, J., & Grzes, J. (1999). Neural mechanisms subserving the perception of human actions.
Trends in Cognitive Sciences, 3, 172178.
Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: Basic
evidence and a workspace framework. Cognition, 79, 137.
Del Cul, A., Baillet, S., & Dehaene, S. (2007). Brain dynamics underlying the nonlinear threshold
for access to consciousness. PLoS Biology, 5, e260.
Dennett, D. C. (1991). Consciousness explained. Boston: Little, Brown.
Page 23 of 37
Desender, K., van Lierde, E., & van den Bussche, E. (2013). Comparing conscious and
unconscious conflict adaptation. PLoS ONE, 8(2), Article e55976. Retrieved from .
Desender, K., van Opstal, F. V., & van den Bussche, E. (2014). Feeling the conflict: The crucial
role of conflict experience in adaptation. Psychological Science, 25, 675683.
Desmurget, M., Reilly, K. T., Richard, N., Szathmari, A., Mottolese, C., & Sirigu, A. (2009).
Movement intention after parietal cortex stimulation in humans. Science, 324(5928), 811813.
Desmurget, M., & Sirigu, A. (2010). A parietal-premotor network for movement intention and
motor awareness. Trends in Cognitive Sciences, 13, 411419.
DeSoto, M. C., Fabiani, M., Geary, D. C., & Gratton, G. (2001). When in doubt, do it both ways:
brain evidence of the simultaneous activation of conflicting responses in a spatial Stroop task.
Journal of Cognitive Neuroscience, 13, 523536.
Doesburg, S. M., Green, J. L., McDonald, J. J., & Ward, L. M. (2009). Rhythms of consciousness:
binocular rivalry reveals large-scale oscillatory network dynamics mediating visual perception.
PLoS, 4, 114.
Doherty, M. J., Wilensky A. J., Holmes, M. D., Lewis, D. H., Rae, J., & Cohn, G. H. (2002). Singing
seizures. Neurology, 59, 14351438.
Dormashev, Y. (2010). Flow experience explained on the grounds of an activity approach to

attention. In B. Bruya (Ed.), Effortless attention (pp. 287333). Cambridge, MA: MIT Press.
Edelman, G. M., & Tononi, G. (2000). A universe of consciousness: how matter becomes
imagination. New York: Basic Books.
Research, 229, 475484.
Ellis, R. (2009). Interactions between action and visual objects. In E. Morsella, J. A. Bargh, & P.
M. Gollwitzer (Eds.), Oxford handbook of human action (pp. 214224). New York: Oxford
University Press.
Enatsu, R., Hantus, S., Gonzalez-Martinez, J., & So, N. (2011). Ictal singing due to left frontal
lobe epilepsy: a case report and review of the literature. Epilepsy & Behavior, 22, 404406.
Engbert, K., Wohlschlager, A. & Haggard, P. (2008). Who is causing what? The sense of agency
is relational and efferent-triggered. Cognition, 107, 693704.
Enger, T., & Hirsch, J. (2005). Cognitive control mechanisms resolve conflict through cortical
amplification of task-relevant information. Nature Neuroscience, 8, 17841790.
Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target
letter in a nonsearch task. Perception and Psychophysics, 16, 143149.
Page 24 of 37
Eriksen, C. W., & Schultz, D. W. (1979). Information processing in visual search: a continuous
flow conception and experimental results. Perception and Psychophysics, 25, 249263.
Fecteau, J. H., Chua, R., Franks, I., & Enns, J. T. (2001). Visual awareness and the online
modification of action. Canadian Journal of Experimental Psychology, 55, 104110.
Fehrer, E., & Biederman, I. (1962). A comparison of reaction time and verbal report in the
detection of masked stimuli. Journal of Experimental Psychology, 64, 126130.
Fehrer, E., & Raab, D. (1962). Reaction time to stimuli masked by metacontrast. Journal of
Experimental Psychology, 63, 143147.
Firestone, C., & Scholl, B. J. (2014). Top-down effects where none should be found: the El
Greco fallacy in perception research. Psychological Science, 25, 3846.
Fodor, J. A. (1983). Modularity of mind: an essay on faculty psychology. Cambridge, MA: MIT
Press.
Franklin, S., & Baars, B. (2009). Consciousness is computational: the LIDA model of global
workspace theory. International Journal of Machine Consciousness, 1, 110.
Franklin, S., Ramamurthy, U., DMello, S. K., McCauley, L., Negatu, A., Silva L., R., & Datla, V.
(2007). LIDA: a computational model of global workspace theory and developmental learning. In
AAAI Fall Symposium on AI and Consciousness: theoretical foundations and current approaches.
Arlington, VA: AAAI.
Freeman, W. J. (1991). The physiology of perception. Scientific American, 264, 7885.
Ganz, L. (1975). Temporal factors in visual perception. In E. C. Carterette & M. P. Friedman

(Eds.), Handbook of perception (Vol. 5, pp. 169231). New York: Academic Press.
Gazzaley, A., Cooney, J.W., Rissman, J., & DEsposito, M. (2005). Top-down suppression deficit
underlies working memory impairment in normal aging. Nature Neuroscience, 8, 12981300.
Georgopoulos, A. P., Caminiti, R., Kalaska, J. F., & Massey J. T. (1983). Spatial coding of
movement: a hypothesis concerning the coding of movement direction by motor cortical
populations. Experimental Brain Research: Supplement, 7, 327336.
Gibson, J. J. (1979). The ecological approach to visual perception. Boston, MA: Houghton-Mifflin.
Godwin, C. A., Gazzaley, A., & Morsella, E. (2013). Homing in on the brain mechanisms linked to
consciousness: buffer of the perception-and-action interface. In A. Pereira & D. Lehmann (Eds.),
The unity of mind, brain and world: current perspectives on a science of consciousness (pp. 43
76). Cambridge: Cambridge University Press.
Page 25 of 37
Gold, J. I., & Shadlen, M. N. (2007). The neural basis of decision making. Annual Review of
Neuroscience, 30, 535574.
Goldberg, I. I., Harel, M., & Malach, R. (2006). When the brain loses its self: prefrontal
inactivation during sensorimotor processing. Neuron, 50, 329339.
Goldrick, M., & Blumstein, S. E. (2006). Cascading activation from phonological planning to
articulatory processes: evidence from tongue twisters. Language and Cognitive Processes, 21,
649683.
Goodale, M., & Milner, D. (2004). Sight unseen: an exploration of conscious and unconscious
vision. New York: Oxford University Press.
Gould, S. J. (1977). Ever since Darwin: reflections in natural history. New York: Norton.
Grafman, J., & Krueger, F. (2009). The prefrontal cortex stores structured event complexes that
are the representational basis for cognitively derived actions. In E. Morsella, J. A. Bargh, & P. M.
Gollwitzer (Eds.), Oxford handbook of human action (pp. 197213). New York: Oxford University
Press.
Gray, J. A. (1995). The contents of consciousness: a neuropsychological conjecture. Behavioral

and Brain Sciences, 18, 659676.
Gray, J. A. (2004). Consciousness: creeping up on the hard problem. New York: Oxford
University Press.
Gray, J. R., Bargh, J. A., & Morsella, E. (2013). Neural correlates of the essence of conscious
conflict: fMRI of sustaining incompatible intentions. Experimental Brain Research, 229, 453465.
Grossberg, S. (1999). The link between brain learning, attention, and consciousness.
Guggisberg, A. G., Dalal, S. S., & Nagarajan, S. S. (2009). The neural basis of introspection.
NeuroImage, 47, S39S41.
Haggard, P. (2005). Conscious intention and motor cognition. Trends in Cognitive Sciences, 9,
290295.
Haggard, P. (2008). Human volition: towards a neuroscience of will. Nature Neuroscience

Reviews, 9, 934946.
Haggard, P., Aschersleben, G., Gehrke, J., & Prinz, W. (2002). Action, binding and awareness. In
W. Prinz & B. Hommel (Eds.), Common mechanisms in perception and action: attention and
performance (Vol. XIX, pp. 266285). Oxford: Oxford University Press.
Haggard, P., & Clark, S. (2003). Intentional action: conscious experience and neural prediction.
Page 26 of 37
Nature Neuroscience, 5, 382385.
Hallett, M. (2007). Volitional control of movement: the physiology of free will. Clinical
Neurophysiology, 117, 11791192.
Harless, E. (1861). Der apparat des willens [The apparatus of the will]. Zeitshrift fr Philosophie
und philosophische Kritik, 38, 499507.
Hawkins, J., & Blakeslee, S. (2005). On intelligence. New York: Henry Holt.
Hazy, T. E., Frank, M. J., & OReilly, R. C. (2006). Banishing the homunculus: making working
memory work. Neuroscience, 139, 105118.
Heath, M., Neely, K. A., Yakimishyn, J., & Binsted, G. (2008). Visuomotor memory is independent
of conscious awareness of target features. Experimental Brain Research, 188, 517527.
Heckhausen, H., & Beckmann, J. (1990). Intentional action and action slips. Psychological
Review, 97, 3648.
Helmholtz, H. v. (1856/1925). Treatise of physiological optics: concerning the perceptions in

general. In T. Shipley (Ed.), Classics in psychology (pp. 79127). New York: Philosophy Library.
Holland, O., & Goodman, R. (2003). Robots with internal models: a route to machine
consciousness. Journal of Consciousness Studies, 10, 77109.
Hommel, B. (2009). Action control according to TEC (theory of event coding). Psychological
Research, 73, 512526.
Hommel, B. (2010). Grounding attention in action control: the intentional control of selection. In
B. Bruya (Ed.), Effortless attention (pp. 121140). Cambridge, MA: MIT Press.
Hommel, B. (2013). Dancing in the dark: no role for consciousness in action control. Frontiers in
Psychology, 4, doi: 10.3389/fpsyg.2013.00380
Hommel, B., & Elsner, B. (2009). Acquisition, representation, and control of action. In E.
Morsella, J. A. Bargh, & P. M. Gollwitzer (Eds.), Oxford handbook of human action (pp. 371398).
New York: Oxford University Press.
Hommel, B., Msseler, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding: a
framework for perception and action planning. Behavioral and Brain Sciences, 24, 849937.
Hume, D. ([1739] 1888). A treatise on human nature. Ed. L. A. Selby-Bigge. London: Oxford
University Press. (Original work published 1739.)
Huxley, T. H. (1874). Animal automatism, and other essays. New York: J. Fitzgerald.
Page 27 of 37
Iacoboni, M., & Dapretto, M. (2006). The mirror neuron system and the consequences of its
dysfunction. Nature Reviews Neuroscience, 7, 942951.
Ito, M., Miyashita, Y., & Rolls, E. T. (2007). Cognition, computation, and consciousness. New
York: Oxford University Press.
Jackendoff, R. S. (1990). Consciousness and the computational mind. Cambridge, MA: MIT Press.
James, W. (1890). The principles of psychology. New York: Dover.
Jeannerod, M. (2003). Simulation of action as a unifying concept for motor cognition. In S. H.

Johnson-Frey (Ed.), Taking action: cognitive neuroscience perspectives on intentional acts (pp.
139166). Cambridge, MA: MIT Press.
Jeannerod, M. (2009). The sense of agency and its disturbances in schizophrenia: a reappraisal.
Experimental Brain Research, 196, 527532.
Jersild, A. T. (1927). Mental set and shift. Archive of Psychology, 14, 89.
Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall
Kaido, T., Otsuki, T., Nakama, H., Kaneko, Y., Kubota, Y., Sugai, K., & Saito, O. (2006). Complex
behavioral automatism arising from insular cortex. Epilepsy and Behavior, 8, 315319.
Kececi, H., Degirmenci, Y., & Gumus, H. (2013). Two foreign language automatisms in complex
partial seizures. Epilepsy and Behavior Case Reports, 1, 79.
Kern, M. K., Jaradeh, S., Arndorfer, R. C., & Shaker, R. (2001). Cerebral cortical representation
of reflexive and volitional swallowing in humans. American Journal of Physiology:
Gastrointestinal and Liver Physiology, 280, G354G360.
Klein, D. B. (1984). The concept of consciousness: a survey. Lincoln: University of Nebraska

Press.
Knoblich, G., & Repp, B. H. (2009). Inferring agency from sound. Cognition, 111, 248262.
Koch, C. (2012). Consciousness: confessions of a romantic reductionist. Cambridge, MA: MIT

Press.
Kriegel, U. (2007). A cross-order integration hypothesis for the neural correlate of

consciousness. Consciousness and Cognition, 16, 897912.
Kurzweil, R. (2012). How to create a mind: the secret of human thought revealed. New York:
Viking Adult
Lashley, K. S. (1942). The problem of cerebral organization in vision. In H. Kluver (Ed.), Visual
mechanisms. Biological symposia (Vol. 7, pp. 301322). Lancaster, PA: Cattell Press.
Page 28 of 37
Laureys, S. (2005). The neural correlate of (un)awareness: lessons from the vegetative state.
Leopold, D. A., & Logothetis, N. K. (1996). Activity-changes in early visual cortex reflect
monkeys percepts during binocular rivalry. Nature, 379, 549553.
Levelt, W. J. M. (1989). Speaking: from intention to articulation. Cambridge, MA: MIT Press.
Levine, J. (1983). Materialism and qualia: the explanatory gap. Pacific Philosophical Quarterly,
64, 354361.
Lewin, K. (1935). A dynamic theory of personality. New York: McGraw-Hill.
Lhermitte, F. (1983). Utilization behavior and its relation to lesions of the frontal lobes. Brain,
106, 237255.
Libet, B. (2004). Mind time: the temporal factor in consciousness. Cambridge, MA: Harvard
University Press.
Liu, G., Chua, R., & Enns, J. T. (2008). Attention for perception and action: Task interference for
action planning, but not for online control. Experimental Brain Research, 185, 709717.
Livnat, A., & Pippenger, N. (2006). An optimal brain can be composed of conflicting agents.
Proceedings of the National Academy of Sciences USA, 31983202.
Llins, R. R., & Ribary, U. (2001). Consciousness and the brain: the thalamocortical dialogue in
health and disease. Annals of the New York Academy of Sciences, 929, 166175.
Loewenstein, G. (1996). Out of control: visceral influences on behavior. Organizational Behavior

and Human Decision Processes, 65, 272292.
Logan, G. D. (2003). Executive control of thought and action: in search of the wild homunculus.
Current Directions in Psychological Science, 12, 4548.
Lotze, R. H. (1852). Medizinische Psychologie oder Physiologie der Seele. Leipzig:

Weidmannsche Buchhandlung.
Lynn, M. T., Berger, C. C., Riddle, T. A., & Morsella, E. (2010). Mind control? Creating illusory
intentions through a phony brain-computer interface. Consciousness and Cognition, 19, 1007
1012.
Maestro, I., Carreno, M., Donaire, A., Rumia, J., Conesa, G., Bargallo, N., Falcon, C., Setoain, X.,
Pintor, L., & Boget, T. (2008). Oroalimentary automatisms induced by electrical stimulation of
the fronto-opercular cortex in a patient without automotor seizures. Epilepsy and Behavior, 13,
410412.
Marchetti, C., & Della Sala, S. (1998). Disentangling the alien and anarchic hand. Cognitive
Neuropsychiatry, 3, 191207.
Page 29 of 37
Marcus, G. (2008). Kluge: the haphazard construction of the human mind. New York: Houghton
Mifflin.
Marcus, G. (2012, November 15). Ray Kurzweils dubious new theory of mind. The New Yorker.
Mattler, U. (2005). Flanker effects on motor output and the late-level response activation
hypothesis. The Quarterly Journal of Experimental Psychology, 58A, 577601.
McClelland, J. L. (1979). On the time-relations of mental processes: an examination of systems of

processes in cascade. Psychological Review, 86, 287330.
McDermott, D. (2007). Artificial intelligence and consciousness. In P. D. Zelazo, M. Moscovitch,

& E. Thompson (Eds.), The Cambridge handbook of consciousness (pp. 117150). Cambridge,
UK: Cambridge University Press.
McFarland, D. J., & Sibly, R. M. (1975). The behavioural final common path. Philosophical
Transactions of the Royal Society (London), 270, 265293.
Merker, B. (2007). Consciousness without a cerebral cortex: a challenge for neuroscience and
medicine. Behavioral and Brain Sciences, 30, 63134.
Merker, B. (2013). The efference cascade, consciousness, and its self: naturalizing the first
person pivot of action control. Frontiers in Psychology, 4, doi: 10.3389/fpsyg.2013.00501
Miall, R. C. (2003). Connecting mirror neuron and forward models. Neuroreport, 14, 13.
Mikati, M. A., Comair, Y. G., & Shamseddine, A. N. (2005). Pattern-induced partial seizures with
repetitive affectionate kissing: an unusual manifestation of right temporal lobe epilepsy.
Epilepsy and Behavior, 6, 447451.
Miller, E. K. (2000). The prefrontal cortex and cognitive control. Nature Reviews Neuroscience,
1, 5965.
Miller, N. E. (1959). Liberalization of basic S-R concepts: extensions to conflict behavior,

motivation, and social learning. In S. Koch (Ed.), Psychology: a study of science (Vol. 2, pp. 196
292). New York: McGraw-Hill.
Milner, A. D., & Goodale, M. (1995). The visual brain in action. New York: Oxford University
Press.
Minsky, M. (2006). The emotion machine: commonsense thinking, artificial intelligence, and the
future of the human mind. New York: Simon and Schuster.
Molapour, T., Berger, C. C., & Morsella, E. (2011). Did I read or did I name? Process blindness
from congruent processing outputs. Consciousness and Cognition, 20, 17761780.
Moore, J. W., Wegner, D. M., & Haggard, P. (2009). Modulating the sense of agency with
external cues. Consciousness and Cognition, 18, 10561064.
Page 30 of 37
Morsella, E. (2005). The function of phenomenal states: supramodular interaction theory.

Morsella, E., & Bargh, J. A. (2010). What is an output? Psychological Inquiry, 21, 354370.
Morsella, E., & Bargh, J. A. (2011). Unconscious action tendencies: sources of un-integrated
action. In J. T. Cacioppo & J. Decety (Eds.), The handbook of social neuroscience (pp. 335347).
New York: Oxford University Press.
Morsella, E., Berger, C. C., & Krieger, S. C. (2011). Cognitive and neural components of the
phenomenology of agency. Neurocase, 17, 209230.
Morsella, E., Gray, J. R., Krieger, S. C., & Bargh, J. A. (2009a). The essence of conscious conflict:
subjective effects of sustaining incompatible intentions. Emotion, 9, 717728.
Morsella, E., Hoover, M. A., & Bargh, J. A. (2013). Functionalism redux: how adaptive action
constrains perception, simulation, and evolved intuitions. In K. L. Johnson & M. Shiffrar (Eds.),
People watching: social, perceptual, and neurophysiological studies of body perception (pp. 256
282). New York: Oxford University Press.
Morsella, E., Krieger, S. C., & Bargh, J. A. (2010). Minimal neuroanatomy for a conscious brain:
homing in on the networks constituting consciousness. Neural Networks, 23, 1415.
Morsella, E., Lanska, M., Berger, C. C., & Gazzaley, A. (2009b). Indirect cognitive control
through top-down activation of perceptual symbols. European Journal of Social Psychology, 39,
11731177.
Morsella, E., & Miozzo, M. (2002). Evidence for a cascade model of lexical access in speech
production. Journal of Experimental Psychology: Learning, Memory, and Cognition, 28, 555563.
Morsella, E., Wilson, L. E., Berger, C. C., Honhongva, M., Gazzaley, A., & Bargh, J. A. (2009c).
Subjective aspects of cognitive control at different stages of processing. Attention, Perception, &
Psychophysics, 71, 18071824.
Morsella, E., Zarolia, P., & Gazzaley, A. (2012). Cognitive conflict and consciousness. In B.
Gawronski & F. Strack (Eds.), Cognitive consistency: a unifying concept in social psychology (pp.
1946). New York: Guilford Press.
Mller, J. (1843). Elements of physiology. Philadelphia: Lea and Blanchard.
Nagel, T. (1974). What is it like to be a bat? Philosophical Review, 83, 435450.
Navarrete, E., & Costa, A. (2004). How much linguistic information is extracted from ignored
pictures? Further evidence for a cascade model of speech production. Journal of Memory and
Language, 53, 359377.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: verbal reports on mental
processes. Psychological Review, 84, 231259.
Page 31 of 37
Norman, D. A., & Shallice, T. (1980). Attention to action: willed and automatic control of
behavior. In R. J. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and self-
regulation (pp. 118). New York: Plenum Press.
Ortinski, P., & Meador, K. J. (2004). Neuronal mechanisms of conscious awareness. Archives of
Neurology, 61, 10171020.
Pacherie, E. (2008). The phenomenology of action: a conceptual framework. Cognition, 107, 179
217.
Perky, C. W. (1910). An experimental study of imagination. American Journal of Psychology, 21,

422452.
Pilon, M., & Sullivan, S. J. (1996). Motor profile of patients in minimally responsive and
persistent vegetative states. Brain Injury, 10, 421437.
Pinker, S. (1997). How the mind works. New York: W. W. Norton.
Plazzi, G., Vetrugno, R., Provini, F., & Montagna, P. (2005). Sleepwalking and other ambulatory
behaviors during sleep. Neurological Sciences, 26, S193S198.
Poehlman, T. A., Jantz, T. K., & Morsella, E. (2012). Adaptive skeletal muscle action requires
anticipation and conscious broadcasting. Frontiers in Psychology, 3, 369.
Prinz, W. (2003). How do we know about our own actions? In S. Maasen, W. Prinz, & Roth (Eds),
Voluntary action: brains, minds, and sociality (pp. 2133). London: Oxford University Press.
Prinz, J. (2007). The intermediate level theory of consciousness. In M. Velmans & S. Schneider
(Eds.), The Blackwell companion to consciousness (pp. 248260). Malden, MA: Blackwell
Publishing.
Proctor, R. W., & Vu, K.-P. L. (2010). Action selection. In I. B. Weiner & E. Craighead (Eds.), The
Corsini encyclopedia of psychology (Vol. 1, pp. 2022). Hoboken, NJ: John Wiley.
Raghavendra, S., Mirsattari, S., & McLachlan, R. (2010). Ictal whistling: a rare automatism
during temporal lobe seizures. Epileptic Disorders, 12, 133135.
Reid, T. ([1785] 1855). Essays on the intellectual powers of man (6th ed.). Boston, MA: Phillips,
Sampson, & Company. (Originally published in 1785).
Riddle, T. A., & Morsella, E. (2009). Is that me? Authorship processing as a function of intra-
psychic conflict. Proceedings of the Annual Convention of the Association for Psychological
Science, San Francisco, California.
Roach, J. (2005, June 30). Journal ranks top 25 unanswered science questions. National
Geographic News. Retrieved from news.nationalgeographic.com.
Roe, A., & Simpson, G. G. (1958). Behavior and evolution. New Haven, CT: Yale University Press.
Page 32 of 37
Roepstorff, A., & Frith, C. D. (2004). Whats at the top in the top-down control of action? Script-
sharing and top-top control of action in cognitive experiments. Psychological Research, 68,
189198.
Rosenbaum, D. A. (2002). Motor control. In H. Pashler (Series Ed.) & S. Yantis (Vol. Ed.),
Stevens handbook of experimental psychology, Vol. 1: Sensation and perception (3rd ed., pp.
315339). New York: Wiley.
Rossetti, Y. (2001). Implicit perception in action: short-lived motor representation of space. In P.

G. Grossenbacher (Ed.), Finding consciousness in the brain: a neurocognitive approach (pp.
133181). Netherlands: John Benjamins Publishing.
effect contribute to explicit judgment of agency. Cognition, 110, 7483.
Schacter, D. L., & Addis, D. R. (2007). The cognitive neuroscience of constructive memory:
remembering the past and imagining the future. Philosophical Transactions of the Royal Society
of London, Series B: Biological Sciences, 362, 773786.
Schenk, C. H., & Mahowald, M. W. (1995). A polysomnographically documented case of adult

somnambulism with long-distance automobile driving and frequent nocturnal violence:
parasomnia with continuing danger as a noninsane automatism? Sleep, 18, 765772.
Sergent, C., & Dahaene, S. (2004). Is consciousness a gradual phenomenon? Evidence for an all-
or-none bifurcation during the attentional blink. Psychological Science, 15, 720728.
Shallice, T. (1972). Dual functions of consciousness. Psychological Review, 79, 383393.
Simpson, G. G. (1949). The meaning of evolution. New Haven, CT: Yale University Press.
Skinner, B. F. (1953). Science and human behavior. New York: Macmillan.
Spencer, S. S., Spencer, D. D., Williamson, P. D., & Mattson, R. H. (1983). Sexual automatisms
in complex partial seizures. Neurology, 33, 527533.
Sperry, R. W. (1952). Neurology and the mind-brain problem. American Scientist, 40, 291312.
Squire, L. R. (1987). Memory and brain. New York: Oxford University Press.
Srinivasan, R., Russel, D. P., Edelman, G. M., & Tononi, G. (1999). Increased synchronization of
neuromagnetic responses during conscious perception. The Journal of Neuroscience, 19, 5435
5448.
Starreveld, P. A., Theeuwes, J., & Mortier, K. (2004). Response selection in visual search: the
influence of response compatibility of nontargets. Journal of Experimental Psychology: Human
Perception & Performance, 30, 5678.
Page 33 of 37
Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental

Suhler, C. L., & Churchland, P. S. (2009). Control: conscious and otherwise. Trends in Cognitive
Sciences, 13, 341347.
Suzuki, T., Itoh, S., Arai, N., Kouno, M., Noguchi, M., Takatsu, M., & Takeda, K. (2012). Ambient
echolalia in a patient with germinoma around the bilateral ventriculus lateralis: a case report.
Neurocase, 18, 330335.
Synofzik, M., Vosgerau, G., & Lindner, A. (2009). Me or not me: an optimal integration of agency
Tallon-Baudry, C. (2012). On the neural mechanisms subserving attention and consciousness.

Frontiers in Psychology, 2, 397.
Taylor, J. A., & Ivry, R. B. (2013). Implicit and explicit processes in motor learning. In W. Prinz,
M. Beisert, & A. Herwig (Eds.), Action science (pp. 6387). Cambridge, MA: MIT Press.
Taylor, J. L., & McCloskey, D. I. (1990). Triggering of preprogrammed movements as reactions

to masked stimuli. Journal of Neurophysiology, 63, 439446.
Taylor, J. L., & McCloskey, D. I. (1996). Selection of motor responses on the basis of unperceived
stimuli. Experimental Brain Research, 110, 6266.
Tononi, G. (2012). Phi: a voyage from the brain to the soul. New York: Pantheon.
Tononi, G., & Edelman, G. M. (1988). Consciousness and complexity. Science, 282, 18461851.
Tsakiris, M., Schtz-Bosbach, S., & Gallagher, S. (2007). On agency and body-ownership:
phenomenological and neurocognitive reflections. Consciousness and Cognition, 16, 645660.
Uhlhaas, P. J., Pipa, G., Lima, B., Melloni, L., Neuenschwander, S., Nikolic, D., & Singer, W.
(2009). Neural synchrony in cortical networks: history, concept and current status. Frontiers in
Integrative Neuroscience, 3, 17.
van Gaal, S., & Lamme, V. A. F. (2012). Unconscious high-level information processing:
implication for neurobiological theories of consciousness. Neuroscientist, 18, 287301. doi:
10.1177/1073858411404079
van Gaal, S., Lamme, V. A. F., & Ridderinkhof, K. R. (2010). Unconsciously triggered conflict
adaptation. PLoS ONE, 5(7), Article e11508. Retrieved from
van Veen, V., Cohen, J. D., Botvinick, M. M., Stenger, V. A., & Carter, C. C. (2001). Anterior
cingulate cortex, conflict monitoring, and levels of processing. Neuroimage, 14, 13021308.
Page 34 of 37
Varela, F., Lachaux, J. P., Rodriguez, E., & Martinerie, J. (2001). The brainweb: phase
synchronization and large-scale integration. National Review of Neuroscience, 2, 229239.
Vierkant, T. (2013). Managerial control and free mental agency. In A. Clark, J. Kiverstein, & T.
Vierkant (Eds), Decomposing the will (pp. 285297). New York: Oxford University Press.
Wegner, D. M. (2002). The illusion of conscious will. Cambridge, MA: MIT Press.
Wegner, D. M. (2003). The minds best trick: how we experience conscious will. Trends in
Cognitive Science, 7, 6569.
Wegner, D. M., & Wheatley, T. P. (1999). Apparent mental causation: sources of the experience
of will. American Psychologist, 54, 480492.
Weiskrantz, L. (1992). Unconscious vision: the strange phenomenon of blindsight. The Sciences,
35, 2328.
Weiskrantz, L. (1997). Consciousness lost and found: a neuropsychological exploration. New

Wood, W., Quinn, J., & Kashy, D. (2002). Habits in everyday life: thought, emotion, and action.
Journal of Personality and Social Psychology, 83, 12811297.
Woodworth, R. S. (1915). A revision of imageless thought. Psychological Review, 22, 127.
Wortzel, H., Strom, L., Anderson, A., Maa, E., & Spitz, M. (2011). Disrobing associated with
epileptic seizures and forensic implications. Journal of Forensic Science, 57, 550552.
Yamadori, A. (1997). Body awareness and its disorders. In M. Ito, Y. Miyashita, & E. T. Rolls
(Eds.), Cognition, computation, and consciousness (pp. 169176). Washington, DC: American
Psychological Association.
Zhang, W., & Rosenbaum, D. A. (2008). Planning for manual positioning: the end-state comfort
effect for abduction-adduction of the hand. Experimental Brain Research, 184, 383389.
Zeki, S., & Bartels, A. (1999). Toward a theory of visual consciousness. Consciousness and
Cognition, 8, 225259.
Zorick, F. J., Salis, P. J., Roth, T., & Kramer, M. (1979). Narcolepsy and automatic behavior.
Journal of Clinical Psychiatry, 40, 194197.
Notes:
(1) Sir Karl Popper concluded that the definition of a phenomenonexplaining what something
isis the final stage of scientific inquiry. Thus, establishing the definition of a phenomenon
should not be considered a prerequisite for investigating that phenomenon. From Poppers
standpoint, it is unfair to demand that the scientist first define consciousness in order to study
it, especially at this stage of understanding. Nevertheless, today there are good identifications
Page 35 of 37
and working definitions of what consciousness is. Regarding the former, one can say that
consciousness co-occurs with pain but seldom accompanies, say, peristalsis. This is an
identification of the circumstances co-occurring along with the phenomenon of interest.
Regarding a working definition, the best one to date has been set forth by the philosopher
Thomas Nagel (1974), who proposed that, for an organism to possess consciousness, there must
be something it is like to be that organism. For instance, there is certainly something it is like to
see the color red, experience breathlessness, or have a toothache. This basic form of
consciousness (to be contrasted with higher forms of consciousness, such as self-
consciousness) has fallen under the rubrics of the terms sentience (Pinker, 1997), subjective
experience, qualia, and phenomenal state.
(2) Consistent with a view in which there is no omnipresent decider involved in action selection,
neural evidence demonstrates that, when introspecting about two different kinds of perceptual
events (e.g., a visual percept or auditory percept), there is no common brain region activated
during both acts of introspection (Guggisberg, Dalal, & Nagarajan, 2009), as if there were no
single ever-present observing region.
(3) James (1890) observed that urges to avoid are associated with avoidant actions and that
urges to approach are associated with approach behaviors. To him, this systemic mapping
between conscious states and the associated behavioral inclinations supports the view that
consciousness is not epiphenomenal (i.e., serving no functional role whatsoever) but must serve
a purpose. Otherwise, he argued, why would the mappings between conscious states and overt
action be so systematic.
(4) This integration seems to be associated with only a subset of neural circuits and processes.
There are many regions whose nonparticipation does not render the nervous system incapable
of sustaining a basic form of consciousness. Regarding biological insults, for example, lesion
studies reveal that the non-participation of areas such as the spinal cord, cerebellum, basal
ganglia, hippocampus, amygdala, and corpus callosum lead to major deficits but not to the
eradication of consciousness (see reviews in Godwin et al., 2013; Morsella, Krieger, & Bargh,
2010). This observation has led investigators to attempt to isolate the brain regions constituting
consciousness. Such a distillation of a neural correlate of consciousness remains controversial.
Several hypotheses have been proposed, including that the substrate involves (a) subcortical
areas, (b) frontal cortex, (c) parietal cortex, (d) a network connecting frontal to parietal cortex,
or (e) connections between frontal and temporal cortex (see review in Godwin et al., 2013).
(5) In addition to its effects on consciousness (i.e., authorship processing), agency also leads to
effects that appear to be implicit, as in intentional binding (Haggard, Clark, Kalogeras, 2002). In
this effect, the perceived elapsed time between a voluntary action and its consequence is shorter
than the actual time span, as if the two events were temporally attracted to each other. Thus,
when pressing a button, for example, the experience of the button press and of hearing the
clicking sound of the act are perceived to occur more closely together in time than they actually
did. Although there may have been a half-second delay between pressing the button and hearing
the click, one would perceive the delay as shorter than a half-second. In this way, one also binds
the actions and outcomes performed by others (Engbert, Wohlschlger, Thomas, & Haggard,
Page 36 of 37
2007). Other implicit effects of agency include sensory attenuation (Blakemore, Wolpert, &
Frith, 2000) and other experimental effects (see Eitam, Kennedy, & Higgins, 2013).
(6) Bernstein ([1950] 1996) proposed that complex actions involving dexterity (e.g., piano
playing, deep sea diving) involve action-related processes that must be quick, reflex-like, and
sensitive to the current external environment (Wood, Quinn, & Kashy, 2002). As expertise
increases, action becomes more and more driven by external stimuli (Jersild, 1927). Other
activity theorists (e.g., Dobrynin and Leontiev) developed theories of attention that are related
to flow experiences. Of note, according to these theorists, flow states are associated with
postvoluntary attentiona notion based on the work of Bernstein (see Dormashev, 2010).
Leontievs theory of activity makes action selection in relation to specific cues the kernel of
psychological analysis. Dormashev (2010) interprets Leontievs theory as an activity gestalt
account, one that seeks to clarify the relation between automatic responses to action sets and
conscious content. Thus, the activity theorists may have anticipated the important theoretical
insight that goal selection is crucial to understanding the mechanisms underlying attention.
From this standpoint, attention can serve as the mediator between the selection of action effects
and unconscious automatic behavior (e.g., automatized motor control). The views of the activity
theorists are relevant to contemporary analyses of systematic associations between external
cues for action selection and task switching (see Hommel, 2010, pp. 134136).
(7) Regarding conceptual processing, though a continuous conflicting urge seems very different
phenomenologically from an intention-outcome mismatch, perhaps the conflicting urge is
nothing more than the reiterative cycling of a mismatch process (Morsella et al., 2011), the kind
embodied in comparator models of agency (Berti & Pia, 2006; David et al., 2008; Haggard,
2008). If future findings indicate that the most basic sense of agency requires such conceptual
(e.g., cause-and-effect reasoning) processes, then perhaps the processes must join the bundle of
sensations identified by Hume.
Page 37 of 37
Action Control by If-Then Planning

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Explicating the Mechanisms of Strategic Automaticity in Regard to Objective and Subjective

Agency
Torsten Martiny-Huenger
Sarah E. Martiny
Peter M. Gollwitzer
DOI:10.1093/acprof:oso/9780190267278.003.0003

This chapter explicates the mechanisms of if-then planning (i.e. implementation intentions),
namely, how verbal action planning leads to subsequent stimulus-driven action controla
mechanism referred to as strategic automaticity. The chapter first reviews evidence from
behavioral studies indicating that action control by if-then planning can exhibit features
associated with automatic action initiation; it is fast, effortless, and does not depend on another
in situ conscious intent. The chapter explicates underlying mechanisms of if-then planning by
integrating current simulation theories of cognition and action control into implementation
intention theory. The chapter also focuses on subjective agency: on the basis of currently
available evidence on how humans gain a feeling of control, it analyzes how if-then planning may
affect ones sense of agency. The chapter focuses on if-then planning as an ideal type of
planning; the mechanisms proposed, however, may be relevant to action planning in a more
general sense.
Keywords: if-then planning, implementation intentions, strategic automaticity, agency, sense of agency, action
control
Page 1 of 27
Introduction
Over the last decades, evidence from different research areas has accumulated that casts doubt
on the very intuitive idea that human actions are caused by conscious intentions (i.e., objective
agency; Bargh, Chen, & Burrows, 1996; Haggard & Eimer, 1999; Libet, Gleason, Wright, &
Pearl, 1983; Soon, Brass, Heinze, & Haynes, 2008; but see Baumeister, Masicampo, & Vohs,
2011). At the same time, there has been an increased interest in why we have such a pervasive
feeling that our intentions cause our actions (i.e., subjective agency; Bayne, 2008; Gallagher,
2000; Haggard & Tsakiris, 2009). In this chapter, we approach the question of objective and
subjective agency from a self-regulation perspective. We will focus on how actions can be
caused by conscious planning, that is, how future behavior can be intentionally automated by if-
then planning (i.e., implementation intentions; Gollwitzer, 1993, 1999)a process we refer to as
strategic automaticity (Gollwitzer & Schaal, 1998). We will argue that humans can willfully exert
automatic action control by an anticipatory process of consciously linking a goal-directed
response to an anticipated situation (i.e., if-then planning).
The chapter is divided into two sections concerned with objective agency and a third section
concerned with subjective agency. In the first section, we (p.64) will broadly introduce the
concept of if-then planning and present empirical evidence that action initiation by if-then
planning exhibits features of automaticity (e.g., immediacy, efficiency, and redundancy of
another in situ conscious intent; Bargh, 1989; Shiffrin & Schneider, 1977). Since the early
formulation of a theory of implementation intentions (Gollwitzer, 1993), new developments have
been made in areas of action control and language comprehension. We pick up these new
developments and explicate possible mechanisms behind the strategic automaticity created by
if-then planning in the second section of the chapter. Whereas the first two sections relate to
objective agency, that is, how action is controlled (by planning), in the final section we will
discuss issues related to subjective agency (i.e., the sense of agency). A self-regulation strategy
should not undermine a persons feeling of control, as this feeling provides important
information about ones capabilities and influences action-outcome expectancies (Heckhausen,
2008; Rotter, 1966). Therefore, as we argue that self-regulation by if-then planning leads to
action initiation that exhibits features of automaticity, in the last section of the chapter, we will
discuss how (if at all) these automaticity features affect the sense of agency over the if-then
planned actions.
Overview of Existing Work on Implementation Intentions

Goal Intentions and Implementation Intentions
Most of the time, we cannot immediately implement the behavior that we want or ought to
engage in. For example, a student may sit in class and realize that she has to start working on a
class assignment soon or else she will fail the class requirements. Another person, reading an
article about health issues, may intend to buy an apple instead of a chocolate muffin at the
cafeteria in the afternoon to promote his goal to eat more healthily. These examples illustrate an
important characteristic of intentions: intentions can have different levels of specificity. For
example, the intention to eat more healthy food is a very broad description of what one wants
to do, whereas to buy an apple in the cafeteria is a more specific intention. This specificity
results not only from the obvious reference to a distinct type of healthy food (i.e., goal specificity
in the sense of Locke & Latham, 2006), but to the description of where and when to buy it: at the
Page 2 of 27
cafeteria, in the afternoon. This aspectwhere and when to perform the intended behavioris
at the heart of an important distinction for intentions proposed by Gollwitzer (1993, 1999).
Based on ideas extending back to Narzis Ach (1910) and Kurt Lewin (1926, 1951), Gollwitzer
(1993) differentiated between goal intentions (p.65) and implementation intentions. Whereas
goal intentions simply specify a desired outcome (I want to be healthy) or a desired action (I
want to eat more healthy food), implementation intentions combine a critical anticipated
situation with an intended goal-directed action in the form of If I stand in front of the cafeteria
shelf, then I will grab an apple! Mentally forming a link between a situational cue and a goal-
directed response has been demonstrated to increase the likelihood of actually implementing the
intended behavior in numerous studies (meta-analysis by Gollwitzer & Sheeran, 2006). From
early on, Gollwitzer (1993) proposed two main mechanisms to explain how implementation
intentions achieve these superior outcomes compared to goal intentions. First, the anticipated
critical situation becomes highly accessible, and second, a link is created between the critical
situation and the intended behavior. These mechanisms underlie the notion of strategic
automaticity (Gollwitzer, 1993, 1999; Gollwitzer & Schaal, 1998): an agent intentionally
formulates an if (situation)-then (action) plan geared toward achieving a higher order goal (i.e.,
strategic). The if-then plan results in a perceptual preparedness for the critical situation and a
behavioral readiness to engage in the planned behavior (i.e., automaticity). These propositions
laid the groundwork for a huge research program successfully testing the heightened
accessibility of the critical situation (e.g., Achtziger, Bayer, & Gollwitzer, 2012; Parks-Stamm,
Gollwitzer, & Oettingen, 2007; Webb & Sheeran, 2004, Studies 2 and 3; Wieber & Sassenberg,
2006) and the link between situation and action (Aarts, & Dijksterhuis, 2000; Aarts, Dijksterhuis,
& Midden, 1999; Adriaanse, Gollwitzer, de Ridder, de Wit, & Kroese, 2011; Bayer, Achtziger,
Gollwitzer, & Moskowitz, 2009; Brandsttter, Lengfelder, & Gollwitzer, 2001; Gollwitzer &
Brandsttter, 1997, Study 3; Papies, Aarts, & de Vries, 2009; Webb & Sheeran, 2007; Webb,
Sheeran, & Luszczynska, 2009). Even more important are the numerous applied studies testing
the effectiveness of implementation intentions in helping people to achieve their goals, as
reviewed in meta-analyses on eating behavior (Adriaanse et al., 2011) and physical activity
(Blanger-Gravel, Godin, & Amireault, 2013). In the present chapter, we will focus on the second
mechanism of if-then planning, that is, the link between the critical situation and the goal-
directed behavior resulting in action initiation that features characteristics of automaticity (see
Bargh, 1989; Shiffrin & Schneider, 1977): it is fast, efficient, and requires no additional in situ
conscious intent. In the following section, we will present experimental evidence for this
automaticity claim, focusing on evidence pertaining to overt behavior (for other reviews,
including outcomes like emotion suppression or self-affirmation via implementation intentions,
see Gollwitzer & Oettingen, 2011).
(p.66) Empirical Evidence for Strategic Automaticity

Previous overviews of the features of automaticity of if-then planned behavior have structured
the research according to the three features of automaticity (immediacy, efficiency, and no
additional in situ conscious intent; e.g., Gollwitzer & Oettingen, 2011). In line with the general
theme of the present edited volume (i.e., human agency), in the following the evidence is
organized by potential situational impediments for conscious action control. Some situations
may be disadvantageous for conscious, on-line control of behavior because cognitive resources
Page 3 of 27
are deployed elsewhere, temptations undermine intended behavior, the required behavior is
unpleasant, there are conflicting automatic processes, or critical aspects of the planning-to-
action process are not consciously accessible. Finally, we will present evidence from cognitive
neuroscience that complements the behavioral evidence that implementation intentions
intentionally delegate control to the environment (i.e., strategically automate action control).
Cognitive Load: When Resources Are Deployed Elsewhere

Deliberative action control requires cognitive capacity (e.g., Strack & Deutsch, 2004). Thus, the
first evidence that if-then planned behavior is automatically initiated comes from research
showing that the behavior is efficiently and effortlessly initiated, even when cognitive resources
are deployed elsewhere. For example, Brandsttter et al. (2001, Studies 1 & 2) showed that
implementation intentions improved the goal striving (i.e., writing a curriculum vitae) of
psychiatric patients with low cognitive capacities as a result of acute opiate withdrawal
symptoms or acute schizophrenic episodes (Studies 1 & 2). Under more controlled experimental
settings (Study 3 & 4), the authors further demonstrated that the performance of university
students under cognitive load improved when implementation intentions had been formed. More
precisely, participants were asked to work on a dual-task paradigm. The difficulty of the primary
task (to follow a target with the mouse cursor) was manipulated to be either difficult or easy. In
the secondary task, participants were asked to respond as quickly as possible whenever
numbers appeared on the screen, but to withhold their response whenever letters appeared.
Half of the participants formed an implementation intention geared at speeding up their
response to one critical number. Results showed that forming implementation intentions indeed
sped up the responses to the critical cue compared to the non-critical cues (without
compromising the responses to non-critical cues). Importantly, for participants with an
implementation intention, this effect was independent of the task difficulty of the primary task,
thus providing evidence for the prediction that if-then planned action initiation is effortlessness
and efficient (p.67) in terms of required cognitive resources. As we will return to the notion of
effort and if-then planning in regard to the sense of agency, it is important to note that, at this
point, effortlessness does not necessarily refer to perceived effort. The effortlessness of if-then
planned action initiation is deduced from the observation that it can operate successfully even
under conditions of low cognitive resources.
Temptations and the Initiation of Unpleasant Behavior

A particularly difficult situation to exert control involves inhibiting unwanted behaviors that
promise an immediate reward (i.e., temptations; Hofmann, Baumeister, Frster, & Vohs, 2012).
However, implementation intentions have been shown to increase our ability to behave in line
with our goal intentions instead of giving in to temptations. For example, Adriaanse et al. (2010;
Study 1) investigated the behavior of women with the goal of reducing their unhealthy snack
intake. After thinking about the positive effects of their health goal and reflecting on possible
obstacles (i.e., mental contrasting; Oettingen, Pak, & Schnetter, 2001; Oettingen, 2012), they
formulated a personally appropriate implementation intention to reduce unhealthy snacking.
The results showed that the implementation intention group indeed consumed fewer calories
within the following week compared to a control group having the same goal but no related
implementation intention. Thus, implementation intentions helped to overcome the temptation of
snacking in everyday situations.
Page 4 of 27
Not initiating behaviors that promise an immediate reward is one thing, but what about
initiating behaviors that are important but unpleasant? Work by Sheeran and Orbell (2000)
showed that forming implementation intentions increased the percentage of women who
attended a cervical cancer screening from 69% (control group) to 92% (implementation
intention group). Thus, implementation intentions have the potential to initiate actions even
against adverse affective states.
Conflicting Automatic Processes

Another way of demonstrating the automaticity features of if-then planned action initiation is by
showing that they can down-regulate other automatic responses. Evidence for this ability comes
from research on bottom-up triggered spatial congruency paradigms and research on implicit
stereotyping.
Location Congruency Effects
A particularly impressive demonstration of how if-then planned behavior can influence

automatic bottom-up responses is provided by Cohen, Bayer, Jaudas, and Gollwitzer (2006,
Study 2). In this study, participants performed (p.68) a Simon task (Lu & Proctor, 1995) by
responding with a left or right key press to a low or high tone, presented over headphones. The
tone, however, was presented either to the left or right ear. This setup results in spatially
congruent (tone that requires a left button press presented to the left ear) and incongruent (tone
that requires a left button press presented to the right ear) trials. Although the presentation side
of the tone is irrelevant to participants classification responses, there is usually a very robust
congruency effect; left responses to a tone presented to the left ear are faster than left
responses to a tone presented to the right ear (Lu & Proctor, 1995). In the study by Cohen et al.
(2006), participants were given an implementation intention instruction before the task for one
of the incongruent situations (e.g., If I hear a low tone on the right, then I will press the left
button!). For this specific situation and only for this situation, the response times decreased to
the level of congruent trials. Thus, the if-then planned behavior was implemented so quickly that
the influence of the low-level bottom-up location effect was reduced. In a later, more detailed,
investigation on how pre-existing automatic biases and newly formed if-then plans interact,
Miles and Proctor (2008) concluded that if-then plans do not necessarily replace pre-existing
biases but that they are fast enough to match the automatic biases and thus influence resulting
responses in an additive way.
Implicit Stereotypes
In the person perception domain, stereotypes are assumed to be activated in an automatic

fashion. This is important for social interactions, as it implies that conscious control over
stereotype activation and application is problematic (e.g., Payne, 2001). However, in line with
our argument that if-then planning implements automatic control, research has shown that
implementation intentions can be effectively used to prevent stereotyping. Counteracting
automatic stereotype activation via implementation intentions has been demonstrated by
manipulating the mental representation of target categories (e.g., woman; Gollwitzer & Schaal,
1998). Gollwitzer and Schaal showed that if-then planning to ignore the gender of names
resulted in less gender-related stereotype activation. In addition, more recent research has
shown that implementation intentions can reduce the generally observed bias against African
Page 5 of 27
Americans in the so-called shooter paradigm (e.g., Mendoza, Gollwitzer, & Amodio, 2010;
Stewart & Payne, 2008). In this paradigm, participants have to make rapid decisions on whether
a target person is holding a weapon or a tool. Participants are supposed to shoot individuals
with a gun and not shoot those with a tool. Research shows that the shooter paradigm leads to
more erroneous shots of African Americans holding a tool compared to Caucasians holding a
toolpresumably because of the stereotypic association (p.69) between African American
and aggressive. Stewart and Payne (2008) manipulated the mental representation of the target
category (If I see a Black face, then I will think safe!). This manipulation of the concept
African Americans in the direction of reacting with safe rather than the stereotypic
aggressive led to less erroneous shots of African Americans holding a tool. Mendoza et al.
(2010) did not manipulate the representation of the target category but instead provided specific
goal-directed responses (If I see a person with a gun, then I will shoot! and If I see a person
with an object, then I will not shoot!). This second approach also decreased the response bias in
that it resulted in less erroneous shots of African Americans holding a tool. Note that in
general we think that specifying a do not respond in the then-part is problematic, as it may
heighten the activation of the unintended response (Adriaanse et al., 2011). However, in this
particular case, participants had to press one of two buttons labeled Shoot and Dont Shoot.
Thus, do not shoot did not refer to the negation of a response but to the facilitation of pressing
the Dont Shoot button.
Bypassing Conscious Awareness

Finally, the most compelling demonstrations for the automatic nature of if-then planned action
initiation comes from studies in which critical components of the planning-to-action process
were not consciously accessible. For example, Sheeran, Webb, and Gollwitzer (2005) indirectly
manipulated the superordinate goal for which a plan was created. Specifically, in Study 2,
participants formed an implementation intention geared toward speeding up responses in the
Matrix Reasoning Subset of the Wechsler Adult Intelligence Scale. In the task, participants are
presented with an incomplete abstract picture, which they have to complete by choosing from a
number of options. Before the matrix task, the goal to be fast was indirectly activated via a word
puzzle task including words related to speed (e.g., fast, haste, and rapid; speed-goal
activation condition) or words that were neutral regarding speed (e.g., tone, lavish, urban,
polar; control condition). The results showed that the speed-related implementation intention
was effective only if the speed goal was activated. If the superordinate speed goal was not in
place, the implementation intention did not have any effect. As the participants were not
explicitly aware of the activation (or non-activation) of the goal, it is unlikely that conscious
processes modulated the effectiveness of the implementation intentions. We will return to the
issue of why an overarching goal is important for the effectiveness of if-then planned behavior in
the second section, when explicating the mechanisms of implementation intentions.
Undoubtedly, a further critical component of if-then planned action initiation is encountering the
specified situation. If the action initiation is indeed (p.70) automatic (i.e., not requiring another
in situ conscious intent), then the action should be triggered even when the situation is
presented subliminally (i.e., below the threshold of conscious awareness). Bayer et al. (2009)
provided evidence for this assumption in two experiments. In one study (Bayer et al., Study 1),
the activation of behavior-related concepts through the subliminal presentation of the critical
Page 6 of 27
situation was investigated. Participants formed an implementation intention to complain about a

rude experimenter. When this experimenters face was presented subliminally in a word
pronunciation task, concepts (e.g., rude, cheeky, conceited) instrumental to the behavior (i.e.,
complaining) were vocalized faster compared to control words. However, this facilitation was
only observed for participants who had formed an implementation intention with the rude
experimenter as a critical cue and complaining as the intended responsenot for participants
who simply held the goal to complain. Thus, this experiment provides evidence that the critical
cue directly elicited the activation of concepts related to the intended behavior even outside
conscious awareness.
In a second study, Bayer et al. (Study 2) provided more direct evidence that the intended
behavior is elicited automatically (as compared to the related concepts activated in Study 1).
Participants categorized angular and round forms (i.e., by pressing a left or right key,
respectively). One angular form, a triangle, was included in an implementation intention that
read: If I see a triangle, then I will press the left key particularly fast! In this categorization
task, a subliminally presented prime shape preceded the target shape. Response times to the
target shapes revealed a speed-up effect for categorizing angular shapes following a triangular
prime compared to responses to round shapes and responses to angular shapes made after
neutral primes (e.g., shapes not specified in the implementation intention). The authors argued
that the subliminal prime activated the intended response, leading to a faster response if the to-
be-performed response was congruent (e.g., left for triangles and other angular shapes). The
results in both studies were observed only when participants formed an if-then plan prior to the
task and not if the plan was formulated as a goal intention (including all critical information but
not in an if-then format). Most important, as the critical cues were presented subliminally, the
activation of behavior-related concepts (Study 1) and the initiation of the behavior itself (Study
2) could not have been due to conscious control.
First Neuroscientific Support for Strategic Automaticity

We will close this section with one final piece of evidence from cognitive neuroscience that
supports the automaticity claim of implementation intention research using brain-imaging
techniques. Gilbert, Gollwitzer, Cohen, (p.71) Oettingen, and Burgess (2009) compared goal-
driven action initiation with implementation intention-driven action initiation. In line with
evidence that activity at the lateral Brodmann area 10 (lateral BA 10) is associated with action
initiation in a top-down fashion, whereas medial BA 10 activity is associated with more bottom-
up driven behavior (reviewed by Burgess et al., 2008; West, 2008), Gilbert and colleagues found
heightened activity in the lateral BA 10 for goal-driven behavior but heightened activity in the
medial BA 10 for implementation intention-driven behavior. Thus, this evidence from cognitive
neuroscience fits with the behavioral evidence and the assumptions underlying implementation
intentions that if-then planning delegates a persons action control to situational cues.
Summary
The presented overview of research highlights the notion of strategic automaticity implemented
by if-then planning. Careful if-then planning of what to do in critical situations has been shown
to be effective in initiating intended behavior in situations that challenge conscious control. We
have provided evidence that implementation intentions are effective in situations in which
cognitive resources are deployed elsewhere, temptations facilitate unintended behavior, the
Page 7 of 27
behavior to be initiated is unpleasant, there are conflicting automatic processes, and critical
aspects of the planning-to-action process are not accessible to conscious awareness.
Further Explication of Implementation Intention Mechanisms

As shown in the empirical review above, implementation intentions are very simple plans with
surprisingly strong effects. The principles proposed by Gollwitzer (1993, 1999), increased cue
accessibility and automatic action initiation, provided a theory that instigated a lot of research
testing both these process assumptions (summary by Gollwitzer & Oettingen, 2001), and the
effectiveness of implementation intentions in various applied settings (reviewed by Adriaanse et
al., 2011; Blanger-Gravel et al., 2013; Gollwitzer & Sheeran, 2006). However, further analysis
of the psychological mechanisms underlying the situation-behavior link seems warranted. How
exactly is the verbal if-then plan translated into real action? What is the nature of the mental
representation of the situation and the action created by if-then planning? And does the nature
of the representation provide insights into why if-then planning is so effective? In this section,
we will propose answers to these questions and further explicate mechanisms of if-then planned
action initiation.
(p.72) To set the stage, we will compare action control by implementation intentions to
habitual action control. Both share certain characteristics: as with implementation intentions,
habitual behavior is immediate, efficient, and can occur outside awareness (Aarts &
Dijksterhuis, 2000; Verplanken & Aarts, 1999; Wood & Neal, 2007). However, habits are created
differently from implementation intentions (e.g., Sheeran, Webb, & Gollwitzer, 2005). Habitual
associations between situations and behaviors are created by the repeated co-occurrence of
certain situations and executing certain responses. How can the similarities (in terms of action
execution) between implementation intentions and habitual behavior be explained, when
forming an implementation intention involves neither the actual perception of the situational
cue, nor the execution of the realrepeatedly performedaction? Let us take a closer look at
the following example of an implementation intention: If I stand in front of the cafeteria shelf,
then I will grab an apple! Formulated under the umbrella of a corresponding superordinate
goal (e.g., to eat more healthy food), this plan increases the likelihood of actually grabbing an
apple when standing in front of the cafeteria shelf, compared to a mere goal intention (e.g., I
want to eat more apples!). Basically, the two components of the plan, the if-part and the then-
part, are only verbal descriptions of a situation (cafeteria shelf) and an action (grab an
apple). Thus, the question that must be addressed is how verbal self-instructions can achieve
what in habit formation is done through repeated co-occurrence of a real situation and a real
action. Our answer is based on what we refer to as the pattern-overlap principle. We propose
that planning effects (i.e., successful action initiation upon perception of the critical cue) are
successful to the degree that the activation patterns at the time of planning and the necessary
activity pattern to initiate the action upon encountering the critical cue overlap. This overlap
includes similarity between the activation patterns in form (i.e., similarity between what one is
thinking [i.e., simulating] and what one is encountering and intending to act) and location (i.e.,
what brain resources the mental representations draw on). We assume a specific planning event
can be mapped to a certain point in a continuum of no overlap at all to a perfect overlap with the
necessary activity to initiate the action. We expect that the formation of implementation
intentions is an ideal form of planning that results in a comparatively high overlap because the
specific form of the implementation intention activates important aspects (e.g., situation and
Page 8 of 27
action, respecting the causal order) that are not activated in more mundane plans. The
components described in the next sub-section contribute to the activation pattern at the time of
planning and action initiation.
(p.73) Five Components of (If-Then) Planned Action Execution

We propose four components that, when activated simultaneously, lead to the automatic
initiation of the if-then planned behavior, and a fifth component overseeing the execution of the
behavior. The first component (A) is the superordinate goal, which provides the context for the
stimulus-response association. The second component (B) is the representation of the critical
situation, which is the link and the trigger for the following two components: the direct priming
of motor components (C) and the activation of the anticipated behavioral outcome (D), which
together initiate the intended action either directly (motor priming) or by action-effect principles
(anticipated outcome). The last component (E) includes processes that guide and adjust the
behavior according to the current environment (e.g., location of target objects). We will address
each component separately and outline how they may fulfill their function in the execution of (if-
then) planned behavior.
Goal State
If-then plans are formed in the service of a superordinate goal. The goal provides the internal
environment, or context, in which the action planning and execution take place. The active goal
is one important feature contributing to the activity pattern that is necessary for the action
initiation when the critical situation is encountered. If-then planned action initiation is thus
conditional, that is, it depends on the unique context provided by the active goal (e.g., Sheeran
et al., 2005). This unique context is shaped by different aspects, including goal commitment, as
well as the desirability and feasibility of reaching the goal.
The conditional automaticity associated with if-then planning has analogies in other areas of
psychology. For example, there is increasing evidence for conditional automaticity in attitude
activation. Implicit attitude measures (assumed to measure automatic attitude activation) show
that an African-American person in the context (i.e., environment) of a church automatically
activates a different attitude than an African-American person in the context of a street corner
(Dasgupta & Greenwald, 2001; Wittenbrink, Judd, & Park, 2001). As the environmental context
(church vs. street corner) determines the automatic link between African-American faces and
attitudes, we assume that a superordinate goal can similarly provide the (internal) environment
for the automatic initiation of if-then planned behavior, creating context (i.e., goal-) dependent
automaticity (Bargh, 1989). Thus, for implementation intentions to influence behavior (i.e.,
automatic action initiation), the same goal must be active when the critical situation is
encountered as when the plan was formed. The active (p.74) goal contributes to the general
state (internal environment) that leads to the automatic action initiation.
As detailed previously, implementation intention effects are seen only when a respective goal
(i.e., speed) is activated and in line with the plan (Sheeran et al., 2005). This provides evidence
for our assumption that implementation intentions offer a kind of goal-dependent automaticity,
as well as support for the pattern-overlap principle as decreased similarity (presence vs.
absence of goal) between the mental state at the time of planning and the time of cue perception
eliminated implementation intention effects.
Page 9 of 27
Perceptual Simulation of the Critical Situation (If-Part)

With the superordinate goal active, the first step in forming an implementation intention is to
anticipate an adequate situation and formulate this situation in the if-part of the plan. For
example, with the goal of eating more healthy food, one could anticipate that a good
opportunity to eat something healthy is the reoccurring confrontation with the cafeteria shelf in
ones university or company. Thus, one may specify: If I stand in front of the cafeteria shelf
The complexity of specifying an adequate situation is explained elsewhere (Gollwitzer, Wieber,
Myers, & McCrea, 2010). Here we will concentrate on how a verbal formulation of a situation
(e.g., in front of the cafeteria shelf) may be sufficient to provide a mental representation that is
likely to be reactivated upon contact with the real situation, thereby triggering the processes
that set the action in motion. To bridge the gap between language and perception (and
subsequent action), recent theories of language comprehension (reviewed by Barsalou, 2003;
Glenberg, 2007; Zwaan, 2004) provide intriguing insights that may advance our understanding
of how a mental act can link a behavioral response to a situation.
In line with general simulation theories of cognition (Barsalou, 1999, 2008; Kiefer &
Pulvermller, 2012), recent theories of language comprehension assume that comprehending
verbal content relies on the re-enactment of analog sensorimotor experiences (i.e., simulations;
e.g., Glenberg & Kaschak, 2002; Glenberg & Robertson, 1999; Kaschak & Glenberg, 2000;
Stanfield & Zwaan, 2001; Zwaan, Stanfield, & Yaxley, 2002). In contrast to traditional theories
of mental representations as abstract symbols representing feature lists, semantic networks,
and frames, simulation theories assume that a mental representation of an object is a re-
enactment of sensorimotor experiences associated with the real object (Barsalou, 2003). Thus,
reading verbal material activates simulations of the read content in both perceptual and motor
areas. For example, reading about an eagle in the sky has been shown to activate a mental
representation of an eagle with (p.75) outstretched wings, whereas reading about an eagle in
the nest does not. Pure analyses based on syntax and semantics do not necessarily predict this
differentiationhowever, a model that includes re-enactments of prior perceptual experiences
with these two situations certainly would (Stanfield & Zwaan, 2001; Zwaan et al., 2002). That
these simulations indeed recruit perceptual brain areas is supported by neuroscientific research
showing, for example, that reading words that refer to a smell (e.g., cinnamon) activate primary
olfactory areas (Gonzlez et al., 2006) and seeing food activates gustatory processing areas
(Simmons, Martin, & Barsalou, 2005; see also the sub-section below on motor simulations).
What does this mean for the formation of the if-part of an implementation intention? The above-
described research suggests that although a real situation and a verbal description of the same
situation may intuitively seem different, what is going on in our brain may not be so different at
all. The critical situation serves as a link and trigger for the action initiation. If reading about (or
thinking about) the critical situation activates a perceptual simulation of the situation, this
perceptual simulation should overlap with the perceptual activity triggered on contact with the
real situation. Thus, instead of the need to assume questionable translation processes that map
perceptual states to abstract mental representations and re-map these abstract representations
to perceptual states (i.e., transduction and symbol grounding problem; Barsalou, 1999; Searle,
1980), the perceptual state of the critical situation itself becomes the link and trigger for the
action initiation. This perspective can explain why no additional in situ conscious intent is
Page 10 of 27
necessary for the if-then planned action initiation (i.e., why conscious recognition of the
situation is not necessary; Bayer et al., 2009). The efficiency of this account becomes evident
when considering the various translation processes required by an explanation based on more
traditional accounts of mental representations (e.g., one translation from verbal content to
abstract representation during plan formation, another translation from perception to abstract
representation on encountering the cue, and additional translation processes in the action
component of the if-then plan).
Motor Simulation (Then-Part 1)

But how are actions cognitively represented in the then-part of implementation intentions? We
propose two components that become linked to the critical situation that hand in hand lead to
the initiation of the intended behavior: the priming of motor components (motor simulations)
and the simulation of the anticipated behavioral outcome. As with the if-part, the then-part of an
implementation intention is a verbal description, this time (p.76) of a behavior (e.g., grab an
apple). Simulation theories of cognition are not limited to sensory perceptual simulations.
Motor simulations also play an important role in understanding verbal descriptions of behavior
(e.g., Glenberg & Kaschak, 2002; Zwaan & Taylor, 2006; cf. Jeannerod, 2001). William James
(1890, vol. II, p. 526) nicely stated this idea: Every representation of a movement awakens in
some degree the actual movement which is its object. Forming a verbal description of an action
(as is the case in the then-part of an implementation intention) would thus to some degree
activate the actual movement. This old idea finds empirical support in modern research. For
example, participants in a study by Glenberg and Kaschak (2002) read sentences that included
certain arm movements and reported whether the sentence made sense or not by button presses
that required moving their arm away or toward their body. Response times were faster for
button presses that matched the movement included in the respective sentence (e.g., a sentence
about closing a drawer was associated with faster responses to buttons that required them to
extend their arm away from the body). This supports the assumption that processing verbal
material about a motor movement includes the enactment of analog motor simulations that
overlap with the activity patterns associated with actually performing the movement. This view
is also supported by neuroscience studies. For example, event-related fMRI measurements of
brain activity in frontocentral motor regions showed partially overlapping activation patterns
(according to the somatotopic organization of the area) when participants performed specific
movements and passively read corresponding words (e.g., moving their foot and reading the
word kick; Hauk, Johnsrude, & Pulvermller, 2004; Pulvermller, Hauk, Nikulin, & Ilmoniemi,
2005).
Similar to the arguments made above for the mental representation of the critical situation, the
presented research suggests that processing verbal descriptions of a behavior activates analog
simulations in brain areas also involved in actually performing the behavior. Thus, when forming
an implementation intention, a perceptual simulation of the critical situation is activated and
specific motor simulations (covert actions) that reflect specific components of the intended
behavior may thus be linked to the perceptual if-part simulation. With the above plan to grab an
apple, these components could include simulations of arm extension and adjustments of
handgrip size to an ordinary apple. Thus, encountering the critical situation will pre-activate
specific motor components and thereby directly prime the intended behavior. We propose that
Page 11 of 27
this motor priming constitutes another important component that leads eventually to the
effective and effortless action initiation of implementation intentions.
(p.77) Simulation of Anticipated Behavioral Outcome (Then-Part 2)

The motor component explained above may constitute an important function for very concretely
formulated then-parts such as then I will press the right button! or then I will grab an apple!
However, implementation intentions sometimes spell out rather abstract behaviors (e.g.,
Achtziger, Gollwitzer, & Sheeran, 2008, Study 2). Even in the case of then I will grab an apple!
it is unlikely that the primed motor components sufficiently reflect the complexity of the
behavior in varying situations. We propose that another component is important, namely the
anticipation of the behavior outcome. Regardless of how abstractly the then-part of an
implementation intention is formulated, it always implies a certain behavioral outcome, for
example, holding an apple in ones hand. (Note that by anticipated behavioral outcome, we mean
the immediate behavioral outcome of the action specified in the then-part of the plan [e.g.,
holding the apple in ones hand] and not the superordinate goal [eat healthy; but see Wieber,
Sezer, & Gollwitzer, 2014, for implementation intentions including the superordinate goal
attached to the then-component].) Do we represent these behavioral outcomes when forming the
then-part of an implementation intention? If so, this anticipated outcome should also become
associated with the critical situation. Simulation accounts of cognition (e.g., Barsalou, 1999) are
dynamic, meaning that they also include temporal aspects (e.g., Decety, Jeannerod, & Prablanc,
1989; Frak, Paulignan, & Jeannerod, 2001) and outcomes of simulated actions (i.e., the
anticipation assumption, Hesslow, 2002). Thus, processing the verbal description of grabbing
an apple may not only be represented as a static concept (i.e., grabbing and apple) but also
as the movement toward the apple, the adjustment of the handgrip (see the following sub-
section), and also the outcome: the apple in ones hand. Some research in the area of language
comprehension indeed suggests that outcomes of actions are activated when processing verbal
descriptions of behaviors (Horton & Rapp, 2003). From this evidence we deduce that when
forming the then-part of an implementation intention, in addition to any implied low-level motor
simulations, the mental representation includes a simulation of the anticipated behavioral
outcome as well. This outcome simulation, by simple Hebbian mechanisms, is then also
associated with the critical situation and may become re-enacted on encountering the situation.
But how does the anticipated behavior outcome contribute to the initiation of the intended
action? Historical and recent theories of action control in fact state that actions are represented
by their outcomes (Lotze, 1852; ideo-motor principle, James, 1890; action-effect principle, e.g.,
Hommel, 1993; Elsner & Hommel, 2001; Shin, Proctor, & Capaldi, 2010). After the contingency
is learned that a certain action will result in a certain outcome, the activation of the outcome
can initiate the respective action. Experimental evidence supports (p.78) these assumptions
(reviewed by Prinz, 1997; see also Hommel, Chapter 14 of this volume). This action-effect
principle may contribute to the effortless action initiation observed for implementation
intentions. If the critical situation is encountered and the anticipated outcome representation is
triggered, the activation of the outcome may trigger necessary actions to achieve the outcome.
Thus, the action-effect principle provides an explanation of how even rather abstract then-parts
of an implementation intention (which cannot be adequately represented by low-level motor
simulations at the time of planning) are still able to automatically initiate the intended behavior.
Page 12 of 27
We assume that both the low-level motor simulations and the activation of the anticipated
behavioral outcome complement each other in activating the intended action.
On-line Guidance
The final of our five components of if-then planned action control is not directly concerned with
the planning itself; however, it still must be considered. One cannot always anticipate each
aspect of the critical situation. For instance, the location of the apples in the cafeteria shelf in
relation to ones body will never be the same and cannot be perfectly anticipated during the
planning phase. Fortunately, this is not necessary. The only requirement is that the critical
situation is reasonably similar (i.e., there will be apples available). If the previously discussed
components (motor simulations and anticipation of the behavioral outcome) successfully initiate
the response, other processes achieve the guidance of the behavior to its completion. Action-
perception comparisons seem to adjust action control to current environmental circumstances
(e.g., Frith, Blakemore, & Wolpert, 2000). It has been shown that location changes of target
objects are immediately corrected for, even if participants are unaware of the change (Castiello,
Paulignan, & Jeannerod, 1991). Other research on so-called affordances (Grezes, Tucker,
Armony, Ellis, & Passingham, 2003; Tucker & Ellis, 1998; reviewed by Ellis, 2009) demonstrates
that the perception of objects leads to automatic adjustment of low-level motor aspects such as
trajectory of hand movements, handgrip size, and hand orientation (Ellis & Tucker, 2000;
Goodale, Pelisson, & Prablanc, 1986). Together, this research shows that planned behavior only
needs to be initiated; the on-line control of the details are taken care of by our perceptual-motor
system with its years of experience in guiding our body in the environment to produce intended
outcomes.
Summary
Plans are made to achieve a respective goal. Thus, while forming an implementation intention,
the superordinate goal is activated. In the planning (p.79) phase, low-level motor simulations
and the simulation of the intended outcome (then-part) are co-activated with (and thus linked to)
the perceptual simulation of the critical situation (if-part). At the time of action initiation, the
perceptual activity instigated on contact with the critical situation will reactivate the low-level
motor simulations and the intended outcome and thereby prime the intended action. Thus, even
the verbal formulation of an if-then plan co-activates and wires (Hebb, 1949) the necessary
perceptual and motor circuits in the brain for an environmentally controlled action initiation.
The proposed mechanisms provide starting points for further research. For example, the
pattern-overlap principle can be used to predict the effectiveness of if-then planning (and
planning efficiency in general) as it indicates how the critical situation and the intended action
must be specified in line with variables such as the individuals experience level (in regard to the
relevant action domain) or familiarity with the to-be-encountered environment.
The Sense of Agency in If-Then Planned Behavior

The first two sections of this chapter were concerned with our perspective on objective agency,
that is, how an agent is able to control his or her actions. Our answer is that one way of
consciously controlling actions is by planning them in an if-then format. Now we will turn to the
subjective aspect of agency. Feelings or judgments of being in control of ones behavior and
certain aspects of the environment have motivational consequences by informing the agent
about what he or she is capable of doing and influencing contingency judgments about what
Page 13 of 27
actions will bring about which outcomes (Heckhausen, 1989; Rotter, 1966). A self-regulation
strategy should not undermine this feeling of control. Thus, in this last section of the chapter, we
will discuss processes that may be important in creating a feeling and judgment of agency in
regard to the strategic automaticity of if-then planned actions.
Self-Efficacy versus Sense of Agency

Before discussing the sense of agency, we need to distinguish it from a prominent concept in the
realm of motivation psychology: self-efficacy (Bandura, 1982; Schunk & Pajares, 2009). Before
one executes an action and has a sense of having caused the action (i.e., sense of agency),
people have a belief about how well they are able to execute certain actions to deal with
upcoming challenges. These self-efficacy beliefs have been shown to affect choices of activities,
effort expended, persistence, interest, and achievement (Bandura, 1977; Bandura & Schunk,
1981; Pajares, 1996, 1997; Schunk, 1995).
(p.80) Planning the necessary actions to achieve a certain goal (i.e., if-then planning) could
increase self-efficacy and thus goal attainment by the mere confidence gained through the
specification of the necessary steps. However, we expect that the cognitive processes described
in the previous section of this chapter are responsible for if-then planning effects, rather than
factors that influence motivation such as self-efficacy. The results of a meta-analysis (Webb &
Sheeran, 2008, Study 1) are in line with this argument concerning self-efficacy. Self-efficacy
alone could not explain the effects of if-then planning on goal attainment, and the analyses
indicated that if-then planning effects are mostly not driven by factors that influence motivation
(e.g., self-efficacy). However, it may very well be that successful goal attainment (as a
consequence of if-then planning) does increase self-efficacy beliefs. This, however, concerns the
effects of successful goal striving (with or without implementation intentions) on subsequent
goal striving and is a separate question from whether self-efficacy is responsible for planning
effects.
In contrast to the anticipatory belief that one is able to execute certain actions to deal with a
future situation (self-efficacy), the sense of agency has been defined as the sense that I am the
one who is causing or generating an action (Gallagher, 2000). Something that complicates the
analysis of agency experience is that we seldom have an intense feeling of being the agent of an
action, but we certainly feel or become aware of failures of agency (cf. Chambon & Haggard,
2013). This aspect is interesting from a self-regulation perspective. When knowledge (e.g.,
chocolate muffins contain many unhealthy ingredients) in combination with a specific behavior
(e.g., I am eating a chocolate muffin every afternoon) is in conflict with certain goals (e.g., eating
healthy food), we are likely to become aware of this action and our potential role in this
behavior. This may get us started with attempts to self-regulate our behavior. The result of this
self-regulation effort (i.e., reflecting about what I will buy in the cafeteria the next time) is
probably also under heightened scrutiny. Thus, self-regulation processes may highlight certain
aspects of the sense of agency, and if self-regulation is effortful, this feeling of effort may itself
increase to the sense of agency for behaviors related to the self-regulation process (Demanet,
Muhle-Karbe, Lynn, Blotenberg, & Brass, 2013). Provided that the underlying goals for a plan do
not change, self-regulation (by if-then planning) can have two outcomes: either the planned
(intended) behavior is successfully initiated, or the planned behavior is not initiated and some
other, unintended behavior is executed. We will now discuss if-then planning and the sense of
Page 14 of 27
agency with the following two questions in mind. First, are there reasons to believe that the
strategic automaticity implemented by if-then planning undermines the sense of agency for the
planned behaviors? Second, what might the differences be between (p.81) intended actions
(successful self-regulation) and unintended actions (failed self-regulation) concerning the sense
of agency?
Anticipation-Outcome Comparisons
There are multiple processes contributing to the phenomenological experience of being the
agent of ones actions (Gallagher, 2012). The major source of agency information seems to be
derived from comparison processes between anticipated behavioral outcomes and sensory
feedback from actual behavioral outcomes. In general, mismatches (beyond a certain tolerance)
between anticipation and actual outcome decrease the sense of agency (for a critical review, see
Synofzik, Vosgerau, & Newen, 2008; see also Synofzik, Chapter 13 of this volume). The
anticipation component of this comparison can have different sources that we will discuss in the
following.
Low-Level Forward Simulations

Motor signals from the brain are assumed to feed into a simulation mechanism that allows
predictions of the signals consequences (and thus corrections) even before actual perceptual
feedback is available (Frith et al., 2000). This means that when we initiate an action, signals
from the brain are not only sent to the involved muscles, but a simulation mechanism is
triggered that simulates the consequences of these motor signals and helps us to constantly
monitor the success of our movement. Experiments with self versus externally induced
movements indicate that these motor signals and presumably the forward simulation are an
important contributor to the sense of agency (Engbert, Wohlschlaeger, & Haggard, 2008;
Moore, Wegner, & Haggard, 2009; Sato, 2009). As both successful and unsuccessful if-then
planned behaviors are initiated by the organism (and not an external force), they are of course
also based on motor signals from the brain. The information that contributes to ones sense of
agency should be present, just as it is present for non-planned voluntary actions. Thus,
regarding this specific component, there is no reason to assume that the sense of agency is
impaired or otherwise affected for successfully implemented if-then planned actions.
Mental Representation of Action Consequences

Voluntary action is usually preceded by an intention that includes information about the action
outcome. This (conscious or subconscious) mental representation of the outcome prior to the
action is assumed to be a second contributor to the sense of agency. Experiments demonstrated
that both supraliminal and subliminal priming of such representations prior to an action
increase explicit judgments of agency (Aarts, Custers, & Wegner, 2005), as well as (p.82)
implicit measures of the sense of agency (e.g., Haggard, Clark, & Kalogeras, 2002; Moore,
Wegner, & Haggard, 2009).
We proposed that at least two processes are directly involved in the initiation of if-then planned
behaviors (see sub-section above): direct motor priming (i.e., action initiation by direct cue-
behavior associations) and priming of the intended behavioral outcome (i.e., action initiation by
action-effect principles). We thus assume that both processes will contribute to the action
initiation, and the complexity of the intended behavior will influence which process contributes
Page 15 of 27
more. The greater the contribution of cue-initiated motor components, the less may the action
initiation dependent on the action-effect principle (cue-initiated activation of the intended
outcome). Thus, as there is not necessarily the need for another in situ conscious intention,
regarding this specific component, a decrease in the sense of agency may be expected. This
decrease could probably be moderated by the degree to which the action is indeed initiated
without another in situ conscious intention. Note that the argument that implementation
intention-initiated actions do not need another in situ conscious intention does not mean that this
conscious intention is necessarily always absent.
For unintended behaviorswhere unintended refers to the undesired behavior that started the
self-regulation process (in our example, eating chocolate muffins)the case is more
complicated. We will focus on two reasons that the if-then planned behavior may have failed. If
the unintended behavior was executed with no thought, that is, purely habitually, there may
not have been a mental representation of the action outcome prior to the action (see Wood &
Neal, 2007, for a purely cue-motor response account of habits; but see Aarts & Dijksterhuis,
2000, for a habit view that involves goals and thus representations of behavioral outcomes).
Thus, this component may not contribute to a sense of agency for failed self-regulation
behaviors, behaviors that failed because of pure habit: the person who just grabbed the
chocolate muffin out of pure habit (i.e., without an explicit conscious intention) and in light of
having explicitly planned to grab an apple may be left with the feeling that this action was not
initiated by the self.
However, the sense of agency regarding the action of grabbing the chocolate muffin may be
different if the initially unintended action was not habitually initiated, but was undertaken
because one could not resist the temptation. An all-too-vivid representation of the behavioral
consequences (e.g., the delicious taste of the chocolate muffin) may override the initial intention
to grab the apple and make one reach out for the chocolate muffin. Thus, in this case, a mental
representation of the behavioral outcome is present and could make the initially unintended
behavior seem intended, as the actor is aware of the desire for the chocolate muffin. This brings
us directly to the final aspect, a reconstruction of agency after the action is executed.
(p.83) Hindsight Reconstruction and Integration into Higher Order Goals

The construction of a sense of agency in hindsight (e.g., Wegner & Wheatley, 1999; reflective
judgment of agency; Synofzik et al., 2008; higher-order sense of agency; Gallagher, 2012) is an
attributional process that interprets the perceived action and its consequences in light of ones
beliefs and goals. The very nature of if-then planning is to find responses that are not only in line
with ones current goals but that explicitly facilitate these goals. Thus, after executing the
planned action, one is usually very aware of the planning episode in which he or she explicitly
planned to do the action in the specified situation because of its anticipated positive effect on a
currently important goal. In this sense, there is no reason to believe that if-then planning should
impair this attributional processif anything, the explicit planning may increase the likelihood
that the successfully implemented action is attributed to the self.
In the case of undesired behaviors, one actually did not initiate the planned behavior (e.g., grab
an apple) that was supposed to facilitate ones goal (e.g., eat healthily). This apparent mismatch
may decrease the sense of agency and, importantly, could be a signal for the agent that ones
Page 16 of 27
goal achievement is threatened. It may thus emphasize the need to put more thought into a
more effective new plan (e.g., to buy an apple on the way to work to avoid the troublesome
situation of the cafeteria). However, that may be an overly optimistic assumption. Unfortunately
(in this case), humans ability for attributions is very flexible. There are many attributions
possible that allow one to avoid admitting to failure, from thinking that today was a special day
to totally questioning ones health goals because of the observation of ones own behavior (cf.
Bem, 1972). This highlights an important aspect of the process of behavioral change. Action
plans in the form of if-then plans are one important part of behavior regulation. However,
dealing with possible failures may be another important aspect that should not be neglected.
Maybe effective behavioral change needs at least one action plan (to initiate the intended
behavior) and one backup plan that specifies how to deal with a possible failure (e.g.,
prioritizing ones goals or not making self-serving attributions).
Summary
Positive comparisons between anticipated behavioral outcomes and sensory feedback from
actual behavioral outcomes seem to provide us with a sense of agency. The sources of the
anticipated outcome can vary from very low-level forward simulations of motor signals to
conscious or subconscious mental representations of the intended outcome to very high-level
attributions of action outcomes based on ones general belief system. On the lowest level,
namely the (p.84) forward simulations of motor signals, we do not expect differences between
non-planned and if-then planning-initiated actions for the sense of agency, as the same low-level
mechanisms are at work in both cases. At the medium level, conscious or subconscious mental
representations of action outcomes, non-planned and if-then planned actions may indeed vary in
their sense of agency. As if-then planned action initiation is expected to rely to a significant
degree on direct motor primingreducing the need for an explicit mental representation of the
action outcomewe expect the sense of agency to be reduced compared to non-planned
voluntary actions that rely on the mental representation of the action outcome. On the highest
level, however, this may switch, and if-then planned actions may lead to a stronger sense of
agency. As planning is usually a conscious process with attention devoted to ones goals and
actions to achieve the goal, having performed such a planned action will very likely result in self-
serving attributions of being in control. Thus a higher sense of agency will emerge compared to
a non-planned voluntary action that had no planning history.
Given these differences in medium- and high-level factors that contribute to the sense of agency,
the interesting question arises of whether the differences may be captured by different
measures of the sense of agency. Whereas more low-level measures (i.e., implicit measures) may
capture a decrease in the sense of agency for if-then planned actions, higher-level measures
(e.g., explicit judgments) may not differ, or we may even find a stronger sense of agency for if-
then planned actions.
Finally, perceived effort has been shown to contribute to the sense of agency (Demanet et al.,
2013). If-then planned action initiation has been shown to be effortless (i.e., in the sense of
operating even with cognitive load). Thus, aside from the predictions made from the
comparative models discussed earlier, if-then planned actions should lead to a lower sense of
agency compared to non-planned voluntary actions. An interesting question would then be how
Page 17 of 27
the feeling of effort interacts with the information from the comparative models and on what
level (implicit or explicit) the effort information influences the implicit or explicit judgment.
Conclusion
In the present chapter, we have discussed the question of objective and subjective agency from
a self-regulation perspective. We have outlined that implementation intentions have been a
fruitful area of research, as they provide a strategy for humans to regulate their behavior
according to their own goals. In the first section of our chapter, we provided an overview of
empirical research supporting the idea that by using implementation intentions one (p.85) can
strategically automate ones future actions. We provided evidence that the actual action
initiation is fast, efficient, and does not require another in situ conscious intent. This conscious,
intentional planning and the subsequent automatic action initiation are what we refer to as
strategic automaticity.
In the second section of the chapter, we focused on the processes that may underlie if-then
planned action initiation by integrating new developments in research on action control and
language comprehension into implementation intention theory. We proposed five components
relevant to the translation of verbal self-instructions into action. The superordinate goal
provides the context in which the link between the critical situation and action is active.
Furthermore, we proposed that this link is represented by sensorimotor simulation processes,
connecting perceptual simulations of the if-part to motor simulations and simulations of the
anticipated behavioral outcome of the then-part. Finally, automatic processes of on-line
guidance oversee the execution of the initiated intended action. The theoretical explication of
these processes provides a rich basis for future research on planning and action control in
general, and on how to maximize the effectiveness of if-then planning in particular.
In the third section we evaluated what the consequences of our notion of strategic automaticity
might be for the sense of agency regarding if-then planned actions. Our preliminary conclusion
is that if-then planning does not in general impair the sense of agency. However, different
factors contributing to the sense of agency may be affected differently with the consequence
that different measures of the sense of agency are affected differently. These propositions are
empirically unexplored, but their investigation seems necessary to arrive at a full understanding
of self-regulation by if-then planning.
To conclude, even after roughly 20 years of research on if-then planning, there are important
grounds still to be explored. In the current chapter we focused on objective agency by further
explicating the mechanisms of action control by if-then planning and its relation to ones
subjective sense of agency. The scope of research to be investigated in regard to self-regulation
via if-then planning is enormous in our eyes because if-then plans may not simply be one self-
regulation strategy, but (verbal) thinking in if-then formats about future situations and actions
may be a fundamental mechanism of human action control (i.e., not solely as they are used
strategically but as they occur in our natural thinking about the future). In that sense, we have
focused on implementation intentions as an ideal type of planning. The delineated mechanisms
may not only become fruitful in the ultimate endeavor of psychological researchthe prediction
of behaviorbut the notion of strategic automaticity may also provide insights into the more (p.
Page 18 of 27
86) basic philosophical questions of how it is possible that immaterial thoughts can propel our
physical bodies.
References
Aarts, H., Custers, R., & Wegner, D. M. (2005). On the inference of personal authorship:
enhancing experienced agency by priming effect information. Consciousness and Cognition, 14,
439458. doi: 10.1016/j.concog.2004.11.001
Aarts, H., & Dijksterhuis, A. (2000). Habits as knowledge structures: automaticity in goal-
directed behavior. Journal of Personality and Social Psychology, 78, 5363. doi:
10.1037/0022-3514.78.1.53
Aarts, H., Dijksterhuis, A., & Midden, C. (1999). To plan or not to plan? Goal achievement or
interrupting the performance of mundane behaviors. European Journal of Social Psychology, 29,
971979.
Ach, N. (1910). ber den Willen (Vol. 1). Leipzig: Quelle & Meyer.
Achtziger, A., Bayer, U. C., & Gollwitzer, P. M. (2012). Committing to implementation intentions:
attention and memory effects for selected situational cues. Motivation and Emotion, 36, 287
300. doi: 10.1007/s11031-011-9261-6
Achtziger, A., Gollwitzer, P. M., & Sheeran, P. (2008). Implementation intentions and shielding
goal striving from unwanted thoughts and feelings. Personality and Social Psychology Bulletin,
34, 381393. doi: 10.1177/0146167207311201
Adriaanse, M. A., Gollwitzer, P. M., De Ridder, D. T. D., de Wit, J. B. F., & Kroese, F. M. (2011).
Breaking habits with implementation intentions: A test of underlying processes. Personality and
Social Psychology Bulletin, 37, 502513. doi: 10.1177/0146167211399102
Adriaanse, M. A., Oettingen, G., Gollwitzer, P. M., Hennes, E. P., de Ridder, D. T. D. & de Wit, J.
B. F. (2010). When planning is not enough: fighting unhealthy snacking habits by mental
contrasting with implementation intentions (MCII). European Journal of Social Psychology, 40,
12771293. doi: 10.1002/ejsp.730
Adriaanse, M. A., Vinkers, C. D. W., De Ridder, D. T. D., Hox, J. J., & De Wit, J. B. F. (2011). Do
implementation intentions help to eat a healthy diet? A systematic review and meta-analysis of
the empirical evidence. Appetite, 56, 183193. doi: 10.1016/j.appet.2010.10.012
Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological

Review, 84, 191215. doi: 10.1037/0033-295X.84.2.191
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122
147. doi: 10.1037/0003-066X.37.2.122
Page 19 of 27
Bandura, A., & Schunk, D. H. (1981). Cultivating competence, self-efficacy, and intrinsic interest
through proximal self-motivation. Journal of Personality and Social Psychology, 41, 586598. doi:
10.1037/0022-3514.41.3.586
Bargh, J. A. (1989). Conditional automaticity: varieties of automatic influence in social

perception and cognition. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (Vol. 3, pp.
5169). New York: The Guilford Press.
Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: direct effects of
trait construct and stereotype-activation on action. Journal of Personality and Social Psychology,
71, 230244.
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and Brain Sciences, 22, 577660.
doi: 10.1017/S0140525X99532147
Barsalou, L. W. (2003). Situated simulation in the human conceptual system. Language and
Cognitive Processes, 18, 513562. doi: 10.1080/01690960344000026
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617645. doi:
10.1146/annurev.psych.59.103006.093639
Baumeister, R. F., Masicampo, E. J., & Vohs, K. D. (2011). Do conscious thoughts cause
behavior? Annual Review of Psychology, 62, 331361. doi: 10.1146/annurev.psych.
093008.131126
Bayer, U. C., Achtziger, A., Gollwitzer, P. M., & Moskowitz, G. B. (2009). Responding to
subliminal sues: do if-then plans facilitate action preparation and initiation without conscious
intent? Social Cognition, 27, 183201. doi: 10.1521/soco.2009.27.2.183
Bayne, T. (2008). The phenomenology of agency. Philosophy Compass, 3, 182202. doi: 10.1111/
j.1747-9991.2007.00122.x
Blanger-Gravel, A., Godin, G., & Amireault, S. (2013). A meta-analytic review of the effect of
implementation intentions on physical activity. Health Psychology Review, 7, 2354. doi:
10.1080/17437199.2011.560095
Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social

psychology (Vol. 6, pp. 162). New York: Academic Press.
Brandsttter, V., Lengfelder, A., & Gollwitzer, P. M. (2001). Implementation intentions and
efficient action initiation. Journal of Personality and Social Psychology, 81, 946960. doi:
10.1037/0022-3514.81.5.946
Burgess, P. W., Dumontheil, I., Gilbert, S. J., Okuda, J., Schlvinck, M. L., & Simons, J. S. (2008).
On the role of rostral prefrontal cortex (area 10) in prospective memory. In M. Kliegel, M. A.
McDaniel, & G. O. Einstein (Eds.), Prospective memory: cognitive, neuroscience, developmental,
and applied perspectives. (pp. 235260). Mahwah, NJ: Erlbaum Associates.
Page 20 of 27
Castiello, U., Paulignan, Y., & Jeannerod, M. (1991). Temporal dissociation of motor responses
and subjective awareness. Brain, 114, 26392655. doi: 10.1093/brain/114.6.2639
Chambon, V., & Haggard, P. (2013). Premotor or Ideomotor: how does the experience of action
come about? In W. Prinz, M. Beisert, & A. Herwig (Eds.), Action science: foundations of an
emerging discipline. (pp. 359380). Cambridge: MIT Press.
Cohen, A.-L., Bayer, U. C., Jaudas, A., & Gollwitzer, P. M. (2006). Self-regulatory strategy and
executive control: implementation intentions modulate task switching and Simon task
performance. Psychological Research, 72, 1226. doi: 10.1007/s00426-006-0074-2
Dasgupta, N., & Greenwald, A. G. (2001). On the malleability of automatic attitudes: combating
automatic prejudice with images of admired and disliked individuals. Journal of Personality and
Social Psychology, 81, 800814. doi: 10.1037/0022-3514.81.5.800
Decety, J., Jeannerod, M., & Prablanc, C. (1989). The timing of mentally represented actions.
Behavioural Brain Research, 34, 3542. doi: 10.1016/S0166-4328(89)80088-9
Demanet, J., Muhle-Karbe, P. S., Lynn, M. T., Blotenberg, I., & Brass, M. (2013). Power to the
will: How exerting physical effort boosts the sense of agency. Cognition, 129, 574578. doi:
10.1016/j.cognition.2013.08.020
Ellis, R. (2009). Interactions between action and visual objects. In E. Morsella, J. A. Bargh, & P.
M. Gollwitzer (Eds.), Oxford handbook of human action (pp. 214224). New York: Oxford
University Press.
Ellis, R., & Tucker, M. (2000). Micro-affordance: the potentiation of components of action by
seen objects. British Journal of Psychology, 91, 451471. doi: 10.1348/000712600161934
Elsner, B., & Hommel, B. (2001). Effect anticipation and action control. Journal of Experimental
Psychology: Human Perception and Performance, 27, 229240. doi: 10.1037/0096-1523.27.1.229
Engbert, K., Wohlschlger, A., & Haggard, P. (2008). Who is causing what? The sense of agency
is relational and efferent-triggered. Cognition, 107, 693704. doi: 10.1016/j.cognition.
2007.07.021
Frak, V., Paulignan, Y., & Jeannerod, M. (2001). Orientation of the opposition axis in mentally
simulated grasping. Experimental Brain Research, 136, 120127. doi: 10.1007/s002210000583
Frith, C. D., Blakemore, & Wolpert, D. M. (2000). Abnormalities in the awareness and control of
action. Philosophical Transactions of the Royal Society B: Biological Sciences, 355, 17711788.
doi: 10.1098/rstb.2000.0734
Gallagher, S. (2000). Philosophical conceptions of the self: implications for cognitive science.
Trends in Cognitive Sciences, 4, 1421. doi: 10.1016/S1364-6613(99)01417-5
Gallagher, S. (2012). Multiple aspects in the sense of agency. New Ideas in Psychology, 30, 15
31. doi: 10.1016/j.newideapsych.2010.03.003
Page 21 of 27
Gilbert, S. J., Gollwitzer, P. M., Cohen, A.-L., Oettingen, G., & Burgess, P. W. (2009). Separable
brain systems supporting cued versus self-initiated realization of delayed intentions. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 35, 905915. doi: 10.1037/
a0015535
Glenberg, A. M. (2007). Language and action: creating sensible combinations of ideas. In M. G.

Gaskell (Ed.), The Oxford handbook of psycholinguistics (pp. 361370). Oxford: Oxford
University Press.
Glenberg, A. M., & Kaschak, M. P. (2002). Grounding language in action. Psychonomic Bulletin
& Review, 9, 558565. doi: 10.3758/BF03196313
Glenberg, A. M., & Robertson, D. A. (1999). Indexical understanding of instructions. Discourse

Processes, 28, 126. doi: 10.1080/01638539909545067
Gollwitzer, P. M. (1993). Goal achievement: the role of intentions. European Review of Social
Psychology, 4, 141185. doi: 10.1080/14792779343000059
Gollwitzer, P. M. (1999). Implementation Intentions: Strong effects of simple plans. American

Psychologist, 54, 493503. doi: 10.1037/0003-066X.54.7.493
Gollwitzer, P. M., & Brandsttter, V. (1997). Implementation intentions and effective goal
pursuit. Journal of Personality and Social Psychology, 73, 186199. doi:
10.1037/0022-3514.73.1.186
Gollwitzer, P. M., & Oettingen, G. (2011). Planning promotes goal striving. In K. D. Vohs & R. F.
Baumeister (Eds.), Handbook of self-regulation: research, theory, and applications (2nd ed., pp.
162185). New York; London: The Guilford Press.
Gollwitzer, P. M., & Schaal, B. (1998). Metacognition in action: the importance of

implementation intentions. Personality and Social Psychology Review, 2, 124136. doi: 10.1207/
s15327957pspr0202_5
Gollwitzer, P. M., & Sheeran, P. (2006). Implementation intentions and goal achievement: a
metaanalysis of effects and processes. Advances in Experimental Social Psychology, 38, 69119.
Gollwitzer, P. M., Wieber, F., Myers, A. L., & McCrea Sean, M. (2010). How to maximize
implementation intention effects. In C. R. Agnew, D. E. Carlston, W. G. Graziano, & J. R. Kelly
(Eds.), Then a miracle occurs: focusing on behavior in social psychological theory and research
(pp. 137161). New York: Oxford University Press.
Gonzlez, J., Barros-Loscertales, A., Pulvermller, F., Meseguer, V., Sanjun, A., Belloch, V., &
vila, C. (2006). Reading cinnamon activates olfactory brain regions. NeuroImage, 32, 906912.
doi: 10.1016/j.neuroimage.2006.03.037
Goodale, M. A., Pelisson, D., & Prablanc, C. (1986). Large adjustments in visually guided
reaching do not depend on vision of the hand or perception of target displacement. Nature, 320,
748750. doi: 10.1038/320748a0
Page 22 of 27
Grezes, J., Tucker, M., Armony, J., Ellis, R., & Passingham, R. E. (2003). Objects automatically
potentiate action: an fMRI study of implicit processing. European Journal of Neuroscience, 17,
27352740. doi: 10.1046/j.1460-9568.2003.02695.x
Nature Neuroscience, 5, 382385. doi: 10.1038/nn827
voluntary movements. Experimental Brain Research, 126, 128133. doi: 10.1007/s002210050722
Haggard, P., & Tsakiris, M. (2009). The experience of agency: feelings, judgments, and
responsibility. Current Directions in Psychological Science, 18, 242246. doi: 10.1111/j.
1467-8721.2009.01644.x
Hauk, O., Johnsrude, I., & Pulvermller, F. (2004). Somatotopic representation of action words
in human motor and premotor cortex. Neuron, 41, 301307. doi: 10.1016/
S0896-6273(03)00838-9
Hebb, D. O. (1949). The organization of behavior: a neuropsychological theory. New York: Wiley.
Heckhausen, H. (1989). Leistungsmotivation. In H. Heckhausen (Ed.), Motivation und Handeln

(pp. 231278). Berlin, Heidelberg: Springer Berlin Heidelberg. doi:
10.1007/978-3-662-08870-8_8
Heckhausen, H. (2008). Motivation und Handeln. Berlin: Springer-Verlag.
Hesslow, G. (2002). Conscious thought as simulation of behaviour and perception. Trends in

Cognitive Sciences, 6, 242247. doi: 10.1016/S1364-6613(02)01913-7
Hofmann, W., Baumeister, R. F., Frster, G., & Vohs, K. D. (2012). Everyday temptations: an
experience sampling study of desire, conflict, and self-control. Journal of Personality and Social
Psychology, 102, 13181335. doi: 10.1037/a0026545
Hommel, B. (1993). Inverting the Simon effect by intention. Psychological Research, 55, 270
279. doi: 10.1007/BF00419687
Horton, W. S., & Rapp, D. N. (2003). Out of sight, out of mind: occlusion and the accessibility of
information in narrative comprehension. Psychonomic Bulletin & Review, 10, 104110. doi:
10.3758/BF03196473
James, W. (1890). Principles o f psychology. New York: Holt.
Jeannerod, M. (2001). Neural simulation of action: a unifying mechanism for motor cognition.
NeuroImage, 14, S103S109. doi: 10.1006/nimg.2001.0832
Kaschak, M. P., & Glenberg, A. M. (2000). Constructing meaning: the role of affordances and
grammatical constructions in sentence comprehension. Journal of Memory and Language, 43,
508529. doi: 10.1006/jmla.2000.2705
Page 23 of 27
Kiefer, M., & Pulvermller, F. (2012). Conceptual representations in mind and brain: theoretical
developments, current evidence and future directions. Cortex, 48, 805825. doi: 10.1016/
j.cortex.2011.04.006
Lewin, K. (1926). Vorsatz, Wille und Bedrfnis. Psychologische Forschung, 7, 330385. doi:
10.1007/BF02424365
Lewin, K. (1951). Intention, will, and need. In D. Rapaport (Ed.), Organization and pathology of
thought: selected sources (pp. 95153). New York: Columbia University Press.
in relation to onset of cerebral activity (readiness-potential): the unconscious initiation of a
freely voluntary act. Brain, 106, 623642. doi: 10.1093/brain/106.3.623
Locke, E. A., & Latham, G. P. (2006). New directions in goal-setting theory. Current Directions
in Psychological Science, 15, 265268. doi: 10.1111/j.1467-8721.2006.00449.x
Lotze, R. H. (1852). Medicinische Psychologie oder Physiologie der Seele. Leipzig:

Lu, C., & Proctor, R. W. (1995). The influence of irrelevant location information on performance:
a review of the Simon and spatial Stroop effects. Psychonomic Bulletin & Review, 2, 174207.
doi: 10.3758/BF03210959
Mendoza, S. A., Gollwitzer, P. M., & Amodio, D. M. (2010). Reducing the expression of implicit
stereotypes: reflexive control through implementation intentions. Personality and Social
Psychology Bulletin, 36, 512523. doi: 10.1177/0146167210362789
Miles, J. D., & Proctor, R. W. (2008). Improving performance through implementation intentions:
are preexisting response biases replaced? Psychonomic Bulletin & Review, 15, 11051110. doi:
10.3758/PBR.15.6.1105
external cues. Consciousness and Cognition, 18, 10561064. doi: 10.1016/j.concog.2009.05.004
Oettingen, G. (2012). Future thought and behaviour change. European Review of Social
Psychology, 23, 163. doi: 10.1080/10463283.2011.643698
Oettingen, G., Pak, H., & Schnetter, K. (2001). Self-regulation of goal-setting: turning free
fantasies about the future into binding goals. Journal of Personality and Social Psychology, 80,
736753. doi: 10.1037//0022-3514.80.5.736
Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational Research,

66, 543578. doi: 10.3102/00346543066004543
Pajares, F. (1997). Current directions in self-efficacy research. In M. L. Maehr & P. R. Pintrich

(Eds.), Advances in motivation and achievement (Vol. 10, pp. 149). Greenwich, CT: JAI Press.
Page 24 of 27
Papies, E. K., Aarts, H., & de Vries, N. K. (2009). Planning is for doing: implementation
intentions go beyond the mere creation of goal-directed associations. Journal of Experimental
Social Psychology, 45, 11481151. doi: 10.1016/j.jesp.2009.06.011
Parks-Stamm, E. J., Gollwitzer, P. M., & Oettingen, G. (2007). Action control by implementation
intentions: effective cue detection and efficient response initiation. Social Cognition, 25, 248
266. doi: 10.1521/soco.2007.25.2.248
Payne, B. K. (2001). Prejudice and perception: the role of automatic and controlled processes in
misperceiving a weapon. Journal of Personality and Social Psychology, 81, 181192. doi:
10.1037/0022-3514.81.2.181
Prinz, W. (1997). Perception and action planning. European Journal of Cognitive Psychology, 9,
129154. doi: 10.1080/713752551
Pulvermller, F., Hauk, O., Nikulin, V. V., & Ilmoniemi, R. J. (2005). Functional links between
motor and language systems. European Journal of Neuroscience, 21, 793797. doi: 10.1111/j.
1460-9568.2005.03900.x
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of

reinforcement. Psychological Monographs: General and Applied, 80, 128. doi: 10.1037/
h0092976
effect contribute to explicit judgment of agency. Cognition, 110, 7483. doi: 10.1016/j.cognition.
2008.10.011
Schunk, Dale H. (1995). Self-efficacy and education and instruction. In J. E. Maddux (Ed.), Self-
efficacy, adaptation, and adjustment (pp. 281303). Boston: Springer US.
Schunk, D. H., & Pajares, F. (2009). Self-efficacy theory. In K. R. Wentzel & A. Wigfield (Eds.),
Handbook of motivation at school (pp. 3553). New York: Routledge.
Searle, J. R. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3, 417424.
doi: 10.1017/S0140525X00005756
Sheeran, P., & Orbell, S. (2000). Using implementation intentions to increase attendance for
cervical cancer screening. Health Psychology, 19, 283289. doi: 10.1037/0278-6133.19.3.283
Sheeran, P., Webb, T. L., & Gollwitzer, P. M. (2005). The interplay between goal intentions and
implementation intentions. Personality and Social Psychology Bulletin, 31, 8798. doi:
10.1177/0146167204271308
Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human information
processing: II. Perceptual learning, automatic attending and a general theory. Psychological
Review, 84, 127190. doi: 10.1037/0033-295X.84.2.127
Page 25 of 27
Shin, Y. K., Proctor, R. W., & Capaldi, E. J. (2010). A review of contemporary ideomotor theory.
Psychological Bulletin, 136, 943974. doi: 10.1037/a0020541
Simmons, W. K., Martin, A., & Barsalou, L. W. (2005). Pictures of appetizing foods activate
gustatory cortices for taste and reward. Cerebral Cortex, 15, 16021608. doi: 10.1093/cercor/
bhi038
Soon, C. S., Brass, M., Heinze, H.-J., & Haynes, J.-D. (2008). Unconscious determinants of free
decisions in the human brain. Nature Neuroscience, 11, 543545. doi: 10.1038/nn.2112
Stanfield, R. A., & Zwaan, R. A. (2001). The effect of implied orientation derived from verbal
context on picture recognition. Psychological Science, 12, 153156. doi:
10.1111/1467-9280.00326
Stewart, B. D., & Payne, B. K. (2008). Bringing automatic stereotyping under control:
implementation intentions as efficient means of thought control. Personality and Social
Psychology Bulletin, 34, 13321345. doi: 10.1177/0146167208321269
Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior.
Personality and Social Psychology Review, 8, 220247. doi: 10.1207/s15327957pspr0803_1
two-step account of agency. Consciousness and Cognition, 17, 219239. doi: 10.1016/j.concog.
2007.03.010
Tucker, M., & Ellis, R. (1998). On the relations between seen objects and components of
potential actions. Journal of Experimental Psychology: Human Perception and Performance, 24,
830846. doi: 10.1037/0096-1523.24.3.830
Verplanken, B., & Aarts, H. (1999). Habit, attitude, and planned behaviour: is habit an empty
construct or an interesting case of goal-directed automaticity? European Review of Social
Psychology, 10, 101134. doi: 10.1080/14792779943000035
Webb, T. L., & Sheeran, P. (2004). Identifying good opportunities to act: implementation
intentions and cue discrimination. European Journal of Social Psychology, 34, 407419. doi:
10.1002/ejsp.205
Webb, T. L., & Sheeran, P. (2007). How do implementation intentions promote goal attainment?
A test of component processes. Journal of Experimental Social Psychology, 43, 295302. doi:
10.1016/j.jesp.2006.02.001
Webb, T. L., & Sheeran, P. (2008). Mechanisms of implementation intention effects: the role of
goal intentions, self-efficacy, and accessibility of plan components. British Journal of Social
Psychology, 47, 373395. doi: 10.1348/014466607X267010
Webb, T. L., Sheeran, P., & Luszczynska, A. (2009). Planning to break unwanted habits: habit
strength moderates implementation intention effects on behaviour change. British Journal of
Social Psychology, 48, 507523. doi: 10.1348/014466608X370591
Page 26 of 27
Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation: Sources of the experience of
will. American Psychologist, 54, 480492. doi: 10.1037/0003-066X.54.7.480
West, R. (2008). The cognitive neuroscience of prospective memory. In M. Kliegel, M. A.

McDaniel, & G. O. Einstein (Eds.), Prospective memory: cognitive, neuroscience, developmental,
and applied perspectives (pp. 261282). Mahwah, NJ: Erlbaum Associates.
Wieber, F., & Sassenberg, K. (2006). I cant take my eyes off of itattention attraction effects of
implementation intentions. Social Cognition, 24, 723752.
Wieber, F., Sezer, L. A., & Gollwitzer, P. M. (2014). Asking why helps action control by goals
but not plans. Motivation and Emotion, 38, 6578. doi: 10.1007/s11031-013-9364-3
Wittenbrink, B., Judd, C. M., & Park, B. (2001). Spontaneous prejudice in context: variability in
automatically activated attitudes. Journal of Personality and Social Psychology, 81, 815827. doi:
10.1037/0022-3514.81.5.815
Wood, W., & Neal, D. T. (2007). A new look at habits and the habit-goal interface. Psychological
Review, 114, 843863. doi: 10.1037/0033-295X.114.4.843
Zwaan, R. A. (2004). The immersed experiencer: toward an embodied theory of language

comprehension. In B. Ross (Ed.), The psychology of learning and motivation (Vol. 44, pp. 3562).
San Diego: Academic Press.
Zwaan, R. A., Stanfield, R. A., & Yaxley, R. H. (2002). Language comprehenders mentally
represent the shapes of objects. Psychological Science, 13, 168171. doi:
10.1111/1467-9280.00430
Zwaan, Rolf A., & Taylor, L. J. (2006). Seeing, acting, understanding: motor resonance in
language comprehension. Journal of Experimental Psychology: General, 135, 111.
Page 27 of 27
Neural Correlates of Intention

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001

Roee Gilron
Shiri Simon
Roy Mukamel
DOI:10.1093/acprof:oso/9780190267278.003.0004

At the behavioral level, humans are constantly engaged in performing motor acts, but the neural
activity patterns that give rise to the initiation of such acts are poorly understood. Early studies
by Deecke and Libet demonstrated robust changes in neural activity (measured with scalp EEG),
which begin several hundreds of milliseconds prior to the overt performance of a voluntary
motor act. Interestingly, these changes in neural activity start even before the subjective time
point at which the subjects first feel the urge to move. Primate studies, neuroimaging studies in
healthy human subjects, case reports of pathological conditions, and invasive
electrophysiological data obtained from patients all support a model in which the intention to
perform an action results from coordinated pre-conscious activity of neuron assemblies in
parietal and frontal circuits. This chapter discusses neurophysiological studies examining the
relationship between neural activity and the emergence of the subjective feeling of intention to
act.
Keywords: pre-conscious activity, neural activity, subjective feeling, intention to act, motor acts
Introduction
At the core of human society is the belief that we are in control of our actions and are therefore
held accountable for their consequences. A key element underlying accountability is the concept
of intention. Stepping on the foot of the person in front of me when standing in line at the
supermarket might seem to be an act of aggression, but if performed inadvertently the
normative repercussion should be minor. The victims assumption is that had I predicted in
Page 1 of 18
advance that moving my leg would result in stepping on his foot, I wouldnt have moved my leg
in the first place. Thus, intention requires two elements: first, the volitional aspect of wanting to
perform a certain action (moving the foot), and second, the prediction aspectthe ability to
foresee the consequences of the action (stepping on someone else). Young children are usually
not held accountable for their actions since in many cases their ability to predict the
consequences of their actions is not fully developed. At the judicial level, the presence or
absence of intention determines the level of accountability, which in turn evokes different penal
codes, even for identical action consequences.
(p.96) Over the last several decades there has been a growing interest in the neural
underpinnings of agency and volition since they carry special significance for judicial systems,
various pathologies (e.g., anarchic hand, schizophrenia), and a wide variety of fields ranging
from engineering to philosophy. For example, a better understanding of the neurobiology
underlying agency and volition is relevant for building neural prosthetics for paralyzed patients
who are unable to translate intentions into executable actions. Additionally, such knowledge is
informative for philosophical debates about the concept of free will.
Several studies examining these issues have loosely used terms such as urge or desire to
express the subjective feeling of wanting to perform an action, as reported by the subjects.
Throughout this chapter, we use the term intention to describe these subjective feelings. We
define intention as awareness of an imminent execution of a voluntary motor plan aimed at
achieving a certain sensory outcome. In particular, we focus on intentions that are followed by
actions within a relatively short temporal window (on the order of seconds) as opposed to long-
term intentions (e.g., the intention to take a hike on the weekend). We review human studies
examining the neural correlates of intentions using various means, including average evoked
responses, changes in spectral power and firing rates of individual cells, clinical case reports,
and electrical brain stimulation. We conclude by proposing a possible mechanism by which
activity of sensory-motor neurons in medial frontal and parietal regions plays a role in the
emergence of intention.
Physiological Events Preceding Conscious Intention

We begin by discussing three markers of neural activity that have been demonstrated to precede
voluntary actions: event-related potentials, changes in the spectral power of local field
potentials, and changes in the firing rate of individual cells. These markers correspond with
different aspects of neural activity (see Box 4.1).
Box 4.1
EEG (electroencephalography)
A noninvasive method for recording electrical activity from the brain by placing recording
electrodes over the scalp. The measured signal corresponds to ionic current flow from large
neural populations (mostly pyramidal cells). This technique provides high temporal
resolution but suffers from low spatial resolution, thus making exact anatomical claims
regarding the source of the recorded signal difficult.
Page 2 of 18
LFP (local field potentials)

Voltage changes resulting from both dendritic currents and action potentials of large
populations of neurons. These signals are typically recorded invasively using thin
(submillimeter) microwires or contact electrodes (~2 mm) in the case of ECoG.
ECoG (Electrocorticography)
An invasive method for measuring population neural activity. Multiple recording electrodes
(typically ~2 mm contacts) are placed during surgery directly on the surface of the brain of
various clinical patients (e.g., epilepsy). These electrodes allow recording local field
potentials (LFPs) from the underlying brain tissue.
fMRI (functional magnetic resonance imaging)

A noninvasive method that measures changes in blood oxygen content. These changes are
used as a proxy to estimate the underlying neural activity.
ERP (event-related potential)

An EEG signal, locked to stimulus onset (or any other temporal event of interest). This signal
is obtained by averaging the evoked EEG signal across many trials. This signal is typically
divided into different components corresponding to various behavioral/neural aspects.
ERD/ERS (event-related desynchronization/synchronization)

Changes in neural activity are not necessarily time-locked to an external event. Such
changes often appear in the power of the LFP signal at specific frequency bands. These
power changes infer an enhancement or reduction in synchronized neural activity across a
population of neurons. ERD is interpreted as a correlate of enhanced cortical activity since
neurons are not firing in phase with each other. In contrast, ERS is often interpreted as a
deactivated cortical state in which neurons respond in phase with each other.
Single unit recordings

These are extracellular electrical recordings of spiking activity, representing the firing rate
of a specific neuron. In humans, such recordings are performed mostly in epileptic patients
undergoing clinical evaluation for potential surgical treatment.
Affordance
The set of possible actions that can be performed with an object. Many objects have multiple
affordancesfor example, a door knob affords twisting, pulling, and pushing. The same
object can have different affordances for different individuals depending on the set of skills
within their motor repertoire.
EMG (electromyogram)
A measure of the electrical activity produced by muscles. This signal can be recorded by
placing recording electrodes on the muscle.
Page 3 of 18
Event-Related Potentials (ERP)

Early scalp Electroencephalogram (EEG) studies by Deeke, Kornhuber, and Gilden (Deecke et
al., 1969; Gilden et al., 1966; Kornhuber & Deecke, 1965) reported the existence of a slow
negative potential in the EEG signal that precedes the onset of voluntary movements (as
recorded by electromyogram [EMG]). This slow potential, recorded over frontal and central
regions of the brain, appeared as early as 850 ms prior to the onset of voluntary movement (p.
97) (p.98) and was termed Bereitschaftspotential (BP) or readiness potential (RP). In a
seminal study by Libet and colleagues, the onset of this pre-movement potential was compared
with the subjective time in which subjects first felt the urge to perform a quick voluntary flexion
of the finger or wrist (Libet et al., 1983). Subjects were asked to observe a rotating clock dial on
a screen and to make a movement at the time they chose. Following this movement, the subjects
were asked to report the position of the clock dial at the moment when they first felt the urge
to move. This time point was termed W and defined as . . . the time of appearance of [the
subjects] conscious awareness of wanting to perform a given self-initiated movement (Libet,
1983, p. 627). In agreement with the earlier studies, the RP preceded movement onset.
However, it was further found that the RP also preceded W by several hundred milliseconds.
The idea of a physiological marker that precedes the intention to move has tantalized the field of
neuroscience ever since. These studies provided the first realization that a subjective feeling
such as intention can be explored empirically with physiological measures. Throughout this
chapter, we use the term W, in reference to the time point at which intention emerges.
Further characterization of the readiness potential in relation to movement onset has revealed
that it comprises subcomponents, including an early component (termed early RP) starting ~850
ms prior to movement onset, a middle component (termed lateralized RP) starting around ~400
ms, and a very late component (termed motor potential MP) starting ~50100 ms prior to
movement onset (Shibasaki & Hallett, 2006; Yamamoto et al., 2004). While the first component
is bilateral, the latter two components are stronger in the (p.99) hemisphere contralateral to
the limb that is about to move. Although these studies demonstrate that the onset of readiness
potential precedes movement onset, its temporal relationship to W is unclear. Haggard and
Eimer examined this issue further and demonstrated that the onset of the middle RP component
co-varied with how early the subjects reported W relative to movement onset (Haggard & Eimer,
1999). The early RP component did not co-vary with W. This suggests that the onset time of the
lateralized RP component (as opposed to the early RP component) is linked to the emergence of
intention.
In another study, Sirigu and colleagues measured RP using scalp EEG from patients with lesions
in the parietal lobe or cerebellum. Their experimental design was based on Libets original
experiment in which patients performed self-paced simple button presses (Sirigu et al., 2004). At
the behavioral level, the patients were not different from healthy controls in estimating the time
of finger movement (~20 ms from actual time of button press). Similar to healthy controls in this
and other studies, when estimating W, patients with lesions in the cerebellum reported an
intention to press the button ~240 ms prior to movement onset. In contrast, parietal patients
estimated the onset of intention at 55 ms prior to movement onset. Interestingly, these
behavioral results were echoed by the onset time of RP over central and frontal cortical regions.
In the parietal patients, the RP was drastically reduced when they had to report W, while in
cerebellar patients and healthy controls, clear RPs starting ~1 second prior to movement onset
Page 4 of 18
were observed. The EEG studies described thus far point to a neural event in the form of
readiness potentials originating from frontal/parietal regions that precedes and correlates with
the temporal signature of intention (W).
Under certain circumstances, as part of a clinical evaluation, electrodes are placed directly (in
an invasive manner) on the brain of human patients in order to record neural activity. In
addition to the clinical importance, these circumstances provide a unique window of opportunity
to gain a basic understanding of brain function in humans (Mukamel & Fried, 2012). These
invasive studies in patient populations have provided further invaluable insight into the
anatomical localization of the cortical and subcortical sources of the readiness potential (see
Figure 4.1).
Using direct recordings from the brain

(electrocorticography [ECoG]), various
studies have examined the characteristics of
the readiness potential in multiple regions
within the motor system hierarchy. Ikeda
and colleagues recorded a RP in primary
motor area (M1) and the supplementary
motor area (SMA) while patients performed
voluntary movements with the finger, foot,
or tongue, and also during vocalization
(Ikeda et al., 1992). They report a RP
starting between ~500 and ~1500 ms prior Figure 4.1 Cortical regions from which RP
to voluntary movement that originates in was recorded.
both regions. The RP in SMA was obtained SMA: supplementary motor area; ACC:
during movements with (p.100) ipsilateral anterior cingulate cortex.
(same side) and also contralateral effectors,
but was stronger for movements with
effectors in the contralateral side. In M1, RP
was found only for movement on the
contralateral side. Regarding the selectivity
of the RP in relation to the effector that is
moved, converging evidence supports a
somatotopic organization of the RP in SMA
(with a RP for hand movements more rostral
than a RP for leg movements).
Yazawa and colleagues demonstrate that

RPs can be recorded both from SMA and
even more anteriorly in the pre-SMA
(Yazawa et al., 2000), although in the pre- Figure 4.1 Cortical regions from which RP
was recorded.
SMA the RP is much less sensitive to
SMA: supplementary motor area; ACC:
effector type (Ikeda et al., 1992; Yamamoto anterior cingulate cortex.
et al., 2004; Yazawa et al., 2000). Similarly,
in the pre-motor cortex on the lateral side,
Page 5 of 18
Kunieda and colleagues report readiness potentials that are invariant to which effector performs
the action (e.g. finger, foot, lips) (Kunieda et al., 2004). It should be noted that in addition to RPs
recorded from the cortex, subcortical sources have also been demonstrated in structures of the
descending motor pathway, including putamen, globus pallidus, the head of caudate nucleus,
and the posterior thalamus (Rektor et al., 2001a; Rektor et al., 2001b; Rektor et al., 2001c).
The studies described thus far demonstrate changes in brain potentials starting several hundred
milliseconds prior to movement onset. These potentials are characterized by three components
(early, middle, and late RP), with the middle component co-varying with the timing of intention.
Invasive studies demonstrate that this potential is observed in a distributed network of cortical
(primary motor, lateral premotor, SMA, and pre-SMA; see Figure 4.1) and subcortical regions,
within the motor system hierarchy.
(p.101) Changes in Spectral Power

The RP described in the preceding section corresponds to synchronized (evoked) neural activity
that is phase-locked to a specific event (such as movement onset). Changes in neural activity can
also be non-phase-locked (induced). Since these changes are not perfectly timed with an
external event, they are hard to detect when examining evoked potentials averaged across
multiple trials. Such non-phase-locked neural events are better observed by measuring changes
in the spectral power of the EEG signal in different frequency bands. These changes can
manifest as either increases (event-related synchronization [ERS]) or decreases (event related
desynchronization [ERD]) in power at various frequency bands (Buzsaki et al., 2012; Neuper et
al., 2006). Both ERPs and changes in spectral power reflect activity changes of large populations
of neurons.
Ohara and colleagues measured changes in spectral power of the ECoG signal while patients
performed self-paced finger/wrist extensions (Ohara et al., 2000). The authors found ERD in low
frequencies (~1022 Hz) originating from SMA, M1, and primary somatosensory cortex (S1) that
started between 0.5 and 3.4 seconds prior to movement onset. These power changes started
earlier in SMA, and onset times did not depend on the laterality of the moving limb (i.e., ipsi- or
contralateral to the recording electrode). In M1 and S1, the onset of ERDs was later relative to
SMA. In addition, ERD in these regions was more strongly lateralized in terms of latency and
amplitude (i.e., ERD in M1 and S1 for contralateral movement started earlier and was stronger
than ERD for ipsilateral movement).
In another study, Rektor and colleagues compared the anatomical distribution of ERDs in the
alpha/beta range (~820 Hz) and RPs, while patients performed self-paced finger flexions with
their right or left fingers (Rektor et al., 2006). While in sensorimotor regions (such as S1/M1,
and SMA) they found both RPs and ERDs preceding movement, ERDs were more widespread
and could also be found in other regions in which RPs were not found (including orbitofrontal
and inferior parietal cortices, and lateral and mesial temporal cortices).
Taken together, the temporal profile of ERDs and RPs are similarpreceding movement onset
by up to ~3 seconds. While the anatomical distribution of ERDs is more widespread than that of
RPs, there is large overlap of these two measures of neural activity in sensorimotor regions.
Page 6 of 18
Changes in Firing Rates

The RP and changes in spectral power described thus far pertain to activity changes across very
large populations of neurons. However, these measures (p.102) do not allow discrimination of
firing patterns of individual cells. In a recent study, we recorded extracellular single and multi-
unit activity from 12 patients while they performed a variant of the task originally described by
Libet (Fried et al., 2011). The patients observed the dial of an analog clock rotate with a period
of 2.6 seconds and were instructed to press a button on the keyboard whenever they felt the
urge to do so. The timing of the button press was registered together with the neural activity,
and the subjective time-point of urge to press (W) was calculated based on the patients post-hoc
report of the location of the dial on the clocks circumference. In agreement with previous
studies, the patients reported the time of urge as occurring approximately 200 ms prior to the
actual time of button press (e.g., Sirigu et al., 2004). We recorded spiking activity from neurons
in frontal regions including the SMA, pre-SMA, and anterior cingulate cortex (ACC), and also
from neurons in medial temporal regions. We found clear deviations from baseline firing rate of
individual neurons that preceded W by several hundred milliseconds. These changes were more
common in frontal regions (17% of recorded cells) compared with temporal regions (8%), and
within frontal regions, the highest proportion of cells (23%) was observed in the SMA. This is in
agreement with a previous study showing that the majority of cells in SMA and pre-SMA display
increased firing rate (compared with baseline) during a 5-second delay period before execution
of a finger-to-thumb apposition task (Amador & Fried, 2004). In our study, changes in firing
rates were not uniform across neurons. Some cells displayed gradual increases in firing rate,
while others (with more or less equal distribution) displayed gradual decreases in firing rate as
W time was approached. Firing rate in both cell populations started to deviate from baseline
~1500 ms prior to W. Interestingly, 100 ms piror to W, firing rates started to return to baseline.
Both ERPs and changes in spectral power are measures that are based on averaged activity
across multiple trials. On the level of individual trials, we showed deviations in firing rate of
single neurons that allowed prediction (80% accuracy) of an upcoming intention 700
milliseconds in advance. These data suggest that both increases and decreases in firing rate in
frontal regions (predominantly in the SMA) might have a role in determining the time of
intention.
The data reviewed so far points to a correlation between neural activity and the emergence of
intention. This correlation is reported at the level of populations of cells (ERP and changes in
low-frequency spectral power), and also at the level of individual cells (changes in firing rate).
These changes in neural activity are predominant in S1/M1, SMA/pre-SMA, and also pre-motor
cortex.
(p.103) Physiological Markers of Intended Movements

Are the physiological markers discussed thus far a hallmark of intended movements, or are they
associated with any movement type? The coupling between the presence/absence of various
physiological markers and intention has been examined in various pathologies.
Patients with Gilles de la Tourette syndrome (commonly known as Tourette) suffer from
uncontrollable motor and vocal tics. If the source of the RP is motoric in nature, and not linked
to higher-level goals and intentions, one would expect a similar RP during voluntary and non-
Page 7 of 18
voluntary movements. In one study, five out of six patients showed no RP during endogenous
tics, but when the patients had to voluntarily mimic their tics, a RP starting ~500 ms earlier was
evident (Obeso et al., 1981). In another study, similar results were obtained in three out of five
subjects (Karp et al., 1996). These results suggest that intention in controlled movement is
coupled with the presence of an RP.
An additional pathology in which voluntary and non-voluntary movements were compared is

anarchic hand syndrome (AHS). In patients with AHS, the affected hand occasionally performs
meaningful goal-directed actions while the patients deny having an intention to perform these
actions. As opposed to ballistic movements typical of Tourette patients, the actions performed by
the anarchic hand are meaningful but inappropriate to the context. The patients are fully aware
of their inappropriateness and often try to stop the action by using their healthy hand. This
syndrome is often associated with lesions to the parietal or mesial frontal (SMA) regions (Della
Sala & Marchetti, 2005). In a case study, Assal and colleagues report a patient suffering from
AHS in the left hand following stroke in the right parietal lobe. Using fMRI, they found increased
activity in a network of motor regions (including right M1 and premotor cortex, and also left
inferior frontal gyrus and frontopolar cortex) when the patients performed movements with their
left hand experienced as voluntary. However, during left hand movements that were not
experienced as being under conscious control, only activity in right M1 was found (Assal et al.,
2007). These results suggest that although activity in M1 is sufficient to evoke movement, it is
not sufficient for the emergence of intention, which requires the recruitment of additional motor
regions.
An fMRI study conducted in our lab supports the involvement of the right parietal cortex in the
dissociation between intention and movement kinematics. Subjects had to perform hand
movements in different directions across trials in order to obtain the same sensory goal. While
primary and pre-motor regions were sensitive to the specific direction of hand movement, we
found a region in the right parietal cortex that was sensitive to the intended sensory (p.104)
goal but invariant to the specific direction of hand movement that was performed to achieve it
(Krasovsky et al., 2014).
These data further support the involvement of descending pathways controlling movement (at
the anatomical level) and the presence of RP (at the functional level) in the emergence of
intention.
Electrical Stimulation and Feeling of Intention

In some clinical settings, implanted electrodes in patients are used not only for recording ECoG
or single unit activity but also to electrically stimulate and evaluate the functional role of the
underlying tissue. This method allows assessment of the causal role that specific brain regions
might play in generating behavior.
As part of a functional mapping procedure conducted in epileptic patients, Fried and colleagues
stimulated regions of the supplementary motor area (SMA) in 13 patients to determine the
relation of the area of seizure onset to behavioral function (Fried et al., 1991). In this procedure,
different regions of the brain are electrically stimulated while clinicians look for overt motor
responses. The authors reported the existence of a crude somatotopic map along the caudal-
rostral axis of the SMA, with the legs represented in caudal regions and the head and eyes
Page 8 of 18
represented more in rostral regions. This somatotopic organization was also confirmed in later
studies (e.g., Ikeda et al., 1992; Lim et al., 1994; Yazawa et al., 2000). Surprisingly, stimulation
of different sites in some of the patients elicited the urge to move (as reported by the patients),
even though no overt movement was observed. This reported urge was somatotopically
organized and commonly corresponded with overt movement evoked when stimulation intensity
was increased. For example, at low stimulation currents, the patient reported an urge to lift the
right elbow, and increasing the stimulation current evoked overt right arm abduction.
In another set of 15 patients, Lim and colleagues confirmed the somatotopic organization of the
SMA with the leg representation caudal (toward M1) and the head and upper extremities
representation rostral (Lim et al., 1994). Electrically stimulating the SMA evoked sensory
responses including numbness, tingling, or pressure sensation contralateral to the stimulating
electrode. Importantly, occasionally some patients reported a sensation of movement or an urge
to move their limbs but lack of overt movement. Invariably, the authors report that increasing
the stimulation current evoked overt motor responses.
Recently, Desmurget and colleagues electrically stimulated the parietal and pre-motor regions of
tumor patients as part of a functional mapping procedure designed to minimize postoperative
sequelae following tumor resection (Desmurget et al., 2009). Low current stimulation in inferior
parietal regions (p.105) evoked an intention/desire to move (as reported by the patients) with
no overt movement recorded. Increasing the current level produced a sensation of movement,
although again, no electromyogram (EMG) activity was observed. In contrast, stimulation of
lateral pre-motor regions did evoke overt movements at higher stimulation currents. In these
conditions, the patients did not report a desire to move, and in many cases they were not even
aware of such movements.
These studies provide strong causal evidence for the involvement of mesial frontal and lateral
parietal regions in the intention to perform a voluntary motor act.
Discussion
In what follows, we discuss the findings reviewed in the preceding sections and conclude by
suggesting a plausible neural mechanism underlying the emergence of intention to move.
Research over the past few decades detected physiological markers that correlate with the
intention to move a specific effector. These physiological markers include slow changes in
electrical potentials (RPs), changes in spectral power (ERD/ERS) that reflect changes in the
degree of synchronization across large populations of neurons, and also changes (both increases
and decreases) in the firing rate of individual neurons. In some cases these physiological
markers even allow prediction of an upcoming intention to move several hundred milliseconds in
advance.
RPs have been commonly subdivided into different components including an early, middle, and
late component, which start ~800, ~400, and ~100 ms prior to movement onset, respectively.
Since the time of movement onset and W are not independent, it is difficult to disambiguate the
functional role of different RP components with regard to W or movement preparation. However,
it has been shown that the onset of the middle component of RPs co-varies with the temporal
Page 9 of 18
report of W. Furthermore, the presence/absence of RP seems to co-vary with whether the

movement is experienced as voluntary.
With regard to the anatomical distribution of these physiological markers, most evidence for RPs
in humans is ascribed to frontal/parietal regions, including S1/M1, SMA, and pre-SMA
(predominantly on the mesial aspect of the cortex), and also the pre-motor cortex (on the lateral
aspect of the cortex). Electrical stimulation studies further corroborate the involvement of
inferior parietal cortex and SMA in the emergence of intention to move (even in the absence of
actual movement). In contrast, stimulation of the lateral motor regions has not been reported to
evoke the sensation of intention. Stimulation in these regions has been mostly associated either
with evoked movement or (p.106) movement interference (such as speech arrest; Filevich et
al., 2012). Further characterization of the different functional properties of these regions might
shed light on their causal role in generating intention.
Model
This chapter has discussed a wide body of literature that described neural activity preceding
intention of upcoming movement. However, the evidence we presented remained mostly
correlational with regard to neural activation associated with movement and the emergence of
intention. The fact that pre-conscious neural activity is observable across a wide variety of
methodologies makes the absence of a mechanistic proposal with regard to its function all the
more glaring. Our model aims to bridge this gap by proposing an underlying functional role for
such neural activity in the emergence of intention and voluntary action.
We suggest that activation of a special class of neurons with visuomotor properties may hold
functional significance for the emergence of intentions. We describe how these neurons, co-
localized in the same brain regions in which we find RPs, may lay a framework in which
intentions are translated into actions. This framework is supported by several lines of evidence,
including anatomical, behavioral, functional, and lesion data.
Neurons in various regions of the motor system have been found to respond not only during
overt action but also during passive observation of actions performed by others (mirror neurons;
Rizzolatti & Sinigaglia, 2010). Mirror neurons have been classified as either strictly congruent,
responding to one type of action (either executed or observed), or broadly congruent
responding to one type of action during execution but to several types of actions during
observation (Gallese et al., 1996). It should be noted that typically, the various types of observed
actions that broadly congruent mirror neurons respond to are logically related (e.g., different
grips). Another type of cells with visuo-motor properties are canonical neurons. Like mirror
neurons, canonical neurons respond during execution of goal-directed actions. However, while
mirror neurons respond during observation of actions involving an agent interacting with an
object, canonical neurons respond to the mere perception of graspable objects in the absence of
an interacting agent.
Together, these various types of visuomotor neurons (strictly congruent, broadly congruent, and
canonical neurons) have the functional properties to represent possible actions based on
perceived objects/actions in the environment. Their activity could represent the first step in
generating an intention to act. For example, seeing a ball (or someone interact with a ball)
evokes activity in several networks of visuomotor neurons corresponding to specific sets (p.107)
Page 10 of 18
of possible actionssuch as bouncing, throwing, or kicking. Indeed, visuomotor neurons in the

ventral pre-motor cortex of monkeys have been shown to respond in a differential manner to
reachable versus unreachable objects (Caggiano et al., 2009). At the behavioral level, observing
actions performed by others is known to trigger an urge (and increased tendency) to perform
similar actions (e.g., social contagion, the chameleon effect; Chartrand & Bargh, 1999). The
activity of mirror neurons and canonical neurons might play a role in such phenomena by
constantly updating the set of actions afforded by objects perceived in the environment (Cisek &
Kalaska, 2010).
We propose that the affordances discussed earlier (and the activity of mirror neurons) may be
represented by the early phase of the RP. During the early RP phase, our model assumes that
neural representations of multiple affordances are evoked (Figure 4.2A). Several lines of
evidence support this assumption. The anatomical regions where RPs have been recorded
largely overlap with regions containing neurons with mirroring properties (including pre-motor,
SMA, and primary motor cortex). Furthermore, decreases in spectral power (ERD) prior to
voluntary movement are reported in the 820 Hz frequency band. Decreased power of the EEG
signal in the same frequency band is also associated with mirror neuron activity (mu
suppression; Pineda, 2005). Neurons with mirroring properties have also been demonstrated in
the parietal cortex, where electrical stimulation has been shown to elicit an urge to move. In
addition, a large body of research from the monkey literature shows that frontal-parietal circuits
simultaneously code the reaching direction of several movements (Cisek & Kalaska, 2005; Cui &
Andersen, 2011; Klaes et al., 2011). It should be noted that the early RP is not lateralized, in
agreement with the computation of multiple affordances, possibly with different effectors.
Page 11 of 18
Figure 4.2 Model scheme depicting the

functional role of visuo-motor neurons in the
emergence of intended actions. (a)
Affordance Representations. Seeing an
object in the environment evokes multiple
affordances through the activity of visuo-
motor neurons within various regions of the
motor system hierarchy. The various
affordances can be performed with different
effectors and are represented by the early
phase of the RP, which is bilateral. (b) Action
Selection. The neural representation of an
affordance with a specific effector crosses a
selection threshold. This selection process is
represented by the middle RP phase and is
lateralized to the contralateral hemisphere.
During this phase the cognitive processes
that give rise to W gradually evolve
(represented in gray gradient) until W is
finally reported. (c) Action Execution. The
transition toward the late phase of the RP
occurs when an execution threshold
(between W and M) is crossed. The motor
potential is evoked when the selected action
is executed.
Page 12 of 18
At a certain stage, the neural representation

of a specific affordance crosses a selection
threshold. This selection process is
represented by the middle RP phase (Figure
4.2B). Unlike the early phase of the RP, the
middle and also the late phases are
lateralized in agreement with the
interpretation that they represent the
choice of a motor plan carried out with a
specific effector. Interestingly, patients with
parietal lesions or patients with Tourette
report feeling W almost concurrently with
movement initiation. The absence of RP Figure 4.2 Model scheme depicting the
preceding movement in these patients functional role of visuo-motor neurons in the
emergence of intended actions. (A)
suggests that the middle phase RP might Affordance Representations. Seeing an
reflect the culmination of neural processes object in the environment evokes multiple
that give rise to intention. Taken together, affordances through the activity of visuo-
these findings suggest that W correlates motor neurons within various regions of the
motor system hierarchy. The various
with the point in time at which an affordances can be performed with different
affordance was selected. In essence, the effectors and are represented by the early
earliest point in time at which the intention phase of the RP, which is bilateral. (B)
Action Selection. The neural representation
is felt (and subjectively reported) is tied to
of an affordance with a specific effector
the point in time at which a selection crosses a selection threshold. This selection
threshold was crossed. process is represented by the middle RP
phase and is lateralized to the contralateral
Once an affordance has been selected, the hemisphere. During this phase the cognitive
processes that give rise to W gradually
transition to the late phase of the RP occurs
evolve (represented in gray gradient) until W
when an execution threshold is crossed is finally reported. (C) Action Execution. The
(Figure 4.2C). If observing (p.108) a transition toward the late phase of the RP
graspable object or observing someone else occurs when an execution threshold
(between W and M) is crossed. The motor
perform an action automatically evoked potential is evoked when the selected action
responses in motor cortex, there needs to is executed.
be a braking mechanism in order to avoid
automatic (and sometimes dysfunctional)
imitation or uncontrolled manipulation of
surrounding objects. What mechanism prevents the immediate execution/imitation of these
actions? Evidence for such a putative braking mechanism comes from single unit recordings in
monkeys and from epileptic patients which show that a subset of mirror neurons increase their
firing rates during action execution but decrease their firing rates during action observation
(Kraskov et al., 2009; Mukamel et al., 2010). Such functional properties might underlie the
inhibition of unwanted actions. Interestingly, patients with frontal lobe lesions have been shown
to suffer from (p.109) compulsory imitative behavior (Chow & Cummings, 2007) or utilization
behavior (Gazzaley & DEsposito, 2007). The symptoms in these pathologies suggest a deficit in
such a brake mechanism, resulting in the inability of these patients to inhibit unwanted actions.
It would be interesting to characterize RP in such patients.
Page 13 of 18
Some testable predictions emerge from our model. If the early RP component represents action
affordances, an increase in the number of affordances should correspond with increased
amplitude of the early RP. Moreover, interfering with the RP buildup at different stages (early/
late) should modulate the subjects ability to report any experience of emerging intention.
Several questions regarding this putative mechanism remain unanswered. The functional role of
each region within the fronto-parietal network underlying intention is not well established. RP
has been shown in pre-motor, SMA, and parietal regions. However, only electrical stimulation in
parietal regions and SMA elicits an urge to move. Increasing stimulation current evokes
movement in SMA but not in parietal regions. These results point to functional differences
between the regions that need to be elucidated. Furthermore, the temporal dynamics of activity
between the different regions within the network are not known.
We note that our model does not account for actions that are internally generated, presumably
originating from within since they lack any discernable evoking stimulus. Actions are often
classified along an axis ranging from internally to externally generated. Whereas externally
generated actions are often in immediate response to an external stimulus (such as catching a
ball), internally generated actions take place without any evident overt stimulus. We propose
that the distinction between internally and externally generated actions corresponds with the
temporal distance between the action-evoking stimulus and motor act. Within this temporal
framework, internally generated actions are actions in which the temporal distance from the
evoking stimulus is large (e.g., actions that are evoked by long-term memory), whereas
externally triggered actions have a shorter temporal window that separates them from their
evoking stimulus. Much less is known about the processes by which internally generated
intentions emerge. The study of spontaneous neural activity preceding voluntary action is
currently an active field of research taking its first steps in addressing this issue (Rolls & Deco,
2011; Schurger et al., 2012).
The answers to some of these questions will have implications in various fields, including
pathologies (e.g., Tourette and anarchic hand syndrome) and brain machine interfaces, and may
help the judiciary system decide the degree to which the defendant in a trial should be held
accountable for the consequences of his action.
(p.110) Acknowledgments
We thank T. Roberts for fruitful comments. R. M. is supported by the Human Frontiers Science
Project Organization, and the Israeli Center of Research Excellence (ICORE). R. G. and S. S. are
supported by the Sagol School of Neuroscience.
References
Amador, N., & Fried, I. (2004). Single-neuron activity in the human supplementary motor area
underlying preparation for action. Journal of Neurosurgery, 100, 250259.
Assal, F., Schwartz, S., & Vuilleumier, P. (2007). Moving with or without will: functional neural
correlates of alien hand syndrome. Annals of Neurology, 62, 301306.
Page 14 of 18
Bhatia, K. P., & Marsden, C. D. (1994). The behavioral and motor consequences of focal lesions
of the basal ganglia in man. Brain, 117, 859876.
Buzsaki, G., Anastassiou, C. A., & Koch, C. (2012). The origin of extracellular fields and currents:
EEG, ECoG, LFP and spikes. Nature Reviews: Neuroscience, 13, 407420.
Caggiano, V., Fogassi, L., Rizzolatti, G., Thier, P., & Casile, A. (2009). Mirror neurons
differentially encode the peripersonal and extrapersonal space of monkeys. Science, 324, 403
406.
Chartrand, T. L., & Bargh, J. A. (1999). The chameleon effect: the perception-behavior link and
social interaction. Journal of Personality and Social Psychology, 76, 893910.
Chow, T. W., & Cummings, J. L. (2007). Frontal-subcortical circuits. In B. L. Miller & J. L.

Cummings (Eds.), The human frontal lobes: functions and disorders (pp. 2543). New York:
Guilford Press.
Cisek, P., & Kalaska, J. F. (2005). Neural correlates of reaching decisions in dorsal premotor
cortex: specification of multiple direction choices and final selection of action. Neuron, 45, 801
814.
Cisek, P., & Kalaska, J. F. (2010). Neural mechanisms for interacting with a world full of action
choices. Annual Review of Neuroscience, 33, 269298.
Cui, H., & Andersen, R. A. (2011). Different representations of potential and selected motor
plans by distinct parietal areas. Journal of Neuroscience, 31, 1813018136.
Deecke, L., Scheid, P., & Kornhuber, H. H. (1969). Distribution of readiness potential, pre-
motion positivity, and motor potential of the human cerebral cortex preceding voluntary finger
movements. Experimental Brain Research, 7, 158168.
Della Sala, S., & Marchetti, C. (2005). Anarchic hand. In F. Hans-Joachim, M. Jeannerod, M.
Hallett, & R. Leiguarda (Eds.), Higher-order motor disorders: from neuroanatomy and
neurobiology to clinical neurology (pp. 291301). Oxford: Oxford University Press.
Desmurget, M., Reilly, K. T., Richard, N., Szathmari, A., Mottolese, C., & Sirigu, A. (2009).
Movement intention after parietal cortex stimulation in humans. Science, 324, 811813.
Filevich, E., Kuhn, S., & Haggard, P. (2012). Negative motor phenomena in cortical stimulation:
implications for inhibitory control of human action. Cortex: A Journal Devoted to the Study of the
Nervous System and Behavior, 48, 12511261.
Fried, I., Katz, A., McCarthy, G., Sass, K. J., Williamson, P., et al. (1991). Functional organization
of human supplementary motor cortex studied by electrical stimulation. Journal of Neuroscience,
11, 36563666.
Fried, I., Mukamel, R., & Kreiman, G. (2011). Internally generated preactivation of single
neurons in human medial frontal cortex predicts volition. Neuron, 69, 548562.
Page 15 of 18
Gallese, V., Fadiga, L., Fogassi, L., & Rizzolatti, G. (1996). Action recognition in the premotor
cortex. Brain, 119, 593609.
Gazzaley, A., & DEsposito, M. (2007). Unifying Prefrontal Cortex Function. In B. Miller, & J.
Cummings (Eds.), The human frontal lobes function and disorders (pp. 187206). New York: The
Guilford Press.
Gilden, L., Vaughan, H. G., Jr., Costa LD. (1966). Summated human EEG potentials with
voluntary movement. Electroencephalography and Clinical Neurophysiology, 20, 433438.
voluntary movements. Experimental Brain Research, 126, 128133.
Ikeda, A., Luders, H. O., Burgess, R. C., & Shibasaki, H. (1992). Movement-related potentials
recorded from supplementary motor area and primary motor area: role of supplementary motor
area in voluntary movements. Brain, 115(Pt 4), 10171043.
Karp, B. I., Porter, S., Toro, C., Hallett, M. (1996). Simple motor tics may be preceded by a
premotor potential. Journal of Neurology, Neurosurgery and Psychiatry, 61, 103106.
Klaes, C., Westendorff, S., Chakrabarti, S., & Gail, A. (2011). Choosing goals, not rules: deciding
among rule-based action plans. Neuron, 70, 536548.
Kornhuber, H. H., & Deecke, L. (1965). [Changes in the brain potential in voluntary movements
and passive movements in man: readiness potential and reafferent potentials]. Pflugers Archiv
fur die gesamte Physiologie des Menschen und der Tiere, 284, 117.
Kraskov, A., Dancause, N., Quallo, M. M., Shepherd, S., & Lemon, R. N. (2009). Corticospinal
neurons in macaque ventral premotor cortex with mirror properties: a potential mechanism for
action suppression? Neuron, 64, 922930.
Krasovsky, A., Gilron, R., Yeshurun, Y., & Mukamel, R. (2014). Differentiating Intended Sensory
Outcome from Underlying Motor Actions in the Human Brain. Journal of Neuroscience, 34,
1544615454.
Kunieda, T., Ikeda, A., Ohara, S., Matsumoto, R., Taki, W, et al. (2004). Role of lateral non-
primary motor cortex in humans as revealed by epicortical recording of Bereitschaftspotentials.
in relation to onset of cerebral activity (readiness-potential): the unconscious initiation of a
freely voluntary act. Brain, 106(Pt 3), 623642.
Lim, S. H., Dinner, D. S., Pillay, P. K., Luders, H., Morris, H. H., et al. (1994). Functional
anatomy of the human supplementary sensorimotor area: results of extraoperative electrical
stimulation. Electroencephalography and Clinical Neurophysiology, 91, 179193.
Page 16 of 18
Manes, F., & Leiguarda, R. (2005). Frontostriatal circuits and disorders of goal-directed actions
In H.-J. Freund, M. Jeannerod, M. Hallet, & R. Leiguarda (Eds.), Higher-order Motor Disorders
from neuroanatomy and neurobiology to clinical neurology (pp. 414439). New York: Oxford
University Press.
Marin, R. S. (1997). Differential diagnosis of apathy and related disorders of diminished

motivation. Psychiatric Annals, 27, 3033.
Marin, R. S., & Wilkosz, P. A. (2005). Disorders of diminished motivation (vol 20, pg 285, 2005).
Journal of Head Trauma Rehabilitation, 20, CP1CP1.
Mukamel, R., Ekstrom, A. D., Kaplan, J., Iacoboni, M., & Fried, I. (2010). Single-neuron
responses in humans during execution and observation of actions. Current Biology, 20, 750756.
Mukamel, R., & Fried, I. (2012). Human intracranial recordings and cognitive neuroscience.
Annual Review of Psychology, 63, 511537.
Neuper, C., Woertz, M., & Pfurtscheller, G. (2006). ERD/ERS patterns reflecting sensorimotor
activation and deactivation. Event-Related Dynamics of Brain Oscillations, 159, 211222.
Obeso, J. A., Rothwell, J. C., & Marsden, C. D. (1981). Simple tics in Gilles de la Tourette's
syndrome are not prefaced by a normal premovement EEG potential. Journal of Neurology,
Neurosurgery and Psychiatry, 44, 735738.
Ohara, S., Ikeda, A., Kunieda, T., Yazawa, S., Baba, K., et al. (2000). Movement-related change
of electrocorticographic activity in human supplementary motor area proper. Brain, 123(Pt 6),
12031215.
Pineda, J. A. (2005). The functional significance of mu rhythms: translating seeing and

hearing into doing. Brain Research Reviews, 50, 5768.
Rektor, I., Bares, M., Kanovsky, P., & Kukleta, M. (2001a). Intracerebral recording of readiness
potential induced by a complex motor task. Movement Disorders, 16, 698704.
Rektor, I., Bares, M., & Kubova, D. (2001b). Movement-related potentials in the basal ganglia: a
SEEG readiness potential study. Clinical Neurophysiology, 112, 21462153.
Rektor, I., Kanovsky, P., Bares, M., Louvel, J., & Lamarche, M. (2001c). Event-related potentials,
CNV, readiness potential, and movement accompanying potential recorded from posterior
thalamus in human subjects: a SEEG study. Neurophysiologie clinique / Clinical
Rektor, I., Sochurkova, D., & Bockova, M. (2006). Intracerebral ERD/ERS in voluntary movement
and in cognitive visuomotor task. Progress in Brain Research, 159, 311330.
Rizzolatti, G., & Sinigaglia, C. (2010). The functional role of the parieto-frontal mirror circuit:
interpretations and misinterpretations. Nature Reviews Neuroscience, 11, 264274.
Page 17 of 18
Rolls, E. T., & Deco, G. (2011). Prediction of decisions from noise in the brain before the
evidence is provided. Frontiers in Neuroscience, 5, 33.
Schurger, A., Sitt, J. D., & Dehaene, S. (2012). An accumulator model for spontaneous neural
activity prior to self-initiated movement. Proceedings of the National Academy of Sciences of the
U.S.A., 109, E2904E2913.
Shibasaki, H., & Hallett, M. (2006). What is the Bereitschaftspotential? Clinical

Sirigu, A., Daprati, E., Ciancia, S., Giraux, P., Nighoghossian, N., et al. (2004). Altered
awareness of voluntary action after damage to the parietal cortex. Nature Reviews
Yamamoto, J., Ikeda, A., Satow, T., Matsuhashi, M., Baba, K., et al. (2004). Human eye fields in
the frontal lobe as studied by epicortical recording of movement-related cortical potentials.
Brain, 127, 873887.
Yazawa, S., Ikeda, A., Kunieda, T., Ohara, S., Mima, T, et al. (2000). Human presupplementary
motor area is active before voluntary movement: subdural recording of Bereitschaftspotential
from medial frontal cortex. Experimental Brain Research, 131, 165177.
Page 18 of 18
Explicit and Implicit Beliefs, Attitudes, and Intentions

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
The Role of Conscious and Unconscious Processes in Human Behavior
Icek Ajzen
Nilanjana Dasgupta
DOI:10.1093/acprof:oso/9780190267278.003.0005

Willful acts are familiar experiences in daily life. Experiences such as having pancakes for
breakfast, making a doctors appointment, and watching the evening news on television are
perceived as volitional behaviors, which are the result of deliberate decisions or intentions by an
individual. Consistent with this intuition, many contemporary models in social psychology
incorporate the assumption that intentions are the proximal antecedent of human social
behavior. This chapter considers the nature and origins of behavioral intentions, the explicit and
implicit beliefs and attitudes on which they are based, as well as their causal effects on
behavior.
Keywords: intention, explicit attitudes, implicit attitudes, behavioral, social psychology
Willful acts are familiar experiences in daily life. Having pancakes for breakfast, making a
doctors appointment, and watching the evening news on television are perceived as volitional
behaviors, the result of deliberate decisions or intentions. Consistent with this intuition, many
contemporary models in social psychology incorporate the assumption that intentions are the
proximal antecedent of human social behavior. In the present chapter we consider the nature
and origins of behavioral intentions, the explicit and implicit beliefs and attitudes on which they
are based, as well as their causal effects on behavior.
Page 1 of 28
Reasoned Action
It is well recognized that people can be fully aware of only a small fraction of the many stimuli
vying for attention at any given moment and of the vast amount of information stored in
memory. However, attention can shift, and information can be recalled as needed to deal with
the task at hand. The mechanisms underlying these processes lie largely outside conscious
awareness; we only become aware of the resulting mental contents and states. When weighing
whether to go on a diet, for example, dormant beliefs and feelings associated with dieting can
become accessible to conscious awareness, prompt deliberation, and influence the decision.
Description of the ways in which (p.116) explicit beliefs and attitudes influence intentions and
actions is the domain of reasoned action models (see Fishbein & Ajzen, 2010); among them are
social cognitive theory (Bandura, 1986, 1997), the theory of subjective culture and interpersonal
relations (Triandis, 1972), the health belief model (Rosenstock, Strecher, & Becker, 1994), goal-
setting theory (Locke & Latham, 1994), the information-motivation-behavioral skills model
(Fisher & Fisher, 1992), and the technology acceptance model (Davis, 1989). Prominent among
these models is the theory of planned behavior (Ajzen, 1991, 2012), which is discussed in some
detail later in the chapter.
As the term implies, the hallmark of reasoned action models is their reliance on explicit beliefs
and attitudes as the basis of behavioral intentions leading to action. These models emphasize the
controlled aspects of human information processing and decision-making. Their concern is
primarily with behaviors that are goal-directed and steered by conscious self-regulatory
processes. In reasoned action models, behavioral intention is conceptualized as a predictive
process that precedes reasoned action, rather than a post-dictive inference that occurs after an
action has already occurred. In this regard, behavioral intention is better aligned with the
Comparator Model (Blakemore & Frith, 2003; Blakemore, Wolpert, & Frith, 2002) than Wegner
and colleagues (e.g., Wegner & Wheatley, 1999) post hoc inference account of mental causation
and behavior. According to the Comparator Model, the experience of agency over ones action
(intention to act) arises from internal motor representations that precede the action. A mental
representation of the sensory consequences of ones action is generated prior to the action,
which is compared with the actual sensory state after the action has been initiated. If the actual
sensory state matches the predicted one, it is understood to be self-caused. If there is a
mismatch, the action is understood to be externally caused.
Complementing the reasoned action approach, a great deal of research in recent years has
focused on implicit cognitions and their effects on behavior. The general theorizing behind this
line of work is the proposition that dormant beliefs, attitudes, intentions, and other constructs of
this kind can be activated while still remaining below conscious awareness, and that these
implicit reactions can have observable effects on judgments and actions. Consistent with this
idea, research has shown that behavioral intentions measured indirectly, using physiological
measures, occur well before individuals become consciously aware of their intentions and are
able to self-report the desire to act (e.g., Libet, Gleason, Wright, & Pearl, 1983). We consider
this type of evidence and its implications after discussing the role of conscious intentions as
determinants of behavior.
Page 2 of 28
(p.117) Theory of Planned Behavior

In this chapter, we are primarily concerned with volitional, goal-directed behaviors. Physical
activity to improve health and fitness, recycling to protect the environment, wearing a seatbelt
for increased safety, using condoms to prevent sexually transmitted diseases, and so forth, are
voluntary in nature, involving a measure of deliberation and planningat least when first
contemplated. Neither inborn reflexes nor unattended aspects of behavior (e.g., facial
expressions, hand gestures, seating distance, speed of walking) are, in this sense, goal-directed
or volitional and are therefore of secondary interest for our purposes. Like other reasoned action
models, the theory of planned behavior (TPB) (Ajzen, 1991, 2012) posits that the intention to
engage in a certain behavior is the proximal antecedent of voluntary action. It can be defined as
a behavioral disposition (Campbell, 1963), or a readiness to act in a certain way under
appropriate circumstances. Intentions vary in their degree of generality or specificity. At the
lowest level of generality are intentions to engage in a particular action in a given context and
time frame. The intention to get a flu shot at the local pharmacy next Tuesday on the way home
from work is an example of a very specific intention. Much less detailed in terms of context and
time is the general intention to get a flu shot. Whether general or specific, once formed, the
intention can be automatically activated by internal or external cues and thus prompt
performance of the relevant behavior (Bargh, Gollwitzer, Lee-Chai, Barndollar, & Troetschel,
2001; Gollwitzer, 1993). When applying the theory of planned behavior, it is assumed that
implicit or dormant intentions are accessible in memory, can be brought into conscious
awareness, and thus can be measured by means of self-reports.
As we discuss later in this chapter, many factors can influence the predictive validity of
measured intentions. One of these factors is incorporated in the TPB in that the causal effect of
intention is said to depend on the degree to which an individual has control over performance of
the behavior. Given sufficient control, people are expected to carry out their intentions as the
opportunity presents itself. When measures of actual control are unavailable, perceived
behavioral control can serve as a proxy to the extent that perceptions of control accurately
reflect actual control.
According to the TPB, formation of an intention to engage (or not engage) in a given behavior is
said to follow from three kinds of considerations. The first are beliefs about the likely
consequences of the behavior, termed behavioral beliefs. Depending on the subjective value of
these consequences, behavioral beliefs lead to the formation of a positive or negative attitude
toward performance of the behavior. The second consideration is normative in nature. Beliefs as
to what important others expect us to do or (p.118) are themselves doing (normative beliefs),
together with the motivation to comply with the normative referents, produce perceived social
pressure, or a subjective norm, to engage or not to engage in the behavior under consideration.
Finally, individuals are assumed to consider the presence of factors that can facilitate or inhibit
performance of the behavior, such as needed skills and opportunities, time and money,
cooperation by others, and so on. These control beliefs, together with the perceived power of the
control factors to facilitate or interfere with behavioral performance, are assumed to produce an
overall level of perceived control, or what Bandura (1986, 1997) has called self-efficacy. In the
TPB, attitudes, subjective norms, and perceptions of control are postulated jointly to influence
intentions, their relative importance varying as a function of the behavior and the population of
interest.
Page 3 of 28
Empirical support for the theory of planned behavior comes from tests of the model in a great
variety of behavioral domains. Meta-analyses of research findings have confirmed that indexes
composed of behavioral, normative, and control beliefs correlate, as expected, with direct (e.g.,
semantic differential) measures of attitudes, subjective norms, and perceptions of control; and
these variables account for a great deal of the variance in intentions (see Fishbein & Ajzen,
2010, for a review). Moreover, properly operationalized intentions are generally found to be
good predictors of behavior (Sheeran, 2002) and, confirming their causal effects, a meta-
analysis of 47 behavior-change interventions (Webb & Sheeran, 2006) showed that
experimentally induced changes in intentions (mean d = 0.66) are followed by corresponding
changes in later behavior, albeit of smaller magnitude (mean d = 0.36).
The Nature of Reasoned Action

The TPBs focus on controlled aspects of human behavior should not be taken to mean, however,
that people are assumed to be rational. A rational decision-making process would involve careful
review and unbiased evaluation of all available information relevant to the decision, and
dispassionate use of this information to arrive at the best possible decision consistent with
formal rules of logic. The account provided by the theory of planned behavior differs in
important respects from this construal.
Information-Processing Continuum
To start, there is no assumption in the TPB that individuals systematically assemble and
impartially process all relevant information whenever they are contemplating performance of a
behavior. Instead, consistent with popular (p.119) dual-mode processing approaches (see
Carver & Scheier, 1998; Chaiken & Trope, 1999; Petty & Cacioppo, 1986), the extent to which
people process information prior to forming an intention is assumed to depend on their
motivation and cognitive capacity, varying along a continuum from shallow to deep (see Ajzen &
Sexton, 1999). The intention to engage in a behavior is likely to be preceded by deliberate
review and consideration of available information to the extent that the behavior is of
importance and has rarely been performed before. Buying a home, getting married, joining the
military, and quitting ones job are examples of what are, for most people, important, infrequent
decisions. For decisions of this kind, individuals will, according to the TPB, consider the likely
consequences of the behavior, the normative expectations of significant others, and the
availability of requisite resources, as well as possible impediments to performance of the
behavior. In contrast, such everyday behaviors as brushing ones teeth, taking a shower, going
to work, or reading the morning newspaper are assumed to become routine and to be performed
without much prior deliberation. Attitudes, subjective norms, perceptions of control, and
intentions are assumed to guide these kinds of behaviors with little awareness and deliberation;
these attitudinal and normative influences on behavior are often considered implicit or
automatic.
Cognitive Heuristics, Errors and Biases

In the theory of planned behavior, beliefs are the building blocks for the formation of attitudes
toward a behavior, subjective norms, the perceptions of behavioral control, and ultimately
behavioral intentions. They represent the information that people have in relation to the
behavior under consideration. More often than not, our beliefs reflect reality reasonably well
(Jussim, 2012). This could hardly be otherwise, for if they did not, we would not have survived as
Page 4 of 28
a species. However, people do not act like intuitive scientists in the way they arrive at their
beliefs or draw inferences from them; indeed, their cognitive shortcomings are well
documented. Use of cognitive heuristics can produce systematic errors of judgment, and
motivational biases can lead to the formation of unrealistic or even delusional beliefs (see Fiske
& Taylor, 1991; Kruglanski & Ajzen, 1983; Nisbett & Ross, 1980; Tversky & Kahneman, 1974). A
staggering number of cognitive and motivational biases have been identified over the years:
acquiescence bias, false consensus, in-group bias, just world hypothesis, self-serving bias,
unrealistic optimism, expectancy bias, illusory correlation, hindsight bias, and many more (see
Jussim, 2012).
There is nothing in the TPB to contradict these observations. The theory makes no assumptions
about the ways in which beliefs are formed, or about (p.120) the objectivity or veridicality of
those beliefs. All it stipulates is that peoples intentions and behaviors take account of, and are
consistent with, their beliefs, no matter how the beliefs originated. It is in this sense of internal
consistency, and only in this sense, that behavior is considered to be reasoned.
Habits and Automaticity

As noted earlier, behaviors enacted frequently in the same context can become routine and no
longer require a conscious decision to be enacted. Further, after sufficient practice, some
behaviors are performed automatically without much conscious awareness. A good example is
fastening ones seatbelt before or after starting the car. However, behaviors of interest to
psychologists are rarely such simple unitary acts that can easily habituate. Consider, for
example, eating at a restaurant. This common event typically involves a sequence of decisions
and actions, such as picking a day and time, selecting a restaurant, making a reservation,
walking or driving to the restaurant, waiting to be seated, placing the order, and eating, paying,
and leaving the restaurant. Some aspects of this familiar behavioral script (Abelson, 1981),
perhaps waiting to be seated, require little cognitive effort. Others, however, demand more
mindful attention. Before placing their orders, people usually review the menu and take note of
any specials offered; and before paying, they generally calculate an appropriate tip to leave for
the waiter. These aspects of the script require some measure of awareness and conscious
decision-making. It is doubtful, therefore, that even relatively common human behaviors can be
described as purely automatic, performed without any conscious awareness or intention. Even
when intentions fall below the level of awareness and become implicit, routine behaviors should
still be consistent with the original intentions. Consider, for example, two individuals, one
regularly taking the bus to work, the other invariably driving a car. Once habituated through
repetition, these two individuals do not need to consciously bring to mind their beliefs, attitudes,
and intentions to prompt the usual behavior. However, there is no reason to assume that the
behavioral choices of these two individuals will be inconsistent with their implicit intentions. If
asked, they could, with minimal cognitive effort, retrieve their beliefs, attitudes, and intentions,
and it is very likely that the first individual would indicate an intention to take the bus to work,
while the second would report an intention to go by car.
Yet some investigators (e.g., Aarts & Dijksterhuis, 2000; Gollwitzer, 1999; Neal, Wood, & Quinn,
2006; Ouellette & Wood, 1998; Verplanken & Aarts, 1999) have proposed that once a behavior
has habituated (after being performed repeatedly in the same context), it comes under the direct
control of internal or external cues that activate the behavior automatically. As a result, (p.121)
Page 5 of 28
intentions are assumed to become increasingly irrelevant as a behavior habituates. This

reasoning led to the hypothesis that intentions predict relatively novel or unfamiliar behaviors,
or when circumstances have changed, but that their predictive validity is less for behaviors that
have become routine and when the context remains unchanged.
However, empirical evidence does not seem to support this hypothesis. Ouellette and Wood
(1998) and Sheeran and Sutton (unpublished research) performed meta-analyses of studies
across different behavioral domains, classifying each behavior as one that can be performed
frequently (e.g., using a seatbelt, drinking coffee) or infrequently (e.g., getting a flu shot,
donating blood). Habituation was considered to be more likely for the former than the latter
behaviors. To be sure, the relative contribution of frequency of past behavioral performance to
the prediction of future behavior was found to be greater for high- than for low-opportunity
behaviors. This finding reflects the effect of habituation. However, the correlation between
intention and future behavior remained about the same, showing that the predictive validity of
intentions is undiminished as we go from low- to high-opportunity behaviors. Indeed, contrary to
the habit hypothesis, in the meta-analyses of Ouellette and Wood and of Sheeran and Sutton,
prediction of high-opportunity behaviors from explicit measures of intention was about as
accurate as prediction of low-opportunity behaviors (mean r = .59 and r = .67, respectively, in
Ouellette and Woods meta-analyses [difference not significant] and mean r = .51 and r = .53 in
the Sheeran and Sutton analysis).
Nor is there evidence to support the related hypothesis that, independent of frequency of
performance, intentions are better predictors of behaviors that are performed in an unstable as
opposed to a stable context; the predictive validity of intentions was found to be approximately
the same in both contexts (Ouellette & Wood, 1998) or, contrary to the habit hypothesis,
somewhat better in stable contexts (Sheeran & Sutton, unpublished study; see Fishbein and
Ajzen 2010, pp. 5153, for a discussion). These findings suggest that even when people have had
many opportunities to perform a behavior in a stable context, intentionseven if they have
become implicit in the momentcan be brought to mind, explicitly reported, and retain their
predictive validity.
The Intention-Behavior Gap

It is well known that people do not necessarily act in accordance with their expressed intentions.
The observation that intentions are not always carried out has a long history. It is captured in
such adages as its easier said than done (p.122) or the road to hell is paved with good
intentions, and it was documented very early in empirical research on the attitude-behavior
relation (e.g., LaPiere, 1934; Linn, 1965). With respect to health-related behaviors, Sheeran
(2002) reported that among individuals who express an intention to use condoms, to get a
cancer screening, or to exercise, only about 50% actually do so. Similarly, as noted earlier,
experimentally induced large changes in intentions are followed by much smaller changes in
behavior (Webb & Sheeran, 2006). Findings of this kind have led Sheeran, Gollwitzer, and Bargh
(2013) to conclude that conscious processes are insufficient to explain health-related behaviors,
and that the gap between intentions and behavior could be due to unconscious processes.
While the role of unconscious processes in human behavior is undeniable, and is discussed in
greater detail later in this chapter, it is also the case that observed discrepancies between
Page 6 of 28
conscious intentions and actions can be due to many reasons other than unconscious influences
on behavior. For example, expressed intentions may be biased by self-presentation concerns, as
when people profess that they will go on a diet to lose weight, quit smoking, or hire workers
with disabilities without actually intending to do so. But even when honestly expressed, people
may fail to carry out their intentions for any number of reasons, articulated in the following.
Forgetting
A person may truly intend to return a book on time to the library, yet may forget to do so.
Research suggests that such failures of prospective memory are more likely when specific
aspects of intentions, such as where, when, and how to carry out the behavior, are not encoded
in memory, or when one or more of these situational cues is absent (see Brandimante, Einstein,
& McDaniel, 1996). These conclusions are consistent with findings that people are more likely to
act on their intentions if they form a specific plan (Schifter & Ajzen, 1985) or implementation
intention (Gollwitzer, 1999; Sheeran & Orbell, 2000) as to where, when, and how they will
carry out their intentions.
Instability of Intentions
Clearly, people are free to change their intentions. It stands to reason that if changes in
intentions occur after they have been assessed but prior to observation of the behavior,
predictive validity will suffer. Indirect support comes from research in which the time interval
between measurement of intention and observation of behavior was taken as a proxy for change
in intentionswith the passage of time, an increasing number of events can cause intentions to
change. Consistent with this line of reasoning, meta-analyses of research in (p.123) different
behavioral domains have shown that the correlation between intentions and behavior tends to
decline with the passage of time (Albarracn, Johnson, Fishbein, & Muellerleile, 2001; Randall &
Wolff, 1994; Sheeran & Orbell, 1998).
Perhaps less obvious, people may differ in the extent to which their intentions are stable over
time (even when the time interval is held constant). Russell Fazio and his associates (Fazio,
1990; Fazio & Zanna, 1978) have demonstrated the importance of attitude strength, indicated by
response latency, as a moderator of the attitude-behavior relation. Compared to weak attitudes,
strong attitudesproduced by direct experience or repeated expression of the attitude (Fazio &
Zanna, 1981; M. C. Powell & Fazio, 1984)are more stable over time, more resistant to
persuasion, and better predictors of behavior (see Krosnick & Petty, 1995). Just as attitudes vary
in strength, so too do intentions. Strong intentions, as measured by fast response latencies, tend
to be more stable (Doll & Ajzen, 1992). We would therefore expect that people who have formed
strong, stable intentions are more likely to act in accordance with those intentions than are
people with relatively unstable intentions.
In their research on the temporal stability of behavioral intentions, Conner, Sheeran, and their
associates (Conner, Sheeran, Norman, & Armitage, 2000; Sheeran, Orbell, & Trafimow, 1999)
asked participants to express their intentions on two separate occasions. Responses to the
second intention measure were used to predict subsequent behavior. Supporting the moderating
role of intention stability, in both investigations the correlation between intentions and behavior
was found to be significantly stronger among participants with relatively stable, as opposed to
unstable, intentions.
Page 7 of 28
Conflicting Intentions
People sometimes intend to attain goals that are in conflict with each other. In those instances,
assessing their intentions to achieve one goal but not assessing intentions associated with the
other goal can produce relatively low-intention behavior correlations. An interesting case in
point comes from a program of research on dieting to lose weight (Stroebe, van
Koningsbruggen, Papies, & Aarts, 2013). As is well known, most people fail to adhere to an
intended diet and thus fail to lose weight or, after initial success, regain their weight in short
order (Mann et al., 2007; Powell, Calvin, & Calvin, 2007). To explain the failure of weight-loss
intentions to result in actual weight loss, Stroebe and his associates proposed a goal-conflict
model of eating behavior. The two conflicting intentions in the model are the intention to control
ones weight and the intention to enjoy ones food. According to this model, dieters often
encounter enticing food cues that prime their intention to enjoy their food, produce (p.124)
preferential processing of palatable food stimuli, and inhibit cognitive activation of the
competing intention to control their weight.
Lack of Control
As noted in the description of the theory of planned behavior, intentions are expected to lead to
the corresponding behavior only to the extent that people have sufficient behavioral control.
Lack of requisite resources, such as knowledge, physical stamina, time, and money, as well as
unanticipated obstacles or lack of needed cooperation by others, can prevent people from acting
on their intentions (see Ajzen, 2005, for a discussion of internal and external control factors).
We are unaware of studies that have examined the effect of objective control factors on the
predictive validity of intentions. However, as noted in the description of the TPB, perceived
behavioral control is often used as a proxy for actual control under the assumption that
perceptions of control reflect actual control reasonably well. We therefore expect good
intention-behavior correspondence only when perceived control is relatively high. Support for
this proposition comes from research regarding the effect of self-efficacy beliefs (i.e., perceived
behavioral control) on behavior, in particular studies in which self-efficacy was experimentally
manipulated (see Bandura and Locke, 2003, for a review). In these studies, people who intended
to perform a behavior of interest, or to achieve a certain goal, and who were led to believe that
they had a high level of efficacy, that is, that they had control over performance of the behavior
or over the attainment of the goal, were more likely to act on their intentions than were people
who were led to believe that their level of control was low. The former were more likely to
persevere and to work harder at a task and, thus, they were more likely to obtain a desired
outcome.
In a study of restrained eating, Papies, Stroebe, and Aarts (2008) provided evidence for the
effect of perceived control on the ability of dieting intentions to predict eating behavior. Their
data showed that, among people with a weight-loss goal, intentions to avoid pizza, chocolate,
French fries, cookies, and chips predicted actual avoidance of these foods much better when the
participants had a high (r = .90) as compared to a low (r = .27) level of perceived control over
losing weight1 (see also Schifter & Ajzen, 1985).
Page 8 of 28
Context Incongruity
One final issue to be discussed in relation to the intention-behavior gap has to do with belief
accessibility. According to the TPB, intentions and behavior are, in the final analysis, based on
readily accessible behavioral, normative, (p.125) and control beliefs. However, contextual
factors can have strong effects on the number and kind of beliefs that become readily accessible
in the moment (Eitam & Higgins, 2010; Gold, 1993; Schuman & Presser, 1981; Schwarz, 1999).
Furthermore, intentions are normally assessed in a hypothetical context that differs
considerably from the real context in which behavior is observed. It follows that the behavioral,
normative, and control beliefs that are accessible when intentions are assessed by referencing a
hypothetical situation may well differ from the beliefs that become accessible when the behavior
is to be performed in a real situation. We can expect strong intention-behavior correlations only
when the hypothetical and real contexts activate the same beliefs, or beliefs of equivalent
valence, in relation to the behavior of interest (Ajzen & Sexton, 1999).
LaPieres (1934) well-known study on the attitude-behavior relation was designed to

demonstrate the discrepancy between attitude expression in a hypothetical context and the
constraints of a real situation. Responses to a (hypothetical) inquiry by LaPiere indicated that
members of the Chinese race would not be accepted at restaurants, hotels, and inns across the
United States, but in prior visits to the same establishments, a Chinese couple of considerable
means experienced no such discrimination in reality. It is likely that the hypothetical context
activated beliefs about members of the Chinese race that, at that time, differed considerably
from the beliefs activated in the presence of a real Chinese couple who appeared affluent, well
spoken, and well dressed. A similar conclusion was reached by Linn (1965) in a study of racial
attitudes and behavior in which female participants were first asked to indicate their readiness
(intention) to permit use of their photograph with a black man in support of efforts at racial
integration, followed by a request to actually provide their signed permission. In post-
experimental interviews, participants indicated that when confronted with the actual decision to
sign releases of their photographs, beliefs had come to mind that had not been activated when
they had earlier considered this issue in a hypothetical context.
Direct support for the hypothesis that the antecedents of intentions in the TPB can differ in
hypothetical compared to real behavioral contexts comes from a study on willingness to pay for
a public good (Ajzen, Brown, & Carvajal, 2004). In one part of the study, students in small
groups were asked to vote on a referendum to contribute $8 to a university scholarship fund. In
one ballot, they were told that the vote was hypothetical: that even if the majority voted in favor,
they would not actually have to pay the money, but that they should vote as if it were real. In a
second ballot, they were led to believe that everybody actually would have to pay $8 into the
fund if the majority voted yes. As is usually found in studies of this kind, a much larger
percentage voted in favor of the referendum in the hypothetical situation (70%) than in the real
situation (41%), (p.126) a discrepancy known as hypothetical bias (e.g., Blumenschein,
Johannesson, Blomquist, Liljas, & OConnor, 1998). Prior to casting their votes, the same
participants had completed a TPB questionnaire with respect to voting in favor of the
referendum. As expected, attitudes, subjective norms, perceptions of control, and intentions
regarding a yes vote were significantly more favorable in the hypothetical than in the real
voting context.
Page 9 of 28
To mitigate this hypothetical bias, a corrective entreaty was introduced in a second part of the
study, which exhorted participants in the hypothetical condition to examine carefully how they
would vote if it were real and to consider other possible uses of the money. Following this
entreaty, attitudes, subjective norms, perceptions of control, and intentions in the hypothetical
situation were no more favorable than in the real situation, and the hypothetical bias
disappeared: participants were no more likely to vote yes on the referendum in the
hypothetical than in the real context.
Internal and External Cues Affecting Behavior

Peoples beliefs are based on information obtained from many different sources: direct
experience with the object of the belief; information conveyed by parents, teachers, and friends;
exposure to books, newspapers, and mass media; and so forth. In addition, people often draw
far-reaching inferences that go beyond the basic information provided by direct experience or
external sources (see Fishbein & Ajzen, 2010, for a discussion). As we shall see in the following,
some of these beliefs are acquired without awareness, and in other cases their origins are lost to
memory. Moreover, just as people can be unaware of the source of their beliefs and attitudes, so
too can they be unaware of internal or situational cues that activate their beliefs and attitudes.
Attentional Bias
Research has shown that people preferentially attend to motivationally significant stimuli, and
that this attentional bias can influence behavior. As noted earlier, compared to unrestrained
eaters, people on a diet to lose weight are more likely to attend to attractive food items
following exposure to food cues, and this attentional bias reduces their ability to adhere to their
diets (Papies, Stroebe, & Aarts, 2009). Similarly, alcohol consumption was found to be predicted
from alcohol-related attentional bias (Fadardi & Cox, 2008), and physical activity was shown to
correlate with greater attention to exercise-related stimuli (Calitria, Lowe, Eves, & Bennett,
2009). Shifts in attentional bias are assumed to occur automatically, outside awareness. These
studies are therefore (p.127) interpreted as evidence that unconscious motivational processes
exert their influence by directing attention preferentially to certain kinds of cues, and these cues
in turn automatically activate the behavior (see Eitam & Higgins, 2010; Sheeran et al., 2013).
However, one study in this line of research (Calitria et al., 2009) has shown that attentional bias,
even if outside awareness, does not affect behavior automatically. The investigators in this study
assessed attention to exercise cues and also participants implicit and explicit attitudes toward
exercising. The relation between attention to exercise cues and self-reported physical activity
was found to be moderated by participants explicit attitudes toward exercising, such that higher
attentional bias toward exercise cues was associated with higher levels of physical activity only
for participants who had positive explicit attitudes toward exercising. This finding suggests that
biased attention to certain stimuli increases the likelihood of relevant behavior only if the
attention is associated with approach-oriented motivation in relation to the behavior (but see
Eitam & Higgins, 2010, for a different interpretation).
Unconscious Activation of Goals and Behaviors

Research has demonstrated that unconscious activation of attitudes and beliefs about social
groups and other psychological constructs (e.g., the elderly, African Americans, rudeness,
silence) can produce construct-relevant behavior (Aarts & Dijksterhuis, 2003; Bargh, Chen, &
Page 10 of 28
Burrows, 1996). Similarly, activation of goals (e.g., achievement) without conscious awareness
can motivate goal pursuit (Bargh, Gollwitzer, Lee-Chai, Barndollar, & Troetschel, 2001; Hassin,
Aarts, Eitam, Custers, & Kleiman, 2009; Kruglanski et al., 2002). For example, participants
primed with stereotypes of the elderly subsequently walked more slowly compared to others
who were not primed with elderly stereotypes, and participants primed with the concept of
rudeness were more likely to interrupt the experimenter than participants primed with the
concept of politeness (Bargh et al., 1996). Similarly, priming the concept of silence induced
participants to speak softly, and priming the concept of exclusivity increased the likelihood that
participants would remove crumbs after eating a biscuit (Aarts & Dijksterhuis, 2003). Finally,
unconscious priming of an achievement goal improved subsequent performance on a word-
search task (Bargh et al., 2001).
Findings such as these are typically attributed to automatic enactment of behavior made
accessible by priming a behavior-relevant construct (ideomotor expression; see Stock & Stock,
2004). The question, however, is whether or not these construct-to-behavior effects really occur
without mediation (p.128) by implicit or explicit cognitions. Several theoretical articles have
questioned the assumption of automatic behavior activation (see Blair, 2002; Eitam & Higgins,
2010; Loersch & Payne, 2011), and empirical research provides evidence of cognitive mediation
linking unconscious goal activation to behavior. For example, according to Cesario, Plaks, and
Higgins (2006), priming a construct activates implicit preparatory responses, such as implicit
attitudes, and these implicit responses mediate the effect of the prime on behavior. Consistent
with this proposition, they found, as in previous research, that priming the elderly stereotype
slowed walking speed, but only for participants with positive implicit attitudes toward the
elderly; it increased walking speed for participants with negative implicit attitudes. Also
inconsistent with the assumption of automatic activation of behavior as a result of goal priming,
Klein et al. (2012) reported an experimenter expectancy effect in the experimental paradigm
involving elderly stereotypes. Participants primed with the stereotype of the elderly were found
to reduce their speed of walking only when experimenters expected them to do so, but not when
experimenters expected them to increase their walking speed. This suggests that participants
were sensitive to cues associated with the experimenters expectations, and that these
perceptions mediated the effect of the prime on behavior.
The Question of Awareness

Relying on the theory of planned behavior, we have, up to this point, dealt primarily with the
role of conscious beliefs, attitudes, and intentions, even though we have acknowledged that, in
the case of well-rehearsed behaviors, these variables can become implicit and can be
automatically activated without much conscious effort. However, in the latter part of the
twentieth century, research began to focus on the possibility that people may acquire and hold
beliefs and attitudes outside awareness, and that these beliefs and attitudes may have powerful
effects on their intentions and actions. This type of theorizing, which complements the TPB and
extends beyond it, was initially prompted by the cognitive revolution in psychology that resulted
in the development of new theories and methods of measuring implicit memory (Jacoby, 1991;
Roediger & McDermott, 1993; Schacter, 1987) and semantic associations (Meyer &
Schvanaveldt, 1971; Neely, 1977; and Posner & Snyder, 2004). These cognitive theories and
methods were adapted by social psychologists to investigate implicit social cognition (for more
Page 11 of 28
on the history of implicit social cognition see Banaji, 2001; Bazerman & Banaji, 2004; Greenwald
et al., 2002).
(p.129) Attitudes and Beliefs Formed and Expressed Without Awareness

Because implicit attitudes and beliefs are conceptualized as spontaneous mental associations,
they are typically measured using techniques that bypass respondents deliberation and
introspection. These techniques do not ask individuals to self-report their attitudes and beliefs,
but instead rely on the speed or accuracy of their responses to different categories of stimuli,
with the goal of capturing underlying associations indirectly (Gawronski & Payne, 2010). In
comparison, explicit measures, such as those used to assess the constructs in the TPB, rely on
participants responses to direct questions about their beliefs, attitudes, and intentions in the
form of self-reports on multiple-choice items, feeling thermometers, semantic-differential scales,
or structured interviews. A burgeoning body of empirical evidence confirms the existence of
implicit beliefs and attitudes in addition to explicit beliefs and attitudes (Cunningham, Preacher,
& Banaji, 2001; Hofmann, Gschwendner, Nosek, & Schmitt, 2005; Hofmann, Gawronski,
Gschwendner, Le, & Schmitt, 2005; Nosek, 2005; Nosek & Smyth, 2007).
Most theoretical models consider implicit and explicit attitudes to be two correlated but
conceptually distinct systems of psychological processing (Epstein, 1991; Gawronski &
Bodenhausen, 2006; Greenwald & Banaji, 1995; Greenwald et al., 2002; Kahneman & Frederick,
2005; Sloman, 1996; Strack & Deutsch, 2004; but see Keren & Schul, 2009). Information
processing in the controlled system is relatively effortful and slow, relying on symbolic
representations and reasoning. The processes described in the theory of planned behavior fall
squarely within this mode of operation. Information processing in the spontaneous system is
relatively fast and effortless, characterized by associative connections and broad
generalizations.
As a case in point, one modelthe Associative-Propositional Evaluation (APE) modelcontends

that attitudes arise from two different processes, one associative and the other propositional in
nature (Gawronski & Bodenhausen, 2006). Associative processesthe primary basis of implicit
attitudes and beliefsare simple, spontaneous reactions that occur in response to a stimulus
based on a previously learned association between that stimulus and an attribute (which may be
a trait or simply a good/bad evaluation). These stimulus-attribute associations are often learned
because they co-occurred in time and the co-occurrence is repeated with some frequency.
Several experiments have shown that attitudes and beliefs can be learned automatically when an
object and an attribute are repeatedly paired with each other, consistent with classical
conditioning principles (e.g., Krosnick, Betz, Jussim, & Lynn, 1992; Olson & Fazio, 2001; Seger,
1994; Staats & Staats, 1958). In some studies, (p.130) these repeated pairings involve
subliminal presentations where the perceiver is not consciously aware of the pairings but
nevertheless learns the attitude or belief as a result of such exposure (Rydell et al., 2006;
Zajonc, 1980).
Once learned, the presentation of the stimulus automatically activates the associated attribute
or evaluation. These automatic activations can occur outside awareness and require little
cognitive capacity. Importantly, the likelihood of an association being activated is independent
of its perceived truth value, that is, associations can be activated even when the person
Page 12 of 28
considers them invalid (but see Eitam & Higgins, 2010). For example, a spontaneous negative
attitude may pop into mind when a perceiver sees someone who is African American, even if the
perceiver consciously rejects that negative attitude (Devine, 1989; Nosek, Banaji & Greenwald,
2002).
However, when people are asked directly about their attitudes toward African Americans, the
APE model proposes that an entirely different process is set in motion (Gawronski &
Bodenhausen, 2006). In this case, people engage in deliberative inferential processes similar to
those described in the TPB by considering information that they regard as relevant to their
racial attitudes and beliefs. This information may reflect specific exemplars of the group African
Americans (e.g., I like President Obama); it may also include other considerations, such as
ones values (e.g., I should judge people as individuals, not based on their race) or self-
presentational concerns (e.g., I dont want other people to think I am racist). People might
even consider information based on their spontaneous reactions if they are aware of them (e.g.,
I sometimes feel uncomfortable around Black people). The most important aspect of the
explicit inferential process is a determination of which thoughts and feelings are considered
valid and which are considered invalid for the judgment at hand. The end result is an explicit
attitude or belief based on a set of information that the individual considers valid. As implied in
the foregoing example, a determination that ones spontaneous reactions are invalid will lead to
the exclusion of these reactions from the explicit attitude report, resulting in a discrepancy
between the implicit and explicit attitude measures. However, when ones spontaneous reaction
is consistent with other information one considers valid, that reaction will be integrated into the
explicit attitude, and there is a higher likelihood that implicit and explicit attitude measures will
correspond. In sum, explicit attitudes result from considering various pieces of information that
come to mind, weighing them against each other, and creating consistency among them. Implicit
attitude is one piece of information that plays a variable role in this process. Its effects depend
on individuals awareness of it and whether they consider it a valid piece of information to
include in their explicit attitude reports.
(p.131) When Do Implicit Beliefs and Attitudes Predict Behavior?

In the past 15 years, implicit attitudes and beliefs regarding a wide variety of objects have been
the focus of study. They include implicit attitudes and beliefs about social groups (race, gender,
age, sexual orientation, political parties), political candidates, academic disciplines, consumer
products, food, controlled substances (alcohol, drugs, tobacco), clinical disorders, and self-
concept (personality, relationship style, self-esteem). These implicit attitudes and beliefs have
been associated with various decisions and behaviors, including interpersonal behavior and
judgments, employment decisions, political choices, academic behavior, consumer choices, use
of controlled substances, clinical behavior, medical diagnoses, and other health behaviors (for
reviews see Blair, Dasgupta, & Glaser, 2014; Dasgupta, 2004; Greenwald et al., 2009; Jost et al.,
2009). A recent meta-analysis of 122 research reports (184 independent samples) revealed that
in many cases both implicit and explicit responses incrementally predict behavioral outcomes
controlling for the other. In a few cases, implicit attitudes and beliefs predict behavior better
than their explicit counterparts, whereas in many other cases, explicit responses predict
behavior better than their implicit counterparts (see Greenwald et al., 2009, Figure 2). In the
following we consider the conditions under which implicit attitudes and beliefs are more (vs.
less) likely to predict behavior than their explicit counterparts.
Page 13 of 28
Controlled Versus Automatic Behaviors

People are aware of, and can control, many of their behaviors, but some aspects of behavior
remain outside awareness or are difficult to control. In the early days of implicit social cognition
research, the working hypothesis was that controllable behavior would be better predicted from
explicit beliefs and attitudes, whereas automatic behavior would be better predicted from
implicit beliefs and attitudes (e.g., Dovidio et al., 2002; Fazio et al., 1995; McConnell & Leibold,
2001). In these studies, for example, the data showed that implicit racial attitudes better
predicted spontaneous (less controllable) nonverbal friendliness in interracial interactions,
whereas explicit racial attitudes better predicted deliberate (more controllable) interracial
behavior such as verbal statements. Similarly, in the health domain, some research found that
clinicians implicit racial attitudes better predicted their interpersonal behavior with Black
patients (Blair et al., 2013; Cooper et al., 2012; Penner et al., 2010) than their deliberative
medical decisions (Haider et al., 2011; Sabin et al., 2008; Sabin & Greenwald, 2012).
(p.132) However, as this literature has grown, it has become clear that the early hypothesis is
untenable. Of relevance for our present discussion, implicit attitudes and beliefs were found to
predict not only spontaneous but also controlled behavior (decisions, choices, and judgments;
for reviews see Dasgupta, 2004; Greenwald et al., 2009). For example, in the medical domain,
one study (Green et al., 2007) found that doctors implicit racial attitudes predicted differential
medical diagnostic tests they recommended for Black compared to White patients presenting the
same clinical symptoms, such that more high-quality tests were recommended for White
compared to Black patients. Similarly, in employment settings, people who harbored implicit
bias against racial and ethnic groups, implicit bias against obese people, and implicit gender
stereotypes were less likely to hire members of the stereotyped group despite their
qualifications (Agerstrom & Rooth, 2011; Rooth, 2010; Rudman & Glick, 2001; Yogeeswaran &
Dasgupta, 2010). In academic settings, elementary schoolteachers implicit attitudes toward
ethnic minorities were associated with their differential expectations of minority versus majority
children in their classrooms (van den Bergh et al., 2010). In all these cases, the common theme
is that implicit attitudes and stereotypes predicted behaviors and judgments that were clearly
consciously controllable. It remains an open question as to the conditions under which implicit
attitudes and beliefs will better predict behaviors that are relatively automatic compared to
others that are relatively more controlled.
Cognitive Depletion
When cognitive resources are depleted, peoples implicit beliefs and attitudes better predict
their health-related behavior than explicit beliefs and attitudes. For example, in a series of
studies, Friese, Hofmann and colleagues (Friese, Hofmann & Wanke, 2008; Hofmann & Friese,
2008) manipulated participants cognitive resources by increasing the demands of a secondary
task, depleting self-regulation resources, or increasing alcohol intake. When participants were
low in resources, their consumption of potato chips, candy, or beer was better predicted by their
implicit than their explicit health attitudes. When cognitive resources were not so constrained,
these same behaviors were guided more by participants explicit than their implicit attitudes.
Note that in all conditions the measured behavior was exactly the same, but the ability to control
ones behavior was manipulated by varying inner cognitive resources.
Page 14 of 28
(p.133) Social Desirability

For decades, social scientists have known that social desirability bias is a critical reason that
explicit attitudes and beliefs do not always predict behavioral outcomes in socially sensitive
domains. Concerns about others perception of oneself can prevent honest self-reporting of ones
attitudes and beliefs (Crosby, Bromley, & Saxe, 1980; Sigall & Page, 1971, 1972). Implicit
attitude measures can help overcome social desirability biases by assessing the strength of
attitudes and beliefs using tasks that bypass self-report. For example, a large meta-analysis
(Greenwald et al., 2009; also see Blair et al., 2014) found that when it comes to attitudes and
beliefs about racial and ethnic groups, religious groups, the elderly, and so on, peoples implicit
attitudes and beliefs about these groups better predict their behavior toward group members
than their explicit attitudes and beliefs. Relatedly, the higher the likelihood of social desirability
bias in a given attitude domain, the stronger was the relation between respondents implicit
attitudes and behavior, controlling for their explicit attitudes.
Uncertainty
Feelings of uncertainty also influence the effect of implicit attitudes on behavior. In the realm of
political behavior, right before an election some voters inevitably report that they have not yet
decided which political candidate to vote for. Consistent with the argument that uncertainty
allows implicit attitudes to play a stronger role in behavior, Galdi and colleagues (Galdi, Arcuri,
& Gawronski, 2008; Galdi et al., 2012) found that among voters who were undecided one week
before an election, their ultimate vote was predicted by their implicit but not by their explicit
attitudes, whereas for voters who were decided before the election, their vote was better
predicted by their explicit attitudes.
Behavioral Context
Implicit stereotypic beliefs appear to predict peoples behavior when the social context activates
a relevant stereotype. For example, Yogeeswaran and Dasgupta (2010) found that people who
harbored an implicit stereotype that real Americans are White were less likely to recommend
hiring a non-White job candidate (specifically an Asian American) for a national security job, but
this implicit stereotype did not influence hiring decisions for a virtually identical corporate job. A
subsequent study confirmed that the relation between implicit stereotype about who is
legitimately American and hiring bias in national security was mediated by participants doubts
about Asian (p.134) Americans loyalty to the United States. In a conceptually similar manner,
Ziegert and Hanges (2005) found that implicit racial attitudes predicted hiring discrimination
only when participants had received information that suggested the company encouraged
decisions based on race.
Taken together, research on implicit social cognition shows that peoples judgments, decisions,
and behaviors can be influenced by factors that lie outside their awareness. In many of these
studies (with the possible exception of studies involving nonverbal behavior) people are clearly
aware of engaging in a particular behavior, but they are unaware that their behavior is molded
by implicit attitudes and beliefs. Given their awareness of the behavior, it is logical to infer that
in most cases people consciously intend to engage in that particular behavior (that is, they
intend to hire a new employee, provide a medical diagnosis, interact with a patient or a fellow
student), but they probably do not intend that behavior to be shaped by implicit attitudes and
beliefs about which they are unaware.
Page 15 of 28
Preconscious Intentions
In the preceding discussion we have tried to show that implicit beliefs and attitudes can predict
behavior independent of, or in combination with, explicit beliefs and attitudes. We now consider
evidence for the existence of preconscious intentions to behave in a certain way (Chen & Bargh,
1999; Libet et al., 1983; Miller & Maner, 2011; Miller, Zielaskowski, Maner, & Plant, 2012;
Ozono, Watabe, & Yoshikawa, 2012). In a series of studies, Libet and colleagues investigated the
link between behavioral intention, awareness of ones intention, and action. They measured
cortical EEG from participants while they were engaged in a task in which they moved their
fingers. Participants were asked to indicate the time at which they became aware of their
intention to move their fingers and compared the time-course of self-reported intentions to the
observed action and to EEG signals. Results indicated that self-reported intentions consistently
preceded the actual behavior by 300 ms. But more important for our purpose, participants EEG
responses showed a consistent negative potential arising from the supplementary motor area
well before the self-reported intention, preceding it by 1000 ms or more. From these data Libet
concluded that the brain initiates or prepares to act well before there is any reportable
subjective awareness that such a decision has taken place.
In other research, preconscious intentions can be inferred by measuring the speed with which
people approach or avoid stimuli. For example, Chen and Bargh (1999) asked participants to
respond to positive and negative stimuli shown on a computer screen by pushing or pulling a
lever. Sometimes participants were asked to pull the lever toward them for (p.135) positive
stimuli and push it away for negative stimuli. For other trials, task instructions were reversed.
Results showed that participants were faster at pushing the lever away from themselves than
pulling it toward themselves for negative stimuli, suggesting that these stimuli activated
preconscious avoidance tendencies. In contrast, participants were faster at pulling the lever
toward them than pushing it away for positive stimuli, suggesting that positive stimuli activated
approach tendencies. The differential speed of pulling versus pushing in this experiment is
suggestive of approach versus avoidance intentions, respectively.
Another study illustrates that specific emotions activate theoretically meaningful preconscious
intentions (Miller et al., 2012). For example, fear is known to elicit avoidance intentions. When
White participants were made to feel afraid, eliciting in them the motivation to protect the self
from danger, they displayed nonverbal avoidance tendencies (indicated by pushing away a lever)
upon seeing faces of Black men but not White or Asian men, which was predicted given negative
racial stereotypes associating Black men with danger. This differential speed of pushing
responses among fearful participants in response to Black male faces as compared to White or
Asian faces is suggestive of avoidance intentions. When White participants were made to feel
disgusted, eliciting the motivation to protect the self from contamination, there was no race bias
in avoidance tendency (Miller et al., 2012).
A large body of research has shown that a reliable indirect measure of approach and avoidance
motivation involves using electroencephalography (EEG) to capture asymmetric activity in the
frontal cortex (Davidson, 1992; Harmon-Jones, 2003) such that relative left-sided asymmetry is
associated with approach motivation and right-sided asymmetry is associated with avoidance
motivation (for a review, see Coan & Allen, 2003). For example, using EEG to measure
preconscious motivations, Amodio et al. (2007) had White participants view a multiracial series
Page 16 of 28
of faces while their cortical EEG activity was recorded. Some participants received bogus
feedback suggesting that their responses to these faces were racially biased. Participants in this
condition reported elevated guilt, which is typically associated with the intention to halt a
transgression (avoidance motivation) and a subsequent intention to engage in reparation
(approach motivation). Results showed that elevated guilt was correlated with changes in frontal
cortical asymmetry consistent with a reduction in preconscious approach motivation. When the
same participants were presented with an opportunity to engage in prejudice-reducing behavior,
guilt was associated with another shift in frontal cortical asymmetry, this time consistent with
increased approach motivation. These results reveal the ways in which guilt elicited by an
external event is associated with adaptive changes in preconscious motivation and subsequent
behavior.
(p.136) Summary and Conclusions

We have seen that, in accordance with the theory of planned behavior, explicit beliefs, attitudes,
and intentions can account for considerable variance in behavior. However, we have also seen
that a variety of factors can influence the predictive validity of intentions, among them
forgetting, instability of intentions, and a discrepancy between hypothetical situations to which
intentions are expressed and the real contexts in which behavior occurs. The conscious
processes described in the TPB are likely to be invoked only for novel and/or important
decisions, whereas routine behaviors in everyday life tend to habituate and no longer require
conscious deliberation. Nevertheless, we have seen that implicit beliefs and attitudes can have
strong effects on important decisions.
Complementing the reasoned action approach, theory and research on implicit social cognition
start with the assumption that beliefs and attitudes are sometimes learned and expressed
without peoples awareness, and even when people are aware of their beliefs and attitudes, their
self-reports may not be entirely honest. When this is the case, implicit beliefs and attitudes, and
preconscious intentions and motivations, can add to the prediction of behavior. Furthermore,
just as people may be aware or unaware of the beliefs and attitudes that guide their behavior, so
too may they be aware or unaware of the factors that activate these beliefs and attitudes.
Because most everyday behavior is routine, and the factors guiding it are often outside
awareness, it is easy to construe it as largely automatic (Bargh & Chartrand, 1999).
Consistent with this line of reasoning, activation of constructs or goals below conscious
awareness has been found to influence behavior and goal striving. However, we have tried to
show that even when constructs and goals are activated outside awareness, their effects on
behavior are not completely automatic but are instead mediated by implicit beliefs and attitudes
that for most behaviors ultimately produce an explicit behavioral intention.
Note
References
Aarts, H., & Dijksterhuis, A. (2000). Habits as knowledge structures: automaticity in goal-
directed behavior. Journal of Personality and Social Psychology, 78(1), 5363.
Page 17 of 28
Aarts, H., & Dijksterhuis, A. (2003). The silence of the library: environment, situational norm,
and social behavior. Journal of Personality and Social Psychology, 84(1), 1828.
Abelson, R. P. (1981). Psychological status of the script concept. American Psychologist, 36(7),
715729.
Agerstrm, J., & Rooth, D.-O. (2011). The role of automatic obesity stereotypes in real hiring
discrimination. Journal of Applied Psychology, 96(4), 790805.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision
Processes, 50(2), 179211.
Ajzen, I. (2005). Attitudes, personality, and behavior (2nd ed.). Maidenhead, UK: Open
University Press.
Ajzen, I. (2012). The theory of planned behavior. In P. A. M. Lange, A. W. Kruglanski, & E. T.

Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 438459). London: Sage.
Ajzen, I., Brown, T. C., & Carvajal, F. (2004). Explaining the discrepancy between intentions and
actions: the case of hypothetical bias in contingent valuation. Personality and Social Psychology
Bulletin, 30(9), 11081121.
Ajzen, I., & Sexton, J. (1999). Depth of processing, belief congruence, and attitude-behavior
correspondence. In S. Chaiken & Y. Trope (Eds.), Dual-process theories in social psychology (pp.
117138). New York: Guilford.
Albarracn, D., Johnson, B. T., Fishbein, M., & Muellerleile, P. A. (2001). Theories of reasoned
action and planned behavior as models of condom use: a meta-analysis. Psychological Bulletin,
127(1), 142161.
Amodio, D. M., Devine, P. G., & Harmon-Jones, E. (2007). A dynamic model of guilt: implications
for motivation and self-regulation in the context of prejudice. Psychological Science, 18(6), 524
530.
Banaji, M. R. (2001). Implicit attitudes can be measured. In R. L. Roediger, I. N. Nairne, & A M.

Suprenant (Eds.), The nature of remembering: essays in honor of Robert G. Crowder (pp. 117
149). Washington, DC: American Psychological Association.
Bandura, A. (1986). Social foundations of thought and action: a social cognitive theory.
Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1997). Self-efficacy: the exercise of control. New York: Freeman.
Bandura, A., & Locke, E. A. (2003). Negative self-efficacy and goal effects revisited. Journal of
Applied Psychology, 88(1), 8799.
Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of being. American
Psychologist, 54(7), 462479.
Page 18 of 28
Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social behavior: direct effects of
trait construct and stereotype activation on action. Journal of Personality & Social Psychology,
71(2), 230244.
Bargh, J. A., Gollwitzer, P. M., Lee-Chai, A., Barndollar, K., & Troetschel, R. (2001). The
automated will: nonconscious activation and pursuit of behavioral goals. Journal of Personality
and Social Psychology, 81(6), 10141027.
Bazerman, M. H., & Banaji, M. R. (2004). The social psychology of ordinary ethical failures.
Social Justice Research, 17(2), 111115.
Blair, I. V. (2002). The malleability of automatic stereotypes and prejudice. Personality and
Social Psychology Review, 6(3), 242261. doi: 10.1207/s15327957pspr0603_8
Blair, I. V., Dasgupta, N., & Glaser, J. (2014). Implicit Attitudes. In M. Mikulincer and P. R.
Shaver (Eds.), APA Handbook of Personality and Social Psychology, Volume 1: Attitudes and
social cognition (pp. 665691). Washington DC: American Psychological Association.
Blakemore, S-J., & Frith, C. (2003). Self-awareness and action. Current Opinion in Neurobiology,
13, 219224.
Blakemore, S.-J., Wolpert, D. M., & Frith, C. (2002). Abnormalities in the awareness of action.
Trends in Cognitive Science, 6(1), 237242.
Blumenschein, K., Johannesson, M., Blomquist, G. C., Liljas, B., & OConnor, R. M. (1998).
Experimental results on expressed certainty and hypothetical bias in contingent valuation.
Southern Economic Journal, 65(1), 169177.
Brandimante, M., Einstein, G. O., & McDaniel, M. A. (Eds.). (1996). Prospective memory: theory
and applications. Mahway, NJ: Erlbaum Associates.
Calitria, R., Lowe, R., Eves, F. F., & Bennett, P. (2009). Associations between visual attention,
implicit and explicit attitude and behaviour for physical activity. Psychology & Health, 24(9),
11051123. doi: 10.1080/08870440802245306
Campbell, D. T. (1963). Social attitudes and other acquired behavioral dispositions. In S. Koch
(Ed.), Psychology: a study of a science (Vol. 6, pp. 94172). New York: McGraw-Hill.
Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge:

Cambridge University Press.
Cesario, J., Plaks, J. E., & Higgins, E. T. (2006). Automatic social behavior as motivated
preparation to interact. Journal of Personality and Social Psychology, 90(6), 893910.
Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-process theories in social psychology. New York:
Guilford Press.
Page 19 of 28
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation: immediate behavioral
predispositions to approach or avoid the stimulus. Personality and Social Psychology Bulletin,
25(2), 215224.
Coan, J. A., & Allen, J. J. B. (2003). The state and trait nature of frontal EEG asymmetry in
emotion. In K. Hugdahl and R.J. Davidson (Eds.), The asymmetrical brain (pp. 565615).
Cambridge, MA: MIT Press.
Conner, M., Sheeran, P., Norman, P., & Armitage, C. J. (2000). Temporal stability as a moderator
of relationships in the Theory of Planned Behaviour. British Journal of Social Psychology, 39(4),
469493.
Cooper, L. A., Roter, D. L., Carson, K. A., Beach, M. C., Sabin, J. A., Greenwald, A. G., & Inui, T.
S. (2012). The associations of clinicians implicit attitudes about race with medical visit
communication and patient ratings of interpersonal care. American Journal of Public Health,
102(5), 979987.
Crosby, E., Bromley, S., & Saxe, L. (1980). Recent unobtrusive studies of Black and White
discrimination and prejudice: a literature review. Psychological Bulletin, 87(3), 546563.
Cunningham, W. A., Preacher, K. J., & Banaji, M. R. (2001). Implicit attitude measures:
consistency, stability, and convergent validity. Psychological Science, 12(2), 163170.
Dasgupta, N. (2004). Implicit ingroup favoritism, outgroup favoritism, and their behavioral
manifestations. Social Justice Research, 17(2), 143169.
Davidson, R. J. (1992). Emotion and affective style: hemispheric substrates. Psychological

Science, 3(1), 3943.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of
information technology. MIS Quarterly, 13(3), 319340.
Devine, P. G. (1989). Stereotypes and prejudice: their automatic and controlled components.
Journal of Personality and Social Psychology, 56(1), 518.
Doll, J., & Ajzen, I. (1992). Accessibility and stability of predictors in the theory of planned
behavior. Journal of Personality and Social Psychology, 63(5), 754765.
Dovidio, J. F., Kawakami, K., & Gaertner, S. L. (2002). Implicit and explicit prejudice and
interracial interaction. Journal of Personality and Social Psychology, 82(1), 6268.
Eitam, B., & Higgins, E. T. (2010). Motivation in mental accessibility: relevance of a

Representation (ROAR) as a new framework. Social and Personality Psychology Compass, 4(10),
951967. doi: 10.1111/j.1751-9004.2010.00309.x
Epstein, S. (1991). Cognitive-experiential self-theory: an integrative theory of personality. In R.

Curtis (Ed.), The self with others: convergences in psychoanalytical, social, and pesonality
psychology (pp. 111137). New York: Guilford Press.
Page 20 of 28
Fadardi, J. S., & Cox, W. M. (2008). Alcohol-attentional bias and motivational structure as
independent predictors of social drinkers alcohol consumption. Drug and Alcohol Dependence,
97(3), 247256. doi: 10.1016/j.drugalcdep.2008.03.027
Fazio, R. H. (1990). Multiple processes by which attitudes guide behavior: the MODE model as
an integrative framework. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol.
23, pp. 75109). San Diego, CA: Academic Press.
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic
activation as an unobtrusive measure of racial attitudes: a bona fide pipeline? Journal of
Personality and Social Psychology, 69(6), 10131027.
Fazio, R. H., & Zanna, M. P. (1978). Attitudinal qualities relating to the strength of the attitude-
behavior relationship. Journal of Experimental Social Psychology, 14(4), 398408.
Fazio, R. H., & Zanna, M. P. (1981). Direct experience and attitude-behavior consistency. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 14, pp. 161202). New York:
Academic Press.
Fishbein, M., & Ajzen, I. (2010). Predicting and changing behavior: the reasoned action
approach. New York: Psychology Press.
Fisher, J. D., & Fisher, W. A. (1992). Changing AIDS-risk behavior. Psychological Bulletin,
111(3), 455474.
Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: Mcgraw-Hill.
Friese, M., Hofmann, W., & Wanke, M. (2008). When impulses take over: moderated predictive
validity of explicit and implicit attitude measures in predicting food choice and consumption
behaviour. British Journal of Social Psychology, 47(3), 397419.
Galdi, S., Arcuri, L., & Gawronski, B. (2008). Automatic mental associations predict future
choices of undecided decision-makers. Science, 321(5892), 11001102.
Galdi, S., Gawronski, B., Arcuri, L., & Friese, M. (2012). Selective exposure in decided and
undecided individuals: differential relations to automatic associations and conscious beliefs.
Personality and Social Psychology Bulletin, 38(5), 559569.
Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in

evaluation: an integrative review of implicit and explicit attitude change. Psychological Bulletin,
132(5), 692731.
Gawronski, B., & Payne, B. K. (2010). Handbook of implicit social cognition:Measurement,

theory, and applications. New York: Guilford Press.
Gold, R. S. (1993). On the need to mind the gap: on-line versus off-line cognitions underlying
sexual risk taking. In D. J. Terry, C. Gallois, & M. McCamish (Eds.), The theory of reasoned
action: its application to AIDS-preventive behavior (pp. 227252). Oxford: Pergamon Press.
Page 21 of 28
Gollwitzer, P. M. (1993). Goal achievement: the role of intentions. European Review of Social
Gollwitzer, P. M. (1999). Implementation intentions: strong effects of simple plans. American

Green, A. R., Carney, D. R., Pallin, D. J., Ngo, L. H., Raymond, K. L., Iezzoni, L. I., & Banaji, M.
R. (2007). Implicit bias among physicians and its prediction of thrombolysis decisions for black
and white patients. Journal of General Internal Medicine, 22(9), 12311238.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: attitudes, self-esteem, and
stereotypes. Psychological Review, 102(1), 427.
Greenwald, A., Banaji, M., Rudman, L., Farnham, S., Nosek, B., & Mellott, D. (2002). A unified
theory of implicit attitudes, stereotypes, self-esteem, and self-concept. Psychological Review,
109(1), 325.
Greenwald, A. G., Poehlman, T. A., Uhlmann, E., & Banaji, M. R. (2009). Measuring and using
the Implicit Association Test: III. Meta-analysis of predictive validity. Journal of Personality and
Haider, A. H., Sexton, J., Sriram, N., Cooper, L. A., Efron, D. T., Swoboda, S., Villegas C. V. Haut,
E. R., Bonds, M., Pronovost, P. J. Lipsett, P. A., Freischlag, J. A., Cornwell, E. E., III. (2011).
Association of unconscious race and social class bias with vignette-based clinical assessments by
medical students. JAMA: The Journal of the American Medical Association, 306, 942951
Harmon-Jones, E. (2003). Clarifying the emotive functions of asymmetrical frontal cortical

activity. Psychophysiology, 40(6), 838848.
Hassin, R. R., Aarts, H., Eitam, B., Custers, R., & Kleiman, T. (2009). Non-conscious goal pursuit
and the effortful control of behavior. In E. Morsella, P. M. Gollwitzer & J. A. Bargh (Eds.), Oxford
handbook of human action (pp. 549568). New York: Oxford University Press.
Hofmann, W., & Friese, M. (2008). Impulses got the better of me: alcohol moderates the
influence of implicit attitudes toward food cues on eating behavior. Journal of Abnormal
Psychology, 117(2), 420427.
Hofmann, W., Gawronski, B., Gschwendner, T., Le, H., & Schmitt, M. (2005). A meta-analysis on
the correlation between the Implicit Association Test and explicit self-report measures.
Personality and Social Psychology Bulletin, 31(10), 13691385.
Hofmann, W., Gschwendner, T., Nosek, B. A., & Schmitt, M. (2005). What moderates implicit-
explicit consistency? European Review of Social Psychology, 16(1), 335390.
Jacoby, L. L. (1991). A process dissociation framework: separating automatic from intentional

uses of memory. Journal of Memory and Language, 30(5), 513541.
Page 22 of 28
Jost, J. T., Rudman, L. A., Blair, I. V., Carney, D. R., Dasgupta, N., Glaser, J., Hardin, C. D.
(2009). The existence of implicit bias is beyond reasonable doubt: a refutation of ideological and
methodological objections and executive summary of ten studies that no manager should ignore.
Research in Organizational Behavior, 29, 3969.
Jussim, L. (2012). Social perceptions and social reality: why accuracy dominates bias and self-
fulfilling prophesy. New York: Oxford University Press.
Kahneman, D., & Frederick, S. (2005). A model of heuristic judgment. In K. J. Holyoak, R. G.

Morrison, K. J. Holyoak & R. G. Morrison (Eds.), The Cambridge handbook of thinking and
reasoning. (pp. 267293). New York: Cambridge University Press.
Keren, G., & Schul, Y. (2009). Two is not always better than one: a critical evaluation of two-
system theories. Perspectives on Psychological Science, 4(6), 533550. doi: 10.1111/j.
1745-6924.2009.01164.x
Klein, O., Doyen, S., Leys, C., de Saldanha da Gama, P. A. M., Miller, S., Questienne, L., &
Cleeremans, A. (2012). Low hopes, high expectations: expectancy effects and the replicability of
behavioral experiments. Perspectives on Psychological Science, 7(6), 572584.
Krosnick, J. A., Betz, A. L., Jussim, L. J., & Lynn, A. R. (1992). Subliminal conditioning of
attitudes. Personality and Social Psychology Bulletin, 18(2), 152162.
Krosnick, J. A., & Petty, R. E. (1995). Attitude strength: an overview. In R. E. Petty & J. A.
Krosnick (Eds.), Attitude strength: antecedents and consequences (pp. 124). Mahwah, NJ:
Erlbaum Associates.
Kruglanski, A. W., & Ajzen, I. (1983). Bias and error in human judgment. European Journal of
Social Psychology, 13(1), 144.
Kruglanski, A. W., Shah, J. Y., Fisbach, A., Friedman, R. S., Chun, W. Y., & Sleeth-Keppler, D.
(2002). A theory of goal systems. In M. P. Zanna (Ed.), Advances in experimental social
psychology (pp. 331378). San Diego, CA: Academic Press.
LaPiere, R. T. (1934). Attitudes vs. actions. Social Forces, 13, 230237.
in relation to onset of celebral activity (readiness potential): the unconscious initiation of a freely
voluntary act. Brain, 106, 623642.
Linn, L. S. (1965). Verbal attitudes and overt behavior: a study of racial discrimination. Social
Forces, 43, 353364.
Locke, E. A., & Latham, G. P. (1994). Goal setting theory. In H. F. J. ONeil & M. Drillings (Eds.),
Motivation: theory and research. (pp. 1329). Hillsdale, NJ: Erlbaum Associates.
Page 23 of 28
Loersch, C., & Payne, B. K. (2011). The situated inference model: an integrative account of the
effects of primes on perception, behavior, and motivation. Perspectives on Psychological
Science, 6(3), 234252. doi: 10.1177/1745691611406921
Mann, T., Tomiyama, A. J., Westling, E., Lew, A.-M., Samuels, B., & Chatman, J. (2007).
Medicares search for effective obesity treatments: diets are not the answer. American
Psychologist, 62(3), 220233. doi: 10.1037/0003-066x.62.3.220
McConnell, A. R., & Leibold, J. M. (2001). Relations among the Implicit Association Test,
discriminatory behavior, and explicit measures of racial attitudes. Journal of Experimental Social
Psychology, 37(5), 435442.
Meyer, D. E., & Schvaneveldt, R. W. (1971). Facilitation in recognizing pairs of words: evidence
of a dependence between retrieval operations. Journal of Experimental Psychology, 90(2), 227
234.
Miller, S. L. & Maner, J. K. (2011). Sick body, vigilant mind: the biological immune system
activates the behavioral immune system. Psychological Science, 22(12), 14671471.
Miller, S. L., Zielaskowski, K., Maner, J. K., & Plant, E. A. (2012). Self-protective motivation and
avoidance of heuristically threatening outgroups. Evolution and Human Behavior, 33(6), 726
735.
Neal, D. T., Wood, W., & Quinn, J. M. (2006). Habits: a repeat performance. Current Directions
in Psychological Science, 15(4), 198202.
Neely, L. H. (1977). Semantic priming and retrieval from lexical memory: roles of inhibitionless
spreading activation and limited-capacity attention. Journal of Experimental Psychology:
General, 106(3), 226254.
Nisbett, R., & Ross, L. (1980). Human inference: strategies and shortcomings of social judgment.
Englewood Cliffs, NJ: Prenctice-Hall.
Nosek, B. (2005). Moderators of the relationship between implicit and explicit evaluation.
Journal of Experimental Psychology: General, 134(4), 565584.
Nosek, B. A., Banaji, M. R., Greenwald, A. G. (2002). Harvesting implicit group attitudes and
beliefs from a demonstration web site. Group Dynamics: Theory, Research & Practice, 6(1), 101
115.
Nosek, B. A., & Smyth, F. L. (2007). A multitrait-multimethod validation of the Implicit

Association Test: implicit and explicit attitudes are related but distinct constructs. Experimental
Psychology, 54(1), 1429.
Olson, M. A., & Fazio, R. H. (2001). Implicit attitude formation through classical conditioning.
Psychological Science, 12(5), 413417.
Page 24 of 28
Ouellette, J. A., & Wood, W. (1998). Habit and intention in everyday life: the multiple processes
by which past behavior predicts future behavior. Psychological Bulletin, 124(1), 5474.
Ozono, H., Watabe, M., & Yoshikawa, S. (2012). Effects of facial expression and gaze direction
on approachavoidance behaviour. Cognition and Emotion, 26(5), 943949.
Papies, E. K., Stroebe, W., & Aarts, H. (2008). Healthy cognition: processes of self-regulatory
success in restrained eating. Personality and Social Psychology Bulletin, 34(9), 12901300. doi:
10.1177/0146167208320063
Papies, E. K., Stroebe, W., & Aarts, H. (2009). Understanding dieting: a social cognitive analysis
of hedonic processes in self-regulation. European Review of Social Psychology, 20, 339383. doi:
10.1080/10463280802563723
Penner, L. A., Dovidio, J. F., West, T. V., Gaertner, S. L., Albrecht, T. L., Dailey, R. K., &
Markova, T. (2010). Aversive racism and medical interactions with black patients: a field study.
Journal of Experimental Social Psychology, 46(2), 436440.
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123205). New York:
Academic Press.
Posner, M. I., & Snyder, C. R. R. (2004). Attention and cognitive control. In D. A. Balota, & E. J.
Marsh (Eds.), Cognitive psychology: Key readings (pp. 205223). New York, NY: Psychology
Press.
Powell, L. H., Calvin, J. E., III, & Calvin, J. E., Jr. (2007). Effective obesity treatments. American
Psychologist, 62(3), 234246. doi: 10.1037/0003-066x.62.3.234
Powell, M. C., & Fazio, R. H. (1984). Attitude accessibility as a function of repeated attitudinal
expression. Personality and Social Psychology Bulletin, 10(1), 139148.
Randall, D. M., & Wolff, J. A. (1994). The time interval in the intention-behaviour relationship:
meta-analysis. British Journal of Social Psychology, 33, 405418.
Roediger, H. L., & McDermott, K. B. (1993). Implicit memory in normal human subjects. In F.
Boller & J. Grafman (Eds.), Handbook of neuropsychology (pp. 63131). Amsterdam: Elsevier.
Rooth, D.-O. (2010). Automatic associations and discrimination in hiring: real world evidence.
Labour Economics, 17(3), 523534.
Rosenstock, I. M., Strecher, V. J., & Becker, M. H. (1994). The health belief model and HIV risk
behavior change. In R. J. DiClemente & J. L. Peterson (Eds.), Preventing AIDS: theories and
methods of behavioral interventions. AIDS prevention and mental health (pp. 524). New York:
Plenum Press.
Rudman, L. A., & Glick, P. (2001). Prescriptive gender stereotypes and backlash toward agentic
women. Journal of Social Issues, 57(4), 743762.
Page 25 of 28
Rydell, R. J., McConnell, A. R., Mackie, D. M., Strain, L. M. (2006). Of two minds: forming and
changing valence-inconsistent implicit and explicit attitudes. Psychological Science, 17(11), 954
958.
Sabin, J. A., & Greenwald, A. G. (2012). The influence of implicit bias on treatment
recommendations for 4 common pediatric conditions: pain, urinary tract infection, attention
deficit hyperactivity disorder, and asthma. American Journal Public Health, 102(5), 988995.
Sabin, J. A., Rivara, F. P., & Greenwald, A. G. (2008). Physician implicit attitudes and
stereotypes about race and quality of medical care. Medical Care, 46(7), 678685.
Schacter, D. L. (1987). Implicit memory: history and current status. Joumal of Experimental
Psychology: Learning, Memory, and Cognition, 13(3), 501518.
Schifter, D. E., & Ajzen, I. (1985). Intention, perceived control, and weight loss: an application of
the theory of planned behavior. Journal of Personality and Social Psychology, 49(3), 843851.
Schuman, H., & Presser, S. (1981). Questions and answers in attitude surveys: experiments on
question form, wording, and context. San Diego, CA: Academic Press.
Schwarz, N. (1999). Self-reports: how the questions shape the answers. American Psychologist,
54(2), 93105.
Seger, C. A. (1994). Implicit learning. Psychological Bulletin, 115(2), 163196. doi:

10.1037/0033-2909.115.2.163
Sheeran, P. (2002). Intention-behavior relations: a conceptual and empirical review. European

Review of Social Psychology, 12, 136.
Sheeran, P., Gollwitzer, P. M., & Bargh, J. A. (2013). Nonconscious processes and health. Health
Psychology, 32(5), 460473. doi: 10.1037/a0029203
Sheeran, P., & Orbell, S. (1998). Do intentions predict condom use? Meta-analysis and
examination of six moderator variables. British Journal of Social Psychology, 37, 231250.
Sheeran, P., & Orbell, S. (2000). Using implementation intentions to increase attendance for
cervical cancer screening. Health Psychology, 19, 283289.
Sheeran, P., Orbell, S., & Trafimow, D. (1999). Does the temporal stability of behavioral
intentions moderate intention-behavior and past behavior-future behavior relations? Personality
and Social Psychology Bulletin, 25(6), 721730.
Sigall, H., & Page, R. (1971). Current stereotypes: a little fading, a little faking. Journal of
Sigall, H., & Page, R. (1972). Reducing attenuation in the expression of interpersonal affect via
the bogus pipeline. Sociometry, 35(4), 629642.
Page 26 of 28
Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin,
119(1), 322.
Staats, A. W., & Staats, C. K. (1958). Attitudes established by classical conditioning. The Journal
of Abnormal and Social Psychology, 57(1), 3740.
Stock, A., & Stock, C. (2004). A short history of ideo-motor action. Psychological Research, 68(2
3), 176188. doi: 10.1007/s00426-003-0154-5
Strack, F., & Deutsch, R. (2004). Reflective and impulsive determinants of social behavior.
Personality and Social Psychology Review, 8(3), 220247.
Stroebe, W., van Koningsbruggen, G. M., Papies, E. K., & Aarts, H. (2013). Why most dieters fail
but some succeed: a goal conflict model of eating behavior. Psychological Review, 120(1), 110
138. doi: 10.1037/a0030849
Triandis, H. C. (1972). The analysis of subjective culture. New York: Wiley.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: heuristics and biases.
Science, 185(4157), 11241131.
van den Bergh, L., Denessen, E., Hornstra, L., Voeten, M., & Holland, R. W. (2010). The implicit
prejudiced attitudes of teachers: relations to teacher expectations and the ethnic achievement
gap. American Educational Research Journal, 47(2), 497527.
Verplanken, B., & Aarts, H. (1999). Habit, attitude, and planned behaviour: is habit an empty
construct or an interesting case of goal-directed automaticity? European Review of Social
Psychology, 10(1), 101134.
Webb, T. L., & Sheeran, P. (2006). Does changing behavioral intentions engender behavior
change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132(2), 249268.
will. American Psychologist, 54, 480492.
Yogeeswaran, K., & Dasgupta, N. (2010). Will the real American please stand up? The effect of
implicit stereotypes about nationality on discriminatory behavior. Personality and Social
Psychology Bulletin, 36(10), 13321345.
Zajonc, R. B. (1980). Feeling and thinking: preferences need no inferences. American

Ziegert, J. C., & Hanges, P. J. (2005). Employment discrimination: the role of implicit attitudes,
motivation, and a climate for racial bias. Journal of Applied Psychology. 90(3), 553562.
Notes:
(1) We are grateful to Wolfgang Stroebe for providing us with these correlation coefficients,
which were not reported in the published article.
Page 27 of 28
Page 28 of 28
The Neural Basis Underlying the Experience of Control in the Human Brain

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Lauren A. Leotti
Catherine Cho
Mauricio R. Delgado
DOI:10.1093/acprof:oso/9780190267278.003.0006

Converging evidence suggests the perception of controlor the set of beliefs in ones ability to
exert control over the environment and to produce desired resultsis integral for forming a
sense of agency, hence affecting an individuals general well-being. A large literature has
demonstrated that the presence or absence of a sense of agency can have a significant impact
on the regulation of emotion, behavior, and physiology. This chapter discusses the recent efforts
in neuroscience research on humans which investigates this important subject. Collectively, the
findings lend support to the theory that choice and personal control are inherently rewarding
and motivating, which is highly beneficial for survival. The chapter discusses the implications of
this research for understanding how the presence or absence of personal control influences
emotion regulation and contributes to maladaptive behaviour.
Keywords: perception of control, sense of agency, emotion, behavior, reward, motivation
Introduction
When people hear the word choice, they typically think of decisions that are mundane (what
are we going to eat for dinner tonight?) or exceptionally difficult (do I take a job at University A
or University B?). Yet most things we do involve making choices. When choices do not require
much effort (e.g., choosing where to allocate attention in the environment), we may not even
perceive them as personal decisions. If such choices were removed, however, we would be very
aware of their absence. Choice is important because it affords us the opportunity to be causal
agents, instrumental in achieving desired outcomes.
Page 1 of 31
Through choice, we develop a sense of agency, which refers to beliefs in our ability to exercise
control over the environment. Such beliefs in personal control are known to be highly adaptive,
for their presence or absence can have a profound impact on the regulation of behavior,
emotion, and physiology. While there is a rich history of research on beliefs in personal control
and choice in humans and animals (for reviews, see Bandura, 1997; Leotti, Iyengar, & Ochsner,
2010; Ryan & Deci, 2006; Shapiro, Schwartz, & Astin, 1996), the neural mechanisms
surrounding the experience of control are less well understood in humans. However, recent
human neuroscience research has begun to (p.146) contribute to this gap in our knowledge by
exploring the affective experience of control and its impact on affective and motivational
processes. Disruptions to control beliefs are at the core of many psychiatric disorders (Beck,
1976; Mansell, 2005; Ryan, Deci, & Grolnick, 1995; Shapiro et al., 1996; Strupp, 1970; Taylor &
Brown, 1988, 1994), which may be associated with abnormal processing of affective and
motivational stimuli in the brain related to failures in self-regulation (Heatherton & Wagner,
2011; Johnstone, van Reekum, Urry, Kalin, & Davidson, 2007). As a result, it is critical to
understand how control is experienced in the human brain, because such research has
significant implications for understanding the psychological and neural mechanisms related to
the origin, maintenance, and potential treatment of many psychiatric disorders.
In this chapter we argue that the capacity to choose between alternatives and to decide which of
those options will occur increases feelings of control. In other words, individuals acquire a sense
of agency through accumulated experiences of choosing and being the agent of determining the
flow of events. Therefore, in this chapter we argue that a sense of control, as it is exercised
through removal choice, is inherently valuable and serves adaptive functions. We review the
literature examining the value of exercising control, presenting evidence from behavioral
psychology as well as from recent neuroimaging studies that highlight the role of affective and
motivational brain circuitry in the appraisal of choice opportunity. Additionally, we discuss
potential influences to the value of choice, and address the implications for this line of research
for future exploration.
Perception of Control Theories

There are several theories that attempt to explain the processes underlying humans perception
of controllability, including locus of control (Rotter, 1966), illusion of control (Langer, 1975),
self-efficacy (Bandura, 1997), and self-determination theory (Ryan & Deci, 2000). Drawing
from several theories in psychology, White (1959) proposed competence as a mechanism to
explain human and animal motivation. He defined competence/effectance as the ability to
effectively interact with the environment, which can be considered as one of the sources of
human motivation, drawing upon an important link between effectance and motivation. In line
with Whites argument, Higgins (2012) also refers to effectance as one of the sources of the
motivational aspect of competence. Instead of viewing human behavior as goal directed,
individuals seek to be aroused and to maintain activityhence, acting and thus engaging in a
trend of behaviors can have meaning to an individual in and of themselves. Additionally, a recent
study by Eitam, Kennedy, & Higgins (p.147) (2013) suggests that ones motivation toward
performance is enhanced when there is a direct contingency between ones own action and the
outcome. Thus, motivation levels are heightened when individuals feel in more control over the
outcome.
Page 2 of 31
Despite differences across such theories, the underlying themes uniformly tout the adaptiveness
of perceiving control and agency. Most notably, the term self-efficacy, defined by Albert
Bandura, refers to the set of beliefs in ones capability to perform courses of action to
accomplish ones goals. According to Bandura, these self-efficacy beliefs influence individuals
behavior by interacting with cognitive and affective processes, promoting persistence with
difficult tasks in individuals with high self-efficacy, and increasing susceptibility to stress in
individuals with low-self efficacy (Bandura & Wood, 1989). Extensive research supports the
adaptiveness of perceived self-efficacy in different spheres of psychosocial functioning, including
work-related performance (Stajkovic & Luthans, 1998), child development (Bandura, Caprar,
Barbaranelli, Gerbino, & Pastorelli, 2003), academic achievement and persistence (Multon,
Brown, & Lent, 1991), and health functioning (Holden, 1992).
Beliefs in self-efficacy depend on our ability to actually exert control over our environment.
Averill (1973) posited that there are three main ways that one can exert control: cognitive,
behavioral, and decisional control. Cognitive control refers to regulating the way a potentially
threatening stimulus is interpreted, such as altering the meaning or significance of the event or
stimulus. The capacity for cognitive control is important for self-regulation (Baumeister,
Heatherton, & Tice, 1994; Ochsner & Gross, 2005; Vohs & Baumeister, 2011), and may be
critical for fostering self-efficacy beliefs. The other two types of control involve overt behavior,
and are most relevant to the study of agency. As defined by Averill (1973), behavioral control is
the ability to prevent or modify certain aspects of an event through implementing direct action,
and decisional control is the selection of a single course of action from possible alternatives.
Although Averill sees the source under the influence of control as the event when referring to
behavioral control, it is also worthwhile to note that this definition is arbitrary, depending on
whether the object under the influence of control is the event itself or the behavior according to
ones goal. Both of these types of control involve an individual acting as a causal agent to
achieve a desired goal. The theory of reinforcement learning (Skinner, 1953; Thorndike, 1933)
states that when a specific behavior results in a desired outcome, that behavior is reinforced
(i.e., it is more likely to be repeated in the future). Importantly, while the action is successful at
producing desired results, the agent himself is also successful at choosing the appropriate
action. As a consequence, the opportunity to choose may be reinforced as well, and choice
opportunity, then, becomes desirable in and of itself.
(p.148) The opportunity to choose provides an individual with the opportunity to assert their
preferences, thus enhancing motivation and performance (Patall, 2013; Patall, Cooper, &
Robinson, 2008). Individuals feel more satisfied, competent, and engaged when they are able to
express a preference through choice (Cordova & Lepper, 1996; Grolnick & Ryan, 1987; Langer
& Rodin, 1976; Patall et al., 2008; Patall, Cooper, & Wynn, 2010; Ryan & Deci, 2000). Merely
having an opportunity to choose, even over something inconsequential, has been shown to have
a significant impact on quality and even duration of life (Langer & Rodin, 1976). Although the
current chapter focuses on the idea that the capacity to choose between alternatives is an
important aspect of feeling in control, individuals may also feel a sense of agency when only a
single alternative exists and they decide to pursue that option. For example, consumption of the
only available food in ones refrigerator can also instill a feeling of control in an individual. In
Page 3 of 31
such a situation, they choose to consume the food, rather than not, and may even increase the
value assigned to such food item as a result of having chosen it (albeit the lack of alternatives).
Human behavioral studies suggest that choice is not only highly motivating, but it is also
valuable. A seminal paper by Langer (1975) investigated the illusion of control, which refers to
the phenomenon in which an individual perceives control over an outcome when no true control
exists. The study revealed that choice has a substantial impact on control beliefs, and provided
some of the earliest direct evidence that choice is valuable. In Langers study, participants were
offered the opportunity to purchase a lottery ticket for $1, and were either allowed to freely
select their ticket (choice group) or were assigned a ticket (no choice group). On the day of the
lottery, participants were asked if they would be willing to sell their tickets. Those in the no-
choice group were willing to sell their tickets for an average of $1.96, but those in the choice
group priced their tickets at a whopping $8.67. She also found that people in the choice group
were less willing to switch to another lottery, even though it had better odds of winning. Other
studies have shown that people prefer options that lead to additional choice (Bown, Read, &
Summers, 2003; Leotti & Delgado, 2011; Suzuki, 1997, 2011), despite the fact that a secondary
choice requires greater effort without any additional reward. This suggests that choice itself
confers additional value, making it more desirable.
The simple act of choosing has been shown to elicit significant preference change. In the classic
free-choice paradigm (Brehm, 1956), items that are selected (e.g., blender vs. toaster), as
opposed to rejected, are rated as higher in value after selection, and those that are rejected are
rated as lower in value. This post-choice preference shift cannot be explained simply by
rationalization to minimize cognitive dissonance (i.e., if I chose the blender over the toaster, the
blender must be better than I initially thought). In fact, a study by Lieberman (p.149) and
colleagues (2001) revealed that post-choice preference shift could occur in amnesiacs, who
could not remember their explicit choice. Additionally, such choice-induced preference change
has been demonstrated in preschool-aged children and monkeys (Egan, Santos, & Bloom, 2007),
can occur when the choice is made blindly without reviewing alternatives (Sharot, Velasquez, &
Dolan, 2010), and can be long lasting, persisting years beyond the initial decision (Sharot,
Fleming, Yu, Koster, & Dolan, 2012). These studies suggest that the act of choosing, itself, is an
important modulator of affective valuation processes.
Neural Systems Underlying the Perception of Control

The plethora of data demonstrating the important role of perceiving control and the behavioral
evidence discussed in the preceding section suggest that choice is desirable and may be
inherently rewarding. A complement to the behavioral studies and theories is the introduction of
neural data to the question of perceiving control. Identification of the neural substrates of
control allows us to constrain our understanding of how perceiving and exercising control are
beneficial and adaptive. Specifically, one can hypothesize that if expectations of control, via
choice, are valuable and exert a rewarding feeling, then anticipation of choice opportunity
should recruit brain structures involved in reward-related processes. A highly interconnected
cortical-striatal network, modulated by dopaminergic neurons, has been implicated in
processing reward information and fostering goal-directed behavior (Berridge & Kringelbach,
2008; Delgado, 2007; Haber & Knutson, 2010; Levy & Dubois, 2006; ODoherty, 2004; Schultz et
al., 1997). Brain regions that support such reward processes include subcortical regions such as
Page 4 of 31
the striatum (which includes the caudate, putamen, and nucleus accumbens), and prefrontal
cortical (PFC) structures consisting of the orbitofrontal cortex (OFC) and medial prefrontal
cortex (MPFC), which have some structural overlap (Ongur & Price, 2000). The striatum is the
input unit of the larger basal ganglia complex and receives projections from several structures,
including cortical regions, putting it in a prime position to process cognitive, motor, and
motivational information and to influence behavior (Balleine, Delgado, & Hikosaka, 2007; Haber
& Knutson, 2010; Middleton & Strick, 2000). While both subcortical and cortical regions have
been linked to reward processing, the current discussion will focus largely on the striatum, due
to the preponderance of evidence demonstrating its role in instrumental learning, when rewards
are contingent upon behavior (McClure, Berns, & Montague, 2003; ODoherty, Dayan, Friston,
Critchley, & Dolan, 2003; ODoherty et al., 2004; Yacubian (p.150) et al., 2006), which is most
relevant for the discussion of the neural substrates of exercising behavioral control.
Functional magnetic resonance imaging (fMRI) studies in humans have supported a rich animal
literature (e.g., Robbins & Everitt, 1996) and have implicated the striatum in response to the
receipt and anticipation of rewards. Such studies have found increased striatal activity in
response to the receipt of primary reinforcers such as food and drinks (ODoherty, Rolls,
Francis, Bowtell, & McGlone, 2001) and secondary reinforcers such as monetary rewards
(Delgado, Nystrom, Fissell, Noll, & Fiez, 2000; Knutson & Cooper, 2005), as well as to the mere
anticipation of rewards (Kirsch et al., 2003; Knutson, Adams, Fong, & Hommer, 2001; Knutson,
Taylor, Kaufman, Peterson, & Glover, 2005; ODoherty, Deichmann, Critchley, & Dolan, 2002).
Additionally, the striatum differentiates between rewards and punishments and is sensitive to
the magnitude and probability of rewards (Delgado, Locke, Stenger, & Fiez, 2003; Delgado et
al., 2000; Delgado, Stenger, & Fiez, 2004; Kirsch et al., 2003; Knutson et al., 2005; Tobler,
ODoherty, Dolan, & Schultz, 2007; Yacubian et al., 2006).
Although various divisions of the striatum have been proposed based on anatomy and function in
rodents, such as ventromedial to dorsolateral representing initial learning to habit formation
(Balleine & ODoherty, 2009; Voorn, Vanderschuren, Groenewegen, Robbins, & Pennartz, 2004),
the most basic division in humans involves dorsal and ventral portions. The dorsal striatum
includes the caudate and putamen and is thought to be preferentially activated for tasks or
stimuli with increased motivational incentives (Delgado et al., 2004; Zink, Pagnoni, Martin-
Skurski, Chappelow, & Berns, 2004). Furthermore, whereas the ventral striatum, including
nucleus accumbens and ventral portions of caudate and putamen, responds to reward value
irrespective of the actions leading to rewards (i.e. stimulus-outcome relationships), the dorsal
striatum responds more for rewards that are contingent upon behavior (ODoherty et al., 2004).
Several studies have demonstrated that individuals recruit greater activity in the striatum when
rewards are contingent upon their responses, than when they are just passively delivered,
illustrating the important role of this region in processing contingency (Bjork & Hommer, 2007;
Elliott, Newman, Longe, & William Deakin, 2004; ODoherty et al., 2004; Tricomi, Delgado, &
Fiez, 2004).
One such example is a study by Tricomi and colleagues (2004), which was among the first
human neuroimaging studies to illustrate the role of the dorsal striatum in processing
contingency between choice opportunity and outcomes. In this study, participants were led to
Page 5 of 31
believe that on some trials, outcomes were dependent on their actions (choice condition), and on
other trials, their actions had no effect on potential outcomes (no-choice condition). (p.151)
Participants reported perceiving greater control over outcomes in the choice condition (relative
to no-choice), as well as greater motivation to win money. Further, the dorsal striatum was
recruited only for the choice condition, when participants believed that the outcomes were
dependent on their actions.
Another example is a study by Tanaka and colleagues (Tanaka, Balleine, & ODoherty, 2008),
which found that in addition to the dorsal striatum, the medial prefrontal cortex (MPFC) and
medial OFC are also involved in computing contingency. Specifically, greater reports of
subjective causality were associated with greater activity in MPFC, suggesting that this region is
important for processing agency, consistent with research supporting the role of this region in
processing self-relevance (Heatherton et al., 2006; Johnson et al., 2002; Kelley et al., 2002;
Platek, Keenan, Gallup Jr., & Mohamed, 2004) and preference-based decision-making (Johnson
et al., 2005; Paulus & Frank, 2003).
In sum, initial neuroimaging studies have supported an animal literature highlighting the role of
the striatum and cortical regions in reward-related processing and motivated behavior. Further,
such studies have delineated the involvement of the dorsal striatum and MPFC in action-
contingency and perceiving control. The ventral striatum, a critical structure for reward
processing, has been associated with computing the value of potential rewards and making
predictions to aid goal-directed behavior, expressed through consummatory and anticipatory
signals.
The Value of Choice: Neural and Behavioral Correlates

Studies demonstrating the sensitivity of the striatum to agency or behavioral contingency during
reward processing (e.g., Bjork & Hommer, 2007; ODoherty, Critchley, Deichmann, & Dolan,
2003; Tricomi et al., 2004) provided critical preliminary support to the hypothesis that
exercising control, via choice, modulates activity in reward circuitry. However, these studies
focused on the decision period and the outcomes following choice, where a true difference
existed between choice options, so that having an opportunity to choose clearly afforded the
subject the advantage of selecting the best outcome. From behavioral studies, we know,
however, that choice seems to be desirable even if there is no true difference in outcomes (Bown
et al., 2003; Leotti & Delgado, 2011; Suzuki, 1997, 2011). Therefore, it is critical to dissociate
the positive effects of exercising control by having the opportunity to choose between options
from merely selecting the option that leads to higher reward. Thus, to better characterize the
affective experience of control and choice, it is necessary to directly examine whether
expectancies of control opportunities are rewarding, while controlling for the various affective
and cognitive (p.152) processes involved in decision-making tasks used in the previous studies,
such as computation of expected value of potential outcomes at the time of choice and emotional
responses related to perceived success following outcomes.
Leotti and Delgado (2011a) tested this hypothesis with a simple choice paradigm aimed at
examining the affective experience when anticipating choice opportunity. In this study,
participants viewed cues that predicted free choice or forced-choice, where a response would
lead to a potential monetary reward. Unbeknownst to the participants, both available options
Page 6 of 31
(whether freely chosen or selected by the computer) would lead to the same average reward. We
considered this manipulation of fundamental importance, since our primary hypothesis was
concerned with the idea of the opportunity for choice itself having a rewarding feature, which is
not necessarily dependent upon the outcomes associated with ones choice (see Leotti &
Delgado, 2014). Thus, any differences in reported liking of the different cue types (free vs.
forced choice) and in associated blood-oxygen-level-dependent (BOLD) responses in reward
circuitry could be attributed to differences in the appraisal of the cues themselves, and not to
true differences in outcomes, or differences in perceived success. We found that participants
liked cues predicting free choice better than those predicting forced choice (Figure 6.1a).
Furthermore, choice cues recruited greater activity in reward-related regions previously
implicated in the anticipation of reward, including the ventral striatum, midbrain, and dorsal
anterior cingulate cortex (ACC). These regions have been linked to reward processing more
generally (Delgado, 2007; Knutson et al., 2001; Knutson et al., 2005; ODoherty, 2004), and have
been linked to the voluntary engagement in risky decision-making (Rao, Korczykowski, Pluta,
Hoang, & Detre, 2008).
Choice may be desirable, and may recruit

reward-related circuitry, because it is
inherently valuable, or alternatively,
because it increases the predictability of
outcomes or decreases feelings of risk, both
which have been theorized to contribute to
feelings of control (Thompson, 1999). Figure 6.1 Behavioral and fMRI correlates
of choice value (adapted from Leotti &
Additional experimental control conditions
Delgado, 2011). (a) Participants reported
included in our study (Leotti & Delgado, greater subjective liking for cues predicting
2011) allowed us to tease apart the Choice (C) than those predicting No-choice
contributions of these cognitive and (NC); (b) Main effects of cue type revealed
significant effects in striatum bilaterally,
affective components involved in decision- dorsal ACC, and right amygdala; (c) in the
making. In our Non-informative (NI) right striatum, we observed significantly
condition, a cue was followed by choice 50% greater fMRI BOLD activity for Choice
of the time, thus manipulating the condition relative to No-choice and to the
Non-informative (NI) and Predictable (P)
uncertainty of choice opportunity. In a control conditions. Error bars represent
predictable (P) no-choice condition, the cue standard errors of the mean. The asterisk
indicated that the participant would have no marks a significant difference (p < .05).
choice, but it also revealed which colored
key would be selected by the computer,
controlling for potential differences in
predictability. Figure 6.1b displays the main effects of cue type in a whole-brain analysis,
illustrating effects in the striatum, dorsal ACC, and amygdala. Figure 6.1c reveals that
recruitment of the striatum was related to the probability of having a choice (Choice > Non-
informative > No choice), but was not modulated by the predictability of the outcomes (i.e., no
difference (p.153) (p.154) for the Predictable cue, which indicated what the selected option
would be, in comparison to all other conditions). Hence, striatal activity depended upon the
opportunity for choice, rather than a preference for a particular option that is being reflected in
the predictable cue. In addition to the ventral striatum, the dorsal striatum also was
Page 7 of 31
preferentially activated for choice cues, consistent with previous studies of contingency effects
and increased motivational salience, discussed earlier. The recruitment of both ventral and
dorsal striatum suggests that the anticipation of choice opportunity involves processes related to
action-outcome as well as stimulus-outcome contingencies.
Besides the striatum, our analyses revealed that patterns of BOLD activity differed for our
experimental and control conditions in distinct regions of the brain. Heightened activity in the
amygdala was observed for the Non-informative cue, potentially reflecting risk or uncertainty
(Hsu, Bhatt, Adolphs, Tranel, & Camerer, 2005). Additionally, greater activity in the dorsal ACC
was observed for cues predicting choice and possible choice (Non-informative) relative to the
cues predicting no choice, which may reflect the motivational salience of choice opportunity
when anticipating effortful decision-making (Rushworth, Walton, Kennerley, & Bannerman,
2004). The dissociation in the recruitment of these regions across conditions suggests that we
can interpret the recruitment of reward regions, such as the striatum, to reflect reward
processing, rather than other cognitive processes related to choice and control.
Although there was no true difference between options in our tasks, participants reported a
preference for one option over the other. However, anticipation of the computers selection of
the preferred option did not selectively recruit reward circuitry. Nonetheless, it is possible that
individuals value choice opportunity because they believe, correctly or incorrectly, that choice
will provide them access to the best option available. Although the mathematical expected value
(average rewards) of the two available options are equal in our task, participants mis-
estimations of expected value, due to trial-by-trial fluctuations in rewards, may contribute to
perceived differences in the value of the colored keys. As a result, subjects may believe that
choice is more valuable because it allows for the selection of the key that has the highest
expected value at any given time.
The findings from Leotti and Delgado (2011a) suggest that reward-related brain circuitry is
recruited when anticipating choice opportunity, potentially providing support for the hypothesis
that choice is inherently valuable. Another way to quantify the value of choice opportunity is to
determine how much participants were willing to pay for additional choice. An interesting study
(Fujiwara et al., 2013) asked participants if they would rather receive a specific amount of
money or an opportunity to choose from a set number (p.155) of objects. They found that when
the amount of money offered was held constant, participants preferred to have more choices.
Moreover, the value, or willingness to pay, increased with the number of choices available, as
did associated activity in the ventral striatum. As we mentioned earlier, increasing the number
of choice options should increase the value of choice, because it increases the likelihood of
obtaining the best option. However, individuals in this study seemed to value choice above and
beyond what would be expected due to mathematical increases in expected value, implying that
choice has value in and of itself. One interesting line of research to extend these ideas is the
implementation of computational models to better understand how the choice for preference
may develop. For instance, reinforcement learning models have been extensively used in fMRI
research (for a review, see ODoherty, 2007; Frank & Fossella, 2011; Delgado & Dickerson,
2012; Doll et al., 2012), highlighting the involvement of regions such as the striatum in
processing a prediction error signal. Such computational models may prove to be quite valuable
in explaining the mechanisms through which one comes to prefer choice and to value it. For
Page 8 of 31
instance, a recent study posits that the preference for choice develops due to positive prediction
errors experienced when one makes a successful choice (Cockburn et al., 2014).
The Value of Choice and Negative Context

Collectively, human neuroimaging studies of experiencing control over positive outcomes (e.g.,
contingency, free choice) suggest that brain regions involved in reward-related processes are
important in the affective appraisal of choice opportunity. In the context of positive outcomes,
choice may be desirable because it affords opportunities to optimize rewards. Often, however,
people are faced with the challenge of making decisions to avoid potentially negative outcomes.
In these cases, having an opportunity to exercise agency, through choice, may serve an
important role for reducing the stress associated with uncertainty and threat. In the next
section, we discuss the value of choice and control in the context of potentially negative
outcomes.
When individuals are faced with potentially negative outcomes, they may experience stress and
engage in avoidance behavior. However, perceived control has been shown to buffer the
negative emotional response to aversive events. For example, belief in ones ability to exercise
control over a stressful situation has significant impact on autonomic arousal, release of stress
hormones, and functioning of the immune system (Abelson, Khan, Liberzon, Erickson, & Young,
2008; Bandura, Taylor, Williams, Mefford, & Barchas, 1985; Kamen-Siegel, Rodin, Seligman, &
Dwyer, 1991; Maier, Laudenslager, & Ryan, 1985). Behavioral control has been shown to
mitigate arousal during (p.156) anticipation of aversive noise (Glass, Singer, & Friedman,
1969) or photographs (Geer & Maisel, 1972), and to increase tolerance to electric shock (Staub,
Tursky, & Schwartz, 1971), and pain (Kanfer & Seider, 1973). Perceived control over painful
stimuli also reduces subjective reports of pain and anxiety (Salomons, Johnstone, Backonja, &
Davidson, 2004; Salomons, Johnstone, Backonja, Shackman, & Davidson, 2007; Salomons et al.,
2010; Wiech et al., 2006). Importantly, the mere belief in control opportunities is sufficient to
elicit benefits, even if control is never exercised (Corah & Boffa, 1970; Glass, Reim, & Singer,
1971; Gunnar-vonGnechten, 1978).
Having the opportunity to exercise control, through choice, may reduce negative affect induced
by the threat of an aversive outcome. Thus, we would expect that expectancies of choice
opportunity would lead to better coping, or self-regulation, when outcomes are uncertain and
potentially aversive and, as a result, should recruit brain networks involved in cognitive control
and successful emotion regulation. Regulation of negative affect involves the recruitment of
cortical regions within the lateral and medial PFC that exert a modulatory influence over
responses of regions involved in affective processing, such as the amygdala and insula, resulting
in attenuation of negative emotional expression (Ochsner & Gross, 2008). We might expect a
similar pattern of activity in the brain if the anticipation of control is comparable to other
antecedent-focused emotion-regulation strategies. At the same time, neuroimaging research has
demonstrated the critical role of the striatum as a mediator of the relationship between PFC
activity and successful regulation of negative affect (Hare, Tottenham, Davidson, Glover, &
Casey, 2005; Salomons et al., 2010; Wager, Davidson, Hughes, Lindquist, & Ochsner, 2008). As
a result, we might expect the striatum to also play a key role in the modulation of emotional
responses to choice opportunity when anticipating either potentially positive or negative
outcomes.
Page 9 of 31
Although there are only a handful of human neuroimaging studies investigating perceived
control to date, the preliminary evidence supports the hypotheses that expectations of control
influence brain activity involved in affective and motivational processes important for emotion
regulation. Several studies have demonstrated that the exercise of behavioral control over pain
reduces brain activity in pain-processing regions (Salomons et al., 2004; Wiech et al., 2006).
Moreover, recruitment of the PFC during the anticipation of control over pain may contribute to
these controllability effects, resulting in reduced pain processing and subjective reports of pain.
Controllability over pain and related reductions in anxiety also have been associated with
increased activity in the ventral striatum (Salomons et al., 2010). A related study by Kerr,
McLaren, Mathy, and Nitschke (2012) demonstrated the important role of the ventromedial PFC
in the anticipation of control over an (p.157) aversive event. In that study, when snake phobic
participants had the opportunity to terminate a threatening video (as opposed to having no
control over viewing duration), they recruited greater activity in the ventromedial PFC, a region
that had previously been identified in rodents as critical for supporting controllability effects on
stress regulation (Maier, Amat, Baratta, Paul, & Watkins, 2006). These results relate well with
ideas associated with sensorimotor attenuation (Voss, Ingram, Haggard, & Wolpert, 2006),
which highlight reductions in neural responses in sensory areas following voluntary actions.
Interestingly, the mere expectation of a potential action has a significant effect on
somatosensory perception, even in the absence of the execution of behavioral command (Voss et
al., 2008). Taken together, findings in controllability over pain and sensorimotor attenuation
support the idea that similar neural mechanisms may be involved in the expectancy of control
over behavior and stressors (i.e., pain).
The studies described above represent opportunities to exercise behavioral control, via escape,
such that an action can lead to the avoidance of a negative outcome. These studies highlight the
role of regions of the PFC in the anticipation and exercise of control. On the other hand, when
people anticipate decisional control (when a choice must be made between options), the ventral
striatum seems to play an important role (Leotti & Delgado, 2014), consistent with previous
findings when anticipating decisional control leading to potential gains (Leotti & Delgado,
2011). However, the recruitment of the ventral striatum seems to be dependent on the context
in which losses are incurred. Specifically, when losses were incurred in the context of potential
simultaneous gains, there was tremendous inter-subject variability in the reported liking of
choice, as well as ventral striatum recruitment during anticipation of choice. In fact, half of all
participants in the study reported that they preferred to have no choice if choice could lead to
potential losses, and in these participants, we observed greater ventral striatum activity when
anticipating the no-choice condition. However, in a separate experiment, when we presented
losses in the absence of gains, participants more consistently reported that they preferred
choice, and showed greater choice-related BOLD activity in the ventral striatum. While the
impact of context effects (e.g., framing effects or endowment effects) on loss aversion and
decision-making are well-known (De Martino, Kumaran, Holt, & Dolan, 2009; De Martino,
Kumaran, Seymour, & Dolan, 2006; Kahneman & Tversky, 2000; Tversky & Kahneman, 1981),
the current results extend these findings to the prospect of exercising control. These findings
suggest that the value of choice may depend on various situational contexts, as well as
individual differences, which require further exploration in future human neuroscience research.
Page 10 of 31
(p.158) In summary, the MPFC and

striatum play an important role in the
affective experience of control and choice,
when anticipating both rewarding and
aversive outcomes. The putative role of
these regions with respect to choice is
summarized in Figure 6.2. Recruitment of
the MPFC may be important for increasing
motivational salience in the context of Figure 6.2 Overview of neural circuitry
positive outcomes, and may be responsible involved in experience of choice (adapted
for reducing threat in the context of from Leotti, Iyengar, & Ochsner, 2010). The
negative outcomes. While choice opportunity to exercise control over positive
outcomes, via choice, recruits brain regions
opportunity seems to be rewarding in the commonly involved in reward processing,
context of positive outcomes, personal and such as the striatum, which comprises the
situational factors may influence choice caudate (CAU), putamen (PUT), and nucleus
accumbens (nAcc). Other reward regions in
value, and associated modulation of the
the medial prefrontal cortex (MPFC) may
striatum, when anticipating potentially also be preferentially activated to reflect
negative outcomes. increased reward value under increased self-
relevance and general processing of
Influences to the Value of Choice and contingency. In the context of potential
threats, the opportunity to exercise control
Control may recruit regions of the brain that are
In addition to the context effects described critical for adaptive emotion regulation, such
above, there are many potential influences as the lateral PFC, MPFC, and striatum, and
to the value of choice. A recent study by consequently result in decreases in brain
regions involved in negative emotional
Wenke, Fleming, and Haggard (2010) responses, such as the amygdala (AMY).
observed that subliminally priming one When faced with negative outcomes, the
particular response over an alternative value of choice may be sensitive to individual
differences and situational context.
enhanced the sense of control. Intuitively,
the results may (p.159) seem to imply that
subjects felt heightened levels of agency
when in fact they were least in control of the outcome, or that the sense of control is heightened
when there is less interference in the selection process. One interesting idea is that subliminal
priming of a specific key response may cause greater fluency in behavior; that is, a subliminally
primed stimulus that is compatible with a motor response promotes easier and faster action
selection, which thereby increases feelings of control (Chambon et al., 2013). People believe
they are responsible for either the improvement (in the case of positive outcomes) or worsening
(for negative outcomes) of their well-being (Beattie, Baron, Hershey, & Spranca, 1994; Botti,
McGill, & Iyengar, 2003; Burger, 1989). As a result, choice seems to magnify the anticipated
outcome. Because choice opens up the possibility of making a bad decision, and being
responsible for a negative outcome, people may find choice undesirable in the context of
potentially negative outcomes. This discrepancy in the preference for choice is consistent with
the self-serving bias, such that individuals take credit for successes but tend to attribute
failures to external sources beyond ones control (Brewin & Shapiro, 1984; D. T. Miller & Ross,
1975; Rotter, 1966). Interestingly, part of the dorsal striatum, the caudate nucleus, has been
Page 11 of 31
associated with the self-serving bias attribution (Blackwood, Bentall, Simmons, Murray, &
Howard, 2003).
There is also a tendency for people to prefer to accept the status quo, as opposed to overtly
acting in a way that could lead to an error and potential regret (Baron & Ritov, 1994; Feldman,
Miyamoto, & Loftus, 1999; Samuelson & Zeckhauer, 1988; Tsiros & Mittal, 2000). This is
consistent with research demonstrating that people experience greater elation following actions
that lead to positive events and greater regret following actions that lead to negative outcomes
(Kahneman & Tversky, 1982; Landman, 1987). Regions involved in anticipation of control, such
as the anterior insula and MPFC, have also been linked to the anticipation of regret (Coricelli et
al., 2005; Fleming, Thomas, & Dolan, 2010; Nicolle, Fleming, Bach, Driver, & Dolan, 2011). It is
unclear, however, whether this ambivalence toward, or even dislike of, control applies only to
circumstances of decisional control, or whether it also extends to opportunities to exercise
behavioral control.
If we argue that choice itself is valuable, then we might assume that the more choice we have,
the happier we will be. While this may be true to a degree (Fujiwara et al., 2013), an excessive
amount of choice is burdensome. In fact, a surfeit of choice proves to be highly demotivating, a
phenomenon that has been referred to as the tyranny of choice (Schwartz, 2000). Though
people may initially find a larger choice assortment to be more attractive, they tend to be less
satisfied with their choice or to defer choice altogether (Iyengar & Lepper, 2000). Too much
choice may be cognitively burdensome (Dhar, 1997; Iyengar & Lepper, 2000; Shiv & Fedorikhin,
1999) and depleting of self-control (p.160) resources (Vohs et al., 2008), particularly when it is
difficult to differentiate between equally attractive options (Fasolo, Hertwig, Huber, & Ludwig,
2009; Sela, Berger, & Liu, 2009), which may engender feelings of dissatisfaction in choice and
loss of confidence (Iyengar, Wells, & Schwartz, 2006) or buyers remorse and regret (Inman &
Zeelenberg, 2002; Sagi & Friedland, 2007; Tsiros & Mittal, 2000). While this research suggests
that excessive choice may be undesirable, at the same time, other research has demonstrated
that people will work very hard to leave options open to them, for fear of losing additional choice
(Shin & Ariely, 2004). It is this conflict between desire and disdain for choice that is
appropriately coined the paradox of choice (Schwartz, 2009).
Nonetheless, there is a difference between having choices and making choices. Research
suggests that when it comes to difficult decisions (e.g., medical decisions), whereas many people
do not like making the final decision, the large majority of individuals prefer to have a choice, at
the very least (Levinson, Kao, Kuby, & Thisted, 2005; Ogden, Daniells, & Barnett, 2008).
However, choosing may be more stressful when there is insufficient information to make an
informed selection (Paterson & Neufeld, 1995). At these times, it may be preferable to defer
choice to someone with greater expertise, such as a doctor (Levinson et al., 2005). Individuals
may still experience a sense of control, however, if they believe someone is acting in their best
interest (i.e., control by proxy).
Other social factors may also contribute to the value of choice. An interesting study by Ybarra
and colleagues (2012) suggests that security associated with positive social relationships may
decrease the value of choice. In one experiment, participants were instructed to write about
either a supportive or unsupportive social relationship, and then were asked to engage in a
Page 12 of 31
decision-making task in which they could purchase a cell phone. Participants decided whether
they wanted to be assigned a cell phone (default mode option), or they could pay a small
incremental fee to have additional options made available to them. Participants primed to think
about a supportive relationship were less willing to pay for additional choice. Furthermore, a
second experiment revealed that feelings of calmness and security associated with the social
relationship were significant mediators of the observed decreased choice preference (Ybarra et
al., 2012). One explanation for this finding is that feelings of calmness and security may result in
a reduced desire to explore through choice.
Finally, there are individual differences that may contribute to the affective experience of
choice. Differences in decision-making strategies may influence the value of choice and control.
Individuals who are maximizers (i.e., those who look to find the best option available (Iyengar
et al., 2006; Schwartz et al., 2002), for example, tend to experience greater dissatisfaction in
choice (p.161) and greater regret than do people who are satisfiers (i.e., those who choose
options that are good enough). Culturally motivated beliefs may also influence the value of
choice, such that individuals with an independent focus (commonly associated with Western
cultures) may prefer to make their own choices, whereas individuals with an interdependent
focus (commonly associated with Eastern cultures) may prefer choices to be made by benevolent
others (Iyengar & Lepper, 1999). More generally, individual differences in optimism, self-
esteem, and mood have been linked to the tendency to demonstrate an illusion of control (Alloy
& Abramson, 1982; Fontaine, Manstead, & Wagner, 1993; Taylor & Brown, 1988).
With respect to the brain, there are a few studies that highlight the existence of individual
differences in the experience of control. For example, as we previously mentioned, individuals
who reported liking choice better than no choice demonstrated greater recruitment of the
striatum (Leotti & Delgado, 2014). Relatedly, individual differences in preference changes
following choice (e.g., liking a vacation spot better after you chose it) correlate with extent of
recruitment of the ventral striatum (Sharot, De Martino, & Dolan, 2009; Sharot, Shiner, &
Dolan, 2010). Furthermore, pre-existing beliefs about ones ability to exercise control over life
events have been linked to the recruitment of PFC regions during controllable pain (Salomons et
al., 2007; Wiech et al., 2006). An individuals tendency to avoid cognitive demand in decision-
making is associated with greater activity in the lateral PFC (McGuire & Botvinick, 2010), and
may contribute to the value of choice when cognitive resources are taxed. Additional work in
this area is necessary to better understand how individual differences, as well as situational
factors, may contribute to behavioral and brain correlates of the affective experience of choice
and control.
The Significance of Understanding Control and Its Neural Basis

Further research investigating the affective experience of control and its neural correlates is
critical for understanding healthy functioning as well as maladaptive behavior. Struggles for
control are at the core of many psychiatric disorders (Beck, 1976; Mansell, 2005; Ryan, Deci, &
Grolnick, 1995; Shapiro et al., 1996; Taylor & Brown, 1988, 1994), including anxiety disorders
(Abramson, Garber, & Seligman, 1980; Bandura, 1988; Reuven-Magril, Dar, & Liberman, 2008),
depression (Deutsch, 1978; Mathews, 1977; Schwartz et al., 2002; Seligman, 1975), drug and
alcohol addictions (Bandura, 1999; Shapiro Jr. & Zifferblatt, 1976), eating disorders (Fairburn,
Page 13 of 31
Shafran, & Cooper, 1999; Favaro & Santonastaso, 1998; Jeffrey, 1987; King, 1989; Shapiro Jr,
Blinder, Hagman, & Pituck, 1993), and self-injurious behavior (Favazza, 1989; Herpertz, 1995).
(p.162) Beliefs about self-efficacy not only influence affect and behavior, but also physical
health. Control has been shown to reduce stress hormone production following a stressful
procedure (Abelson et al., 2008). Individuals with low perceived self-efficacy report and
demonstrate lesser functional capacities as a result of chronic pain (Lackner & Corosella, 1999),
are more likely to frequent the emergency room for asthmatic symptoms (Nouwen, Freeston,
Labbe, & Boulet, 1999), are less likely to follow prescribed rehabilitation guidelines following a
cardiac event (Lau-Walker, 2004), and are more likely to relapse following treatment for
substance dependence (Litt, Kadden, Kabela-Cormier, & Petry, 2008; Wilson, 1987).
Interestingly, while a lack of control is detrimental to adaptive functioning, overestimating ones
own control, or experiencing an illusion of control (Langer, 1975), may be somewhat
protective from depression (Alloy & Abramson, 1979; Presson & Benassi, 2003).
When animals or humans perceive a sense of lack of control over a stressor, they exhibit
exaggerated fear responses, increase in stress levels, and greater negative affect (Amat et al.,
2005; Maier & Watkins, 2005; Mohr et al., 2012). Indeed, uncontrollability over stressful stimuli
has been linked with a variety of negative consequences such as negative emotions and harmful
psychological and motivational side effects (Amat et al., 2005; Jensen & Karoly, 1991; Jensen et
al., 1991; Maier & Watkins, 2005), suggesting that perceived control might be important for
regulating negative emotional responses (Delgado et al., 2008, 2009). More recently,
neuroimaging studies have supported this idea by showing decreases in negative affect in a
variety of tasks as a function of control, which is coupled with the modulation of brain regions
involved in emotion regulation or control. For instance, when subjects were provided with either
an uncontrollable or controllable cue indicating the duration of heat (long or short) to be applied
to their forearms, brain regions associated with pain were significantly more activated during
uncontrollable compared to controllable conditions (Salomons et al., 2004). Moreover, self-
controlled stimulation of noxious stimuli induced less pain and anxiety based on subjective
ratings, and pain intensity negatively correlated with enhanced neural processing in lateral
prefrontal regions involved in emotion regulation (Wiech et al., 2006). Finally, perceived control
of expected negative affect due to pain (Delgado et al., 2008, 2009; Jensen & Karoly, 1991;
Jensen et al., 2003) or fear (Delgado et al., 2008; 2009) is decreased when participants perceive
control, which can recruit striatum mechanisms when participants have the opportunity to avoid
a negative outcome (LeDoux & Gorman, 2014; Jensen et al., 2003; Delgado et al., 2009) and
more cortical-based mechanisms when using emotion regulation strategies (e.g., Kalisch et al.,
2005; Delgado et al., 2008; for a review, see Ochsner & Gross, 2005).
(p.163) If disturbances to perceived control negatively influence psychosocial well-being, then

methods to adaptively restore appropriate levels of control and feelings of self-efficacy should be
an effective therapeutic technique, and indeed form a central tenet of various schools of
psychotherapy (Ryan, Lynch, Vansteenkiste, & Deci, 2011; Strupp, 1970). Provision of choice
may be one such way to reduce stress and bolster feelings of control. In fact, we know that when
patients have control over analgesic administration (as opposed to nurse-dependent analgesic
administration), they tend to consume less analgesic medicine, report less pain, and experience
Page 14 of 31
greater satisfaction in treatment (Ballantyne et al., 1993; Shiloh et al., 2003). Participants who
are given choice over treatment show enhanced placebo analgesia as compared to participants
who are given no choice over treatment (Geers & Rose, 2011; Geers, Rose, Fowler, Rasinski,
Brown, & Helfer, 2013; Rose, Geers, Rasinski, & Fowler, 2012). Even simply giving participants
a choice in coping strategies had been shown to improve tolerance for experimentally induced
pain (Rokke & alAbsi, 1992; Rokke, Fleming-Ficek, Siemens, & Hegstad, 2004; Rokke & Lall,
1992). Additional research is necessary to extend these findings to address symptoms related to
mood and anxiety, as well as maladaptive behaviors.
Preliminary findings suggest that choice is an important tool for shaping affective and
behavioral responses to stressful events. However, we still have much to learn about how the
experience of exercising control, through choice, may influence self-efficacy beliefs, and how
this influences our neural responses to future opportunities to exercise control. Individuals with
low perceived control demonstrate an impaired ability to control destructive thought patterns
and to produce desired results (Bandura & Jourden, 1991). An inability to exercise control will
lead to feelings of helplessness (Beck, Emery, & Greenberg, 1985) and depressed mood states,
further reinforcing feelings of inefficacy (Kavanagh & Bower, 1985). As a result, individuals can
experience learned helplessness (Seligman, 1972), the phenomenon that occurs when an
individual has the available resources to cope with a stressor, but is unable to effectively use
those resources as a result of previous unsuccessful attempts to exercise control over a similar
stressor. This phenomenon has been extensively demonstrated in animal research (Maier &
Seligman, 1976; Maier & Watkins, 2005), but also has been observed in humans (Hiroto &
Seligman, 1975; I. W. Miller & Norman, 1979) and is predictive of depressive symptoms (Alloy et
al., 1999). Rodents typically demonstrate stress-related behavior (e.g., freezing), and increased
activity in brainstem nuclei mediating escape behavior when faced with uncontrollable stress
(inescapable shock) but not when faced with controllable stress (escapable shock; Maier &
Watkins, 2005). Yet, when rodents are faced with controllable stress after previously (p.164)
having been exposed to uncontrollable stress, they fail to engage in escape behavior,
demonstrating learned helplessness (Maier & Seligman, 1976). Interestingly, however, rodents
respond to inescapable shock as if it were escapable, if they previously had experience with
escapable shock, suggesting that previous experience with controllable stress provides some
future resilience (Amat et al., 2005; Amat, Paul, Zarza, Watkins, & Maier, 2006; Maier et al.,
2006).
The animal literature has made significant contributions to our understanding of the critical role
of the MPFC to the experience of learned helplessness (Maier et al., 2006). In humans, research
has demonstrated that controllability effects depend on regions of the MPFC (Kerr et al., 2012).
Additionally, self-reports of helplessness in chronic pain patients are inversely related to cortical
thickness of the mid-cingulate, which may explain the reduction in escape-motivated behavior in
learned helplessness (Salomons et al., 2012). Moreover, the research outlined in this chapter
suggests that brain regions involved in affective and motivational processes, such as the MPFC
and the ventral striatum, are important in the affective appraisal of choice opportunity, as a
vehicle for exercising control. These regions are also known to play critical roles in successful
affective regulation (Delgado, Jou, Ledoux, & Phelps, 2009; Lieberman et al., 2007; Ochsner &
Gross, 2005; Wager et al., 2008). Thus, the recruitment of these regions in the anticipation of
Page 15 of 31
control may suggest that choice opportunity involves processes important to behavioral and
cognitive emotional regulation strategies.
Conclusion
Collectively, the findings suggest that a sense of agency, and opportunities to exercise control,
through choice, are highly adaptive. Recent human neuroscience research suggests that the
opportunity to control is desirable, and recruits brain circuitry involved in affective and
motivational processes. Choice modulates this reward-related brain circuitry both when
approaching positive outcomes and when avoiding negative outcomes. However, the value of
exercising control may vary across individuals due to person-specific variables and may be
sensitive to situational factors (e.g., context, type of control). Opportunities for control also
recruit brain regions involved in the successful regulation of affect, implying that choice may be
a powerful tool for regulating responses to both appetitive and aversive stimuli. Future human
research will need to address whether simple provisions of choice can be a helpful tool for
augmenting feelings of self-efficacy and agency, to prevent learned helplessness and to enhance
coping under stressors that are both controllable and uncontrollable.
(p.165) Acknowledgments
This work was supported by funding from the National Institute on Drug Abuse to M.R.D.
(DA027764) and a fellowship to L.A.L (F32-DA027308).
References
Abelson, J. L., Khan, S., Liberzon, I., Erickson, T. M., & Young, E. A. (2008). Effects of perceived
control and cognitive coping on endocrine stress responses to pharmacological activation.
Biological Psychiatry, 64(8), 701707.
Abramson, L. Y., Garber, J., & Seligman, M. E. P. (1980). Learned helplessness in humans: an
attributional analysis. Human Helplessness: Theory and Applications, 3, 34.
Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency in depressed and non-depressed
students: sadder but wiser? Journal of Experimental Psychology: General, 108(4), 441485.
Alloy, L. B., & Abramson, L. Y. (1982). Learned helplessness, depression, and the illusion of
control. Journal of Personality and Social Psychology, 42(6), 1114.
Alloy, L. B., Abramson, L. Y., Whitehouse, W. G., Hogan, M. E., Tashman, N. A., Steinberg, D. L.,
et al. (1999). Depressogenic cognitive styles: predictive validity, information processing and
personality characteristics, and developmental origins. Behaviour Research and Therapy, 37(6),
503531.
Amat, J., Baratta, M. V., Paul, E., Bland, S. T., Watkins, L. R., & Maier, S. F. (2005). Medial
prefrontal cortex determines how stressor controllability affects behavior and dorsal raphe
nucleus. Nature Neuroscience, 8(3), 365371.
Page 16 of 31
Amat, J., Paul, E., Zarza, C., Watkins, L. R., & Maier, S. F. (2006). Previous experience with
behavioral control over stress blocks the behavioral and dorsal raphe nucleus activating effects
of later uncontrollable stress: role of the ventral medial prefrontal cortex. Journal of
Neuroscience, 26(51), 1326413272.
Averill, J. R. (1973). Personal control over aversive stimuli and its relationship to stress.
Psychological Bulletin, 80(4), 286.
Ballantyne, J. C., Carr, D. B., Chalmers, T. C., Dear, K. B., Angelillo, I. F., & Mosteller, F. (1993).
Postoperative patient-controlled analgesia: meta-analyses of initial randomized control trials.
Journal of Clinical Anesthesia, 5(3), 182193.
Balleine, B. W., Delgado, M. R., & Hikosaka, O. (2007). The role of the dorsal striatum in reward
and decision-making. Journal of Neuroscience, 27(31), 81618165.
Balleine, B. W., & ODoherty, J. P. (2009). Human and rodent homologies in action control:
corticostriatal determinants of goal-directed and habitual action. Neuropsychopharmacology,
35(1), 4869.
Bandura, A. (1988). Self-efficacy conception of anxiety. Anxiety Research, 1(2), 7798.
Bandura, A. (1997). Self-efficacy: the exercise of control. New York: Freeman.
Bandura, A. (1999). A sociocognitive analysis of substance abuse: an agentic perspective.

Bandura, A., Caprar, G. V., Barbaranelli, C., Gerbino, M., & Pastorelli, C. (2003). Role of
affective self-regulatory efficacy in diverse spheres of psychosocial functioning. Child
Development, 74(3), 769782.
Bandura, A., & Jourden, F. J. (1991). Self-regulatory mechanisms governing the impact of social-
comparison on complex decision making. Journal of Personality and Social Psychology, 55, 479
488.
Bandura, A., Taylor, C. B., Williams, S. L., Mefford, I. N., & Barchas, J. D. (1985). Catecholamine
secretion as a function of perceived coping self-efficacy. Journal of Consulting and Clinical
Bandura, A., & Wood, R. E. (1989). Effect of perceived controllability and performance
standards on self-regulation of complex decision-making. Journal of Personality and Social
Baron, J., & Ritov, I. (1994). Reference points and omission bias. Organizational Behavior and
Human Decision Processes, 59, 475498.
Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: how and why people
fail at self-regulation. San Diego, CA: Academic Press.
Page 17 of 31
Beattie, J., Baron, J., Hershey, J. C., & Spranca, M. D. (1994). Psychological determinants of
decision attitude. Journal of Behavioral Decision Making, 7(2), 129144.
Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International
Universities Press.
Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias. New York:
Basic Books.
Berridge, K. C., & Kringelbach, M. L. (2008). Affective neuroscience of pleasure: reward in

humans and animals. Psychopharmacology (Berlin), 199(3), 457480.
Bjork, J. M., & Hommer, D. W. (2007). Anticipating instrumentally obtained and passively-
received rewards: a factorial fMRI investigation. Behavioural Brain Research, 177(1), 165170.
Blackwood, N. J., Bentall, R. P., Simmons, A., Murray, R. M., & Howard, R. J. (2003). Self-
responsibility and the self-serving bias: an fMRI investigation of causal attributions.
NeuroImage, 20(2), 10761085.
Botti, S., McGill, A., & Iyengar, S. S. (2003). Preference for control and its effect on the
evaluation of consumption experiences. Advances in Consumer Research, 30, 127128.
Bown, N. J., Read, D., & Summers, B. (2003). The lure of choice. Journal of Behavioral Decision
Making, 16, 297308.
Brehm, J. W. (1956). Post-decision changes in the desirability of alternatives. Journal of

Abnormal and Social Psychology, 52, 384389.
Brewin, C. R., & Shapiro, D. A. (1984). Beyond locus of control: attribution of responsibility for
positive and negative outcomes. British Journal of Psychology, 75(1), 4349.
Burger, J. (1989). Negative reactions to increases in perceived personal control. Journal of

Chambon, V., Wenke, D., Fleming, S. M., Prinz, W., & Haggard, P. (2013). An online neural
substrate for sense of agency. Cerebral Cortex, 23(5), 10311037.
Cockburn, J., Collins, A. G. E., & Frank, M. J. (2014). A reinforcement learning mechanism
responsible for the valuation of free choice. Neuron, 83(3), 551557.
Corah, N. L., & Boffa, J. (1970). Perceived control, self-observation, and response to aversive
stimulation. Journal of Personality and Social Psychology, 16(1), 14.
Cordova, D. I., & Lepper, M. R. (1996). Intrinsic motivation and the process of learning:
beneficial effects of contextualization, personalization, and choice. Journal of Educational
Psychology, 88(4), 715730.
Page 18 of 31
Coricelli, G., Critchley, H. D., Joffily, M., ODoherty, J. P., Sirigu, A., & Dolan, R. J. (2005). Regret
and its avoidance: a neuroimaging study of choice behavior. Nature Neuroscience, 8(9), 1255
1262.
De Martino, B., Kumaran, D., Holt, B., & Dolan, R. J. (2009). The neurobiology of reference-
dependent value computation. Journal of Neuroscience, 29(12), 38333842.
De Martino, B., Kumaran, D., Seymour, B., & Dolan, R. J. (2006). Frames, biases, and rational
decision-making in the human brain. Science, 313, 684687.
Delgado, M. R. (2007). Reward-related responses in the human striatum. Annals of the New York
Academy of Sciences, 1104, 7088.
Delgado, M. R., & Dickerson, K. C. (2012). Reward-related learning via multiple memory
systems. Biological Psychiatry, 72(2), 134141.
Delgado, M. R., Jou, R. L., Ledoux, J. E., & Phelps, E. A. (2009). Avoiding negative outcomes:
tracking the mechanisms of avoidance learning in humans during fear conditioning. Frontiers in
Behavioral Neuroscience, 3, 33.
Delgado, M. R., Locke, H. M., Stenger, V. A., & Fiez, J. A. (2003). Dorsal striatum responses to
reward and punishment: effects of valence and magnitude manipulations. Cognitive, Affective, &
Behavioral Neuroscience, 3(1), 2738.
Delgado, M. R., Nearing, K. I., LeDoux, J. E., & Phelps, E. A. (2008). Neural circuitry underlying
the regulation of conditioned fear and its relation to extinction. Neuron, 59, 829838.
Delgado, M. R., Nystrom, L. E., Fissell, C., Noll, D. C., & Fiez, J. A. (2000). Tracking the
hemodynamic responses to reward and punishment in the striatum. Journal of Neurophysiology,
84(6), 30723077.
Delgado, M. R., Stenger, V. A., & Fiez, J. A. (2004). Motivation-dependent responses in the
human caudate nucleus. Cerebral Cortex, 14(9), 10221030.
Deutsch, A.-M. (1978). Self-control and depression: an appraisal. Behavior Therapy, 9(3), 410
414.
Dhar, R. (1997). Consumer preference for a no-choice option. Journal of Consumer Research,
24(2), 215231.
Doll, B. B., Simon, D. A., & Daw, N. D. (2012). The ubiquity of model-based reinforcement
learning. Current Opinion in Neurobiology, 22(6), 10751081.
Egan, L. C., Santos, L. R., & Bloom, P. (2007). The origins of cognitive dissonance: evidence
from children and monkeys. Psychological Science, 18(11), 978983.
Research, 229(3), 475484.
Page 19 of 31
Elliott, R., Newman, J. L., Longe, O. A., & William Deakin, J. F. (2004). Instrumental responding
for rewards is associated with enhanced neuronal response in subcortical reward systems.
NeuroImage, 21(3), 984990.
Fairburn, C. G., Shafran, R., & Cooper, Z. (1999). A cognitive behavioural theory of anorexia
nervosa. Behaviour Research and Therapy, 37(1), 113.
Fasolo, B., Hertwig, R., Huber, M., & Ludwig, M. (2009). Size, entropy, and density: what is the
difference that makes the difference between small and large realworld assortments?
Psychology & Marketing, 26(3), 254279.
Favaro, A., & Santonastaso, P. (1998). Impulsive and compulsive self-injurious behavior in
bulimia nervosa: prevalence and psychological correlates. The Journal of Nervous and Mental
Disease, 186(3), 157165.
Favazza, A. R. (1989). Why patients mutilate themselves. Hospital and Community Psychiatry,
40(2), 137145.
Feldman, J., Miyamoto, J. M., & Loftus, E. F. (1999). Are actions regretted more than inactions?
Organizational Behavior and Human Decision Processes, 78, 232255.
Fleming, S. M., Thomas, C. L., & Dolan, R. J. (2010). Overcoming status quo bias in the human
brain. Proceedings of the National Academy of Sciences USA, 107(13), 60056009.
Fontaine, K. R., Manstead, A. S., & Wagner, H. (1993). Optimism, perceived control over stress,
and coping. European Journal of Personality, 7(4), 267281.
Frank, M. J., & Fossella, J. A. (2011). Neurogenetics and pharmacology of learning, motivation,
and cognition. Neuropsychopharmacology, 36(1), 133152.
Fujiwara, J., Usui, N., Park, S. Q., Williams, T., Iijima, T., Taira, M., Tsutsui, K., & Tobler, P. N.
(2013). Value of freedom to choose encoded by the human brain. Journal of Neurophysiology,
110(8), 1915.
Geer, J. H., & Maisel, E. (1972). Evaluating the effects of the prediction-control confound.
Geers, A., & Rose, J. (2011). Treatment choice and placebo expectation effects. Social and
Personality Psychology Compass, 5(10), 734750.
Geers, A. L., Rose, J. P., Fowler, S. L., Rasinski, H. M., Brown, J. A., & Helfer, S. G. (2013). Why
does choice enhance treatment effectiveness? Using placebo treatments to demonstrate the role
of personal control. Journal of Personality and Social Psychology, 105(4), 549566.
Glass, D. C., Reim, B., & Singer, J. E. (1971). Behavioral consequences of adaptation to
controllable and uncontrollable noise. Journal of Experimental Psychology: General, 7, 244257.
Page 20 of 31
Glass, D. C., Singer, J. E., & Friedman, L. N. (1969). Psychic costs of adaptation to an
environmental stressor. Journal of Personality and Social Behavior, 12, 200210.
Grolnick, W. S., & Ryan, R. M. (1987). Autonomy in childrens learning: an experimental and
individual difference investigation. Journal of Personality and Social Psychology, 52(5), 890.
Gunnar-vonGnechten, M. R. (1978). Changing a frightening toy into a pleasant toy by allowing

the infant to control its actions. Developmental Psychology, 14, 147152.
Haber, S. N., & Knutson, B. (2010). The reward circuit: linking primate anatomy and human
imaging. Neuropsychopharmacology, 35(1), 426.
Hare, T. A., Tottenham, N., Davidson, M. C., Glover, G. H., & Casey, B. J. (2005). Contributions
of amygdala and striatal activity in emotion regulation. Biological Psychiatry, 57(6), 624632.
Heatherton, T. F., & Wagner, D. D. (2011). Cognitive neuroscience of self-regulation failure.

Trends in Cognitive Sciences, 15(3), 132139.
Heatherton, T. F., Wyland, C. L., Macrae, C. N., Demos, K. E., Denny, B. T., & Kelley, W. M.
(2006). Medial prefrontal activity differentiates self from close others. Social Cognitive and
Affective Neuroscience, 1(1), 1825.
Herpertz, S. (1995). Self-injurious behaviour: psychopathological and nosological characteristics

in subtypes of self-injurers. Acta Psychiatrica Scandinavica, 91(1), 5768.
Hiroto, D. S., & Seligman, M. E. (1975). Generality of learned helplessness in man. Journal of
Holden, G. (1992). The relationship of self-efficacy appraisals to subsequent health related

outcomes: a meta-analysis. Social Work in Health Care, 16(1), 5393.
Hsu, M., Bhatt, M., Adolphs, R., Tranel, D., & Camerer, C. F. (2005). Neural systems responding
to degrees of uncertainty in human decision-making. Science, 310(5754), 16801683.
Inman, J. J., & Zeelenberg, M. (2002). Regret in repeat purchase versus switching decisions: the
attenuating role of decision justifiability. Journal of Consumer Research, 29, 116128.
Iyengar, S. S., & Lepper, M. (2000). When choice is demotivating: can one desire too much of a
good thing? Journal of Personality and Social Psychology, 79, 9951006.
Iyengar, S. S., & Lepper, M. R. (1999). Rethinking the value of choice: a cultural perspective on
intrinsic motivation. Journal of Personality and Social Psychology, 76(3), 349.
Iyengar, S. S., Wells, R. E., & Schwartz, B. (2006). Doing better but feeling worse: looking for
the best job undermines satisfaction. Psychological Science, 17(2), 143150.
Jeffrey, R. W. (1987). Behavioral treatment of obesity. Annals of Behavioral Medicine, 9(1), 20

24.
Page 21 of 31
Jensen, M. P., & Karoly, P. (1991). Control beliefs, coping efforts, and adjustment to chronic
pain. Journal of Consulting and Clinical Psychology, 59(3), 431.
Jensen, M. P., Turner, J. A., Romano, J. M., & Karoly, P. (1991). Coping with chronic pain: a
critical review of the literature. Pain, 47(3), 249283.
Jensen, J., McIntosh, A. R., Crawley, A. P., Mikulis, D. J., Remington, G., & Kapur, S. (2003).
Direct activation of the ventral striatum in anticipation of aversive stimuli. Neuron, 40(6), 1251
1257.
Johnson, S. C., Baxter, L. C., Wilder, L. S., Pipe, J. G., Heiserman, J. E., & Prigatano, G. P.
(2002). Neural correlates of selfreflection. Brain, 125(8), 18081814.
Johnson, S. C., Schmitz, T. W., Kawahara-Baccus, T. N., Rowley, H. A., Alexander, A. L., Lee, J.,
et al. (2005). The cerebral response during subjective choice with and without self-reference.
Journal of Cognitive Neuroscience, 17(12), 18971906.
Johnstone, T., van Reekum, C. M., Urry, H. L., Kalin, N. H., & Davidson, R. J. (2007). Failure to
regulate: counterproductive recruitment of top-down prefrontal-subcortical circuitry in major
depression. Journal of Neuroscience, 27(33), 88778884.
Kahneman, D., & Tversky, A. (1982). The psychology of preferences. Scientific American, 246(1),
160173.
Kahneman, D., & Tversky, A. (2000). Choices, values, and frames. Cambridge: Cambridge
University Press.
Kalisch, R., Wiech, K., Critchley, H. D., Seymour, B., ODoherty, J. P., Oakley, D. A., et al. (2005).
Anxiety reduction through detachment: Subjective, physiological, and neural effects. Journal of
Cognitive Neuroscience, 17, 874883.
Kamen-Siegel, L., Rodin, J., Seligman, M. E., & Dwyer, J. (1991). Explanatory style and cell-
mediated immunity in elderly men and women. Health Psychology, 10(4), 229.
Kanfer, F., & Seider, M. L. (1973). Self-control: factors enhancing tolerance of noxious
stimulation. Journal of Personality and Social Behavior, 25, 381.
Kavanagh, D. J., & Bower, G. H. (1985). Mood and self-efficacy: impact of joy and sadness on
perceived capabilities. Cognitive Therapy and Research, 9, 507525.
Kelley, W. M., Macrae, C. N., Wyland, C. L., Caglar, S., Inati, S., & Heatherton, T. F. (2002).
Finding the self? An event-related fMRI study. Journal of Cognitive Neuroscience, 14(5), 785
794.
Kerr, D. L., McLaren, D. G., Mathy, R. M., & Nitschke, J. B. (2012). Controllability modulates the
anticipatory response in the human ventromedial prefrontal cortex. Frontiers in Emotion
Science, 3(557), 111.
Page 22 of 31
King, M. (1989). Locus of control in women with eating pathology. Psychological Medicine, 19,
183187.
Kirsch, P., Schienle, A., Stark, R., Sammer, G., Blecker, C., Walter, B., et al. (2003). Anticipation
of reward in a nonaversive differential conditioning paradigm and the brain reward system: an
event-related fMRI study. NeuroImage, 20(2), 10861095.
Knutson, B., Adams, C. M., Fong, G. W., & Hommer, D. (2001). Anticipation of increasing
monetary reward selectively recruits nucleus accumbens. Journal of Neuroscience, 21(16),
RC159.
Knutson, B., & Cooper, J. C. (2005). Functional magnetic resonance imaging of reward
prediction. Current Opinion in Neurology, 18(4), 411417.
Knutson, B., Taylor, J., Kaufman, M., Peterson, R., & Glover, G. (2005). Distributed neural
representation of expected value. Journal of Neuroscience, 25(19), 48064812.
Lackner, J. M., & Corosella, A. M. (1999). The relative influence of perceived pain control,
anxiety, and functional self efficacy on spinal function among patients with chronic low back
pain. Spine, 24(21), 22542261.
Landman, J. (1987). Regret and elation following action and inaction: affective responses to
positive versus negative outcomes. Personality and Social Psychology Bulletin, 13, 524536.
Langer, E. J. (1975). The illusion of control. Journal of Personality and Social Psychology, 32(2),
311328.
Langer, E. J., & Rodin, J. (1976). The effects of choice and enhanced personal responsibility for
the aged: a field experiment in an institutional setting. Journal of Personality and Social
Psychology, 34(2), 191198.
Lau-Walker, M. (2004). Relationship between illness representation and self-efficacy. Journal of

Advanced Nursing, 48(3), 216225.
LeDoux, J. E., & Gorman, J. M. (2014). A call to action: overcoming anxiety through active
coping.
Leotti, L. A., & Delgado, M. R. (2011a). The inherent reward of choice. Psychological Science,
22(10), 13101318.
Leotti, L. A., & Delgado, M. R. (2014). The value of exercising control over monetary gains and
losses. Psychological Science, 25(2), 596604.
Leotti, L. A., Iyengar, S. S., & Ochsner, K. N. (2010). Born to choose: the origins and value of the
need for control. Trends in Cognitive Sciences, 14(10), 457463.
Levinson, W., Kao, A., Kuby, A., & Thisted, R. A. (2005). Not all patients want to participate in
decision making. Journal of General Internal Medicine, 20(6), 531535.
Page 23 of 31
Levy, R., & Dubois, B. (2006). Apathy and the functional anatomy of the prefrontal cortex-basal
ganglia circuits. Cerebral Cortex, 16(7), 916928.
Lieberman, M. D., Eisenberger, N. I., Crockett, M. J., Tom, S. M., Pfeifer, J. H., & Way, B. M.
(2007). Putting feelings into words: affect labeling disrupts amygdala activity in response to
affective stimuli. Psychological Science, 18(5), 421428.
Lieberman, M. D., Ochsner, K. N., Gilbert, D. T., & Schacter, D. L. (2001). Do amnesics exhibit
cognitive dissonance reduction? The role of explicit memory and attention in attitude change.
Litt, M. D., Kadden, R. M., Kabela-Cormier, E., & & Petry, N. M. (2008). Coping skills training
and contingency management treatments for marijuana dependence: exploring mechanisms of
behavior change. Addiction, 103(4), 638648.
Higgins, E. T. (2012). Beyond pleasure and pain: How motivation works. Oxford University
Press.
Maier, S. F., Amat, J., Baratta, M. V., Paul, E., & Watkins, L. R. (2006). Behavioral control, the
medial prefrontal cortex, and resilience. Dialogues in Clinical Neuroscience, 8(4), 397406.
Maier, S. F., Laudenslager, M. L., & Ryan, S. M. (1985). Stressor controllability, immune
function, and endogenous opiates. In F. R. Brush & J. B. Overmier (Eds.), Affect, conditioning,
and cognition: essays on the determinants of behavior (pp. 183201). Hillsdale, NJ: Erlbaum
Associates.
Maier, S. F., & Seligman, M. E. (1976). Learned helplessness: theory and evidence. Journal of
Experimental Psychology: General, 105(1), 3.
Maier, S. F., & Watkins, L. R. (2005). Stressor controllability and learned helplessness: the roles
of the dorsal raphe nucleus, serotonin, and corticotropin-releasing factor. Neuroscience &
Biobehavioral Reviews, 29(4), 829841.
Mansell, W. (2005). Control theory and psychopathology: an integrative approach. Psychology

and Psychotherapy: Theory, Research and Practice, 78(2), 141178.
Mathews, C. O. (1977). A review of behavioral theories of depression and a self-regulation model

for depression. Psychotherapy: Theory, Research & Practice, 14(1), 79.
McClure, S. M., Berns, G. S., & Montague, P. R. (2003). Temporal prediction errors in a passive
learning task activate human striatum. Neuron, 38(2), 339346.
McGuire, J. T., & Botvinick, M. M. (2010). Prefrontal cortex, cognitive control, and the
registration of decision costs. Proceedings of the National Academy of Sciences USA, 107(17),
79227926.
Middleton, F. A., & Strick, P. L. (2000). Basal ganglia and cerebellar loops: motor and cognitive
circuits. Brain Research Reviews, 31(2), 236250.
Page 24 of 31
Miller, D. T., & Ross, M. (1975). Self-serving biases in attribution of causality: fact or fiction?
Miller, I. W., & Norman, W. H. (1979). Learned helplessness in humans: a review and
attribution-theory model. Psychological Bulletin, 86(1), 93.
Mohr, C., Leyendecker, S., Petersen, D., & Helmchen, C. (2012). Effects of perceived and
exerted pain control on neural activity during pain relief in experimental heat hyperalgesia: a
fMRI study. European Journal of Pain, 16(4), 496508.
Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of self-efficacy beliefs to academic
outcomes: A meta-analytic investigation. Journal of Counseling Psychology, 38(1), 30.
Nicolle, A., Fleming, S. M., Bach, D. R., Driver, J., & Dolan, R. J. (2011). A regret-induced status
quo bias. Journal of Neuroscience, 31(9), 33203327.
Nouwen, A., Freeston, M. H., Labbe, R., & Boulet, L.-P. (1999). Psychological factors associated
with emergency room visits among asthmatic patients. Behavior Modification, 23(2), 217233.
ODoherty, J. P. (2004). Reward representations and reward-related learning in the human

brain: insights from neuroimaging. Current Opinion in Neurobiology, 14(6), 769776.
ODoherty, J. P. (2007). Lights, camembert, action! The role of human orbitofrontal cortex in
encoding stimuli, rewards, and choices. Annals of the New York Academy of Sciences, 1121(1),
254272.
ODoherty, J. P., Critchley, H., Deichmann, R., & Dolan, R. J. (2003). Dissociating valence of
outcome from behavioral control in human orbital and ventral prefrontal cortices. Journal of
Neuroscience, 23(21), 79317939.
ODoherty, J. P., Dayan, P., Friston, K., Critchley, H., & Dolan, R. J. (2003). Temporal difference
models and reward-related learning in the human brain. Neuron, 38(2), 329337.
ODoherty, J. P., Dayan, P., Schultz, J., Deichmann, R., Friston, K., & Dolan, R. J. (2004).
Dissociable roles of ventral and dorsal striatum in instrumental conditioning. Science,
304(5669), 452454.
ODoherty, J. P., Deichmann, R., Critchley, H. D., & Dolan, R. J. (2002). Neural responses during
anticipation of a primary taste reward. Neuron, 33(5), 815826.
ODoherty, J. P., Rolls, E. T., Francis, S., Bowtell, R., & McGlone, F. (2001). Representation of
pleasant and aversive taste in the human brain. Journal of Neurophysiology, 85(3), 13151321.
Ochsner, K. N., & Gross, J. J. (2005). The cognitive control of emotion. Trends in Cognitive
Sciences, 9(5), 242249.
Ochsner, K. N., & Gross, J. J. (2008). Cognitive emotion regulation insights from social cognitive
and affective neuroscience. Current Directions in Psychological Science, 17(2), 153158.
Page 25 of 31
Ogden, J., Daniells, E., & Barnett, J. (2008). The value of choice: development of a new
measurement tool. British Journal of General Practice, 58(554), 614618.
Ongur, D., & Price, J. L. (2000). The organization of networks within the orbital and medial
prefrontal cortex of rats, monkeys and humans. Cerebral Cortex, 10(3), 206219.
Patall, E. A. (2013). Constructing motivation through choice, interest, and interestingness.

Journal of Educational Psychology, 105(2), 522.
Patall, E. A., Cooper, H., & Robinson, J. C. (2008). The effects of choice on intrinsic motivation
and related outcomes: a meta-analysis of research findings. Psychological Bulletin, 134(2), 270
300.
Patall, E. A., Cooper, H., & Wynn, S. R. (2010). The effectiveness and relative importance of
choice in the classroom. Journal of Educational Psychology, 102(4), 896.
Paterson, R. J., & Neufeld, R. W. (1995). What are my options? Influences of choice availability
on stress and the perception of control. Journal of Research in Personality, 29(2), 145167.
Paulus, M. P., & Frank, L. R. (2003). Ventromedial prefrontal cortex activation is critical for
preference judgments. Neuroreport, 14(10), 13111315.
Platek, S. M., Keenan, J. P., Gallup Jr, G. G., & Mohamed, F. B. (2004). Where am I? The
neurological correlates of self and other. Cognitive Brain Research, 19(2), 114122.
Presson, P. K., & Benassi, V. A. (2003). Are depressive symptoms positively or negatively
associated with the illusion of control. Social Behavior and Personality: An International Journal,
31(5), 483495.
Rao, H., Korczykowski, M., Pluta, J., Hoang, A., & Detre, J. A. (2008). Neural correlates of
voluntary and involuntary risk taking in the human brain: an fMRI Study of the Balloon Analog
Risk Task (BART). NeuroImage, 42(2), 902910.
Reuven-Magril, O., Dar, R., & Liberman, N. (2008). Illusion of control and behavioral control
attempts in obsessive-compulsive disorder. Journal of Abnormal Psychology, 117(2), 334.
Robbins, T. W., & Everitt, B. J. (1996). Neurobehavioural mechanisms of reward and motivation.
Current Opinion in Neurobiology, 6(2), 228236.
Rokke, P. D., & alAbsi, M. (1992). Matching pain coping strategies to the individual: a
prospective validation of the cognitive coping strategy inventory. Journal of Behavioral
Medicine, 15, 611625.
Rokke, P. D., Fleming-Ficek, S., Siemens, N. M., & Hegstad, H. J. (2004). Self-efficacy and
choice of coping strategies for tolerating acute pain. Journal of Behavioral Medicine, 27, 343
360.
Page 26 of 31
Rokke, P. D., & Lall, R. (1992). The role of choice in enhancing tolerance to acute pain. Cognitive
Therapy and Research, 16, 5365.
Rose, J. P., Geers, A. L., Rasinski, H. M., & Fowler, S. L. (2012). Choice and placebo expectation
effects in the context of pain analgesia. Journal of Behavioral Medicine, 35(4), 462470.
Rotter, J. (1966). Generalised expectancies for internal versus external control of reinforcement.
Psychological Monographs, 80(1, Whole no. 609), 128.
Rushworth, M. F., Walton, M. E., Kennerley, S. W., & Bannerman, D. M. (2004). Action sets and
decisions in the medial frontal cortex. Trends in Cognitive Sciences, 8(9), 410417.
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic
motivation, social development, and well-being. American Psychologist, 55, 6878.
Ryan, R. M., & Deci, E. L. (2006). Self-regulation and the problem of human autonomy: does
psychology need choice, self-determination, and will? Journal of Personality, 74(6), 15571585.
Ryan, R. M., Deci, E. L., & Grolnick, W. S. (1995). Autonomy, relatedness, and the self: their
relation to development and psychopathology. In D. Cicchetti, & D. J. Cohen (Eds.),
Developmental psychopathology, Vol. 1: Theory and methods. Wiley series on personality
processes (pp. 618655). Oxford, England: John Wiley & Sons, 787 pp.
Ryan, R. M., Lynch, M. F., Vansteenkiste, M., & Deci, E. L. (2011). Motivation and autonomy in
counseling, psychotherapy, and behavior change: a look at theory and practice. The Counseling
Psychologist, 39, 193260.
Sagi, A., & Friedland, N. (2007). The cost of richness: the effect of the size and diversity of
decision sets on post-decision regret. Journal of Personality and Social Psychology, 93(4), 515.
Salomons, T. V., Johnstone, T., Backonja, M. M., & Davidson, R. J. (2004). Perceived
controllability modulates the neural response to pain. Journal of Neuroscience, 24(32), 7199
7203.
Salomons, T. V., Johnstone, T., Backonja, M. M., Shackman, A. J., & Davidson, R. J. (2007).
Individual differences in the effects of perceived controllability on pain perception: critical role
of the prefrontal cortex. Journal of Cognitive Neuroscience, 19(6), 9931003.
Salomons, T. V., Moayedi, M., Weissman-Fogel, I., Goldberg, M. B., Freeman, B. V., Tenenbaum,
H. C., et al. (2012). Perceived helplessness is associated with individual differences in the
central motor output system. European Journal of Neuroscience, 35, 14811487.
Salomons, T. V., Nusslock, R., Detloff, A., Johnstone, T., Abramson, L. Y., & Davidson, R. J.
(2010). Perceived control over pain reduces anxiety: top down mechanisms. Abstract presented
at International Association for the Study of Pain.
Samuelson, W., & Zeckhauer, R. (1988). Status quo bias in decision making. Journal of Risk and
Uncertainty, 1, 759.
Page 27 of 31
Schultz, W., Dayan, P., Montague, P. R. (1997). A neural substrate of prediction and reward.
Science, 275(5306), 15931599.
Schwartz, B. (2000). Self-determination: the tyranny of freedom. American Psychologist, 55(1),

79.
Schwartz, B. (2009). The paradox of choice. New York: HarperCollins.
Schwartz, B., Ward, A., Monterosso, J., Lyubomirsky, S., White, K., & Lehman, D. R. (2002).
Maximizing versus satisficing: happiness is a matter of choice. Journal of Personality and Social
Psychology, 83(5), 1178.
Sela, A., Berger, J., & Liu, W. (2009). Variety, vice, and virtue: how assortment size influences
option choice. Journal of Consumer Research, 35(6), 941951.
Seligman, M. E. P. (1972). Learned helplessness. Annual Review of Medicine, 23(1), 407412.
Seligman, M. E. P. (1975). Helplessness: on depression, development, and death. San Francisco:

W. H. Freeman/Times Books/Henry Holt.
Shapiro, D. H., Jr., Schwartz, C. E., & Astin, J. A. (1996). Controlling ourselves, controlling our
world: psychologys role in understanding positive and negative consequences of seeking and
gaining control. American Psychologist, 51(12), 12131230.
Shapiro, D. H., Jr., Blinder, B. J., Hagman, J., & Pituck, S. (1993). A psychological sense-of-
control profile of patients with anorexia nervosa and bulimia nervosa. Psychological Reports,
73(2), 531541.
Shapiro, D. H., Jr., & Zifferblatt, S. M. (1976). An applied clinical combination of Zen meditation
and behavioral self-management techniques: reducing metadone dosage in drug addiction.
Behavior Therapy, 7(5), 694695.
Sharot, T., De Martino, B., & Dolan, R. J. (2009). How choice reveals and shapes expected
hedonic outcome. Journal of Neuroscience, 29(12), 37603765.
Sharot, T., Fleming, S. M., Yu, X., Koster, R., & Dolan, R. J. (2012). Is choice-induced preference
change long lasting? Psychological Science, 23(10), 11231129.
Sharot, T., Shiner, T., & Dolan, R. J. (2010). Experience and choice shape expected aversive
outcomes. Journal of Neuroscience, 30(27), 92099215.
Sharot, T., Velasquez, C. M., & Dolan, R. J. (2010). Do decisions shape preference? Evidence
from blind choice. Psychological Science, 21(9), 12311235.
Shiloh, S., Zukerman, G., Butin, B., Deutch, A., Yardeni, I., Benyamini, Y., et al. (2003).
Postoperative patient-controlled analgesia (PCA): how much control and how much analgesia?
Psychology and Health, 18(6), 753770.
Page 28 of 31
Shin, J., & Ariely, D. (2004). Keeping doors open: the effect of unavailability on incentives to
keep options viable. Management Science, 50(5), 575586.
Shiv, B., & Fedorikhin, A. (1999). Heart and mind in conflict: the interplay of affect and
cognition in consumer decision making. Journal of Consumer Research, 26(3), 278292.
Stajkovic, A. D., & Luthans, F. (1998). Self-efficacy and work-related performance: a meta-
analysis. Psychological Bulletin, 124(2), 240.
Staub, E., Tursky, B., & Schwartz, G. E. (1971). Self-control and predictability: their effects on
reactions to aversive stimulation. Journal of Personality and Social Behavior, 18, 157162.
Strupp, H. H. (1970). Specific versus non-specific factors in psychotherapy and the problem of
control. Archives of General Psychiatry, 23, 393401.
Suzuki, S. (1997). Effects of number of alternatives on choice in humans. Behavioural Processes,

39, 205214.
Suzuki, S. (2011). Choice between single-response and multichoice tasks in humans. The
Psychological Record, 50(1), 7.
Tanaka, S. C., Balleine, B. W., & ODoherty, J. P. (2008). Calculating consequences: brain
systems that encode the causal effects of actions. Journal of Neuroscience, 28(26), 67506755.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: a social psychological perspective
on mental health. Psychological Bulletin, 103(2), 193.
Taylor, S. E., & Brown, J. D. (1994). Positive illusions and well-being revisited: separating fact
from fiction. Psychological Bulletin, 116(1), 2127.
Thompson, S. (1999). Illusions of control: how we overestimate our personal influence. Current
Directions in Psychological Science, 8(6), 187190.
Thorndike, E. L. (1933). A theory of the action of the after-effects of a connection upon it.
Tobler, P. N., ODoherty, J. P., Dolan, R. J., & Schultz, W. (2007). Reward value coding distinct
from risk attitude-related uncertainty coding in human reward systems. Journal of
Neurophysiology, 97(2), 16211632.
Tricomi, E. M., Delgado, M. R., & Fiez, J. A. (2004). Modulation of caudate activity by action
contingency. Neuron, 41(2), 281292.
Tsiros, M., & Mittal, V. (2000). Regret: a model of its antecedents and consequences in
consumer decision making. Journal of Consumer Research 26, 401417.
Page 29 of 31
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice.
Science, 211(4481), 453458.
Vohs, K. D., Baumeister, R. F., Schmeichel, B. J., Twenge, J. M., Nelson, N. M., & Tice, D. M.
(2008). Making choices impairs subsequent self-control: a limited-resource account of decision
making, self-regulation, and active initiative. Journal of Personality and Social Psychology, 94(5),
883.
Vohs, K. D., & Baumeister, R. F. E. (2011). Handbook of self-regulation: research, theory, and
applications. New York: Guilford Press.
Voorn, P., Vanderschuren, L. J., Groenewegen, H. J., Robbins, T. W., & Pennartz, C. (2004).
Putting a spin on the dorsalventral divide of the striatum. Trends in Neurosciences, 27(8), 468
474.
Voss, M., Ingram, J. N., Haggard, P., & Wolpert, D. M. (2006). Sensorimotor attenuation by
central motor command signals in the absence of movement. Nature Neuroscience, 9(1), 2627.
Voss, M., Ingram, J. N., Wolpert, D. M., & Haggard, P. (2008). Mere expectation to move causes
attenuation of sensory signals. PLoS One, 3(8), e2866.
Wager, T. D., Davidson, M. L., Hughes, B. L., Lindquist, M. A., & Ochsner, K. N. (2008).
Prefrontal-subcortical pathways mediating successful emotion regulation. Neuron, 59(6), 1037
1050.
Wenke, D., Fleming, S. M., & Haggard, P. (2010). Subliminal priming of actions influences sense
of control over effects of action. Cognition, 115(1), 2638.
Wiech, K., Kalisch, R., Weiskopf, N., Pleger, B., Stephan, K. E., & Dolan, R. J. (2006).
Anterolateral prefrontal cortex mediates the analgesic effect of expected and perceived control
over pain. Journal of Neuroscience, 26(44), 1150111509.
Wilson, G. T. (1987). Cognitive processes in addiction. British Journal of Addiction, 82(4), 343
353.
White, R. W. (1959). Motivation reconsidered: the concept of competence. Psychological Review,

66(5), 297.
Yacubian, J., Glascher, J., Schroeder, K., Sommer, T., Braus, D. F., & Buchel, C. (2006).
Dissociable systems for gain- and loss-related value predictions and errors of prediction in the
human brain. Journal of Neuroscience, 26(37), 95309537.
Ybarra, O., Lee, D. S., & Gonzalez, R. (2012). Supportive social relationships attenuate the
appeal of choice. Psychological Science, 23(10), 11861192.
Zink, C. F., Pagnoni, G., Martin-Skurski, M. E., Chappelow, J. C., & Berns, G. S. (2004). Human
striatal responses to monetary reward depend on saliency. Neuron, 42(3), 509517.
Page 30 of 31
Page 31 of 31
Goals and the Sense of Agency

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
The Case of Goal Conflicts
Tali Kleiman
DOI:10.1093/acprof:oso/9780190267278.003.0007

Conflicts have been the center of interest for many researchers coming from utterly different
research fields. The basic question to ask when considering goal conflicts and agency in tandem
is whether a conflict, as compared to a smooth flow of intentions and their execution, changes
ones sense of agency. A fit between ones goals has traditionally been considered as
beneficiary, whereas a misfit is usually less so. From this perspective, conflicts in general and
goal conflicts specifically should have unfavorable effects on agency as well. On the other hand,
the association of conflict and the resulting activation of control to guide goal-directed behavior
may provide a different take on the consequences of conflict for agency. This chapter outlines
research on goal conflicts from these two perspectives, includes consideration of the issue of
consciousness, and makes an attempt to explicitly consider how goal conflicts affect human
sense of agency.
Keywords: conflict, goals, intentionality, consciousness, agency
Introduction
Conflicts are inherent to human experience, and they come in various shapes and forms when
two (or more) thoughts, intentions, or actions are not compatible with one another. Conflicts
have been the center of interest for many researchers from utterly different fields, attempting to
understand how conflicts relate to human functioning in the world: physically, cognitively,
emotionally, and motivationally. The basic question to ask when considering goal conflicts and
agency in tandem is whether a conflict, as compared to a smooth flow of intentions and their
Page 1 of 19
execution, changes ones sense of agency. Specifically, what is there about conflicts that may
produce such changes, and if changes indeed occur, do they weaken or strengthen the sense of
agency? A fit between ones body and mind (Borod, 2000; Huang & Galinsky, 2010; Tamir,
Robinson, Clore, Martin, & Whitaker, 2004), attitudes toward an object (Petty, Tormala, Brinol,
& Jarvis, 2006; Priester & Petty, 1996), a chronic and current state (Higgins, 2005), and most
pertinent to the current chapter, between ones goals (Emmons, 1996; Emmons & King, 1988)
has traditionally been considered as beneficiary, whereas a misfit is usually less so. From this
perspective, conflicts in general and goal conflicts in particular should have unfavorable effects
on agency as well. On the other hand, the association between the detection of conflict and the
subsequent (p.178) activation of control to guide goal-directed behavior may provide a
different take on the consequences of conflict for agency. In this respect, since intention and
conscious experience seem an inherent component of conflicts, goals, and agency, the
spontaneous, unintended, or non-conscious activation of conflict, its resolution, and
subsequently goal-directed behavior is of special interest to the relations between conflicts and
agency. This chapter outlines research on goal conflicts from the two perspectives mentioned
above, includes consideration of the issue of awareness, and makes an attempt to consider how
goal conflicts affect the human sense of agency.
What Is a Goal Conflict?

The simple answer to this question is that goal conflicts are situations in which desired end
states (goals) are incompatible in relation to one another (conflict). However, since both goals
and conflicts can be described on different levels of abstraction (Trope & Liberman, 2010;
Vallacher & Wegner, 1989) and take different forms, the answer may be somewhat more
complicated. For example, not spilling wine on the new carpet and bringing world peace are
both desired goals. But does defining both as being goals mean they share an underlying
process, or that they have the same consequences? Does the conflict activated when performing
incompatible trials of a Stroop task in a lab setting parallel a conflict many of us encounter when
indulgence and dietary concerns are pitted one against the other, or when considering the
consequences of lashing out in anger at our boss?
Moreover, the nature of the conflict between goals can vary as well. In self-control conflicts, a
higher-order goal conflicts with a lower-order goal, thus creating an asymmetric conflict
(Fishbach & Shah, 2006; Fujita & Carnevale, 2012; Hassin, Ochsner, & Trope, 2010; Myrseth &
Fishbach, 2009), for example, the conflict between the short-term goal of enjoying the chocolate
fudge brownie and the long-term, higher-order goal of ones health and physical appearance.
The conflict between ones career and family goals, on the other hand, may be portrayed as a
more symmetric conflict, when the right thing to do is much less obvious. Furthermore, goals
can conflict because they directly compete for limited resources (there are not enough hours in
a day to spend time in a museum and go to the movies, so prioritization is necessary); on the
other hand, goals may conflict irrespective of the amount of resources availablewe cannot be
both faithful partners and have casual sexno amount of resources can alleviate this conflict
(Kleiman & Hassin, 2011, 2013).
Goals, conflicts, and goal conflicts are at the heart of interest and scientific investigation in
many different literatures. These literatures define goal conflicts using different terms, examine
them from different levels of analysis, and (p.179) focus on different origins and consequences
Page 2 of 19
of goal conflicts. More often than not, these literatures do not communicate with one another, so
knowledge bases are rather infrequently shared, not to mention integrated. A taxonomy and
consolidation of these literatures and of goal conflicts is far beyond the reach of the current
chapter. However, throughout this chapter I will describe theories and findings, using examples
from different literatures that investigate goal conflict from different perspectives. I will make
some distinctions, discuss a few definitions, and ponder on some communality, only to facilitate
the comprehension of the ideas furthered in this chapter. I trust that readers will be able to
identify a goal conflict when they see one.
Detrimental Effects of Goal Conflicts

The detrimental consequences of conflicts in general and goal conflicts in particular are perhaps
the most intuitive and well-studied outcome of conflicts. A fit between ones body and mind is
considered to be important for normal functioning (e.g., Huang & Galinsky, 2010). A fit between
ones chronic disposition and situational requirements is beneficial for achieving ones goals
(e.g., Higgins, 2005). States of emotional ambivalence and cognitive dissonance are aversive and
counterproductive (Harmon-Jones, 2000). Moreover, decision-making can become a rather
unpleasant task when the alternatives to choose from conflict with one another, and trade-offs
need to be made. In such cases, higher rates of choice deferral are observed, as well as a
tendency to choose compromise optionsboth of which lead us away from our goals and hinder
progress (Anderson, 2003; Dhar, 1997; Iyengar & Lepper, 2000; Luce, 1998; Tversky & Shafir,
1992). Finally, conflicts between everyday goals have been found to be associated with both
short-term and long-term physical and emotional ill being (Emmons & King, 1988; Kehr, 2003).
Thus, throughout an array of behavioral and experiential phenomena, it is generally the case
that a fit, rather than a misfit, between goals (and with them cognitions, emotions, and
behaviors) is more conducive for achieving ones goals and for ones well-being.
In line with these findings, the emotional and physical ill being associated with conflicts, the
tendency to compromise or defer choices when making decisions, or succumbing to temptations
in self-control conflicts, may be detrimental to the sense of agency as well. Conflicts may
decrease fluency of action selection, which has been shown to be positively related to judgments
of agency (Chambon & Haggard, 2012; but see also discussion below on how control processes
activated by conflict may actually increase sense of agency). On an experiential level, actual and
perceived inaction and the feelings of lack of control that result from intractable conflicts may
reduce ones sense of (p.180) agency for thoughts and actions related to the conflict at hand.
Finally, in more severe cases, when perpetual conflicts lead to depression and anxiety disorders
(Emmons & King, 1988; King & Emmons, 1990), their effects may generalize to other domains
as well, leading to a wide-ranging decrease in the sense of agency.
Goal Conflicts Facilitate Subsequent Goal-Directed Behavior

Some theoretical perspectives on intra-individual goal conflicts link it to human growth and
development. These perspectives have emphasized that alongside the potentially detrimental
effects of goal conflicts to physical and emotional well-being, conflicts are also an important
driving force for self enhancement (Brim & Kagan, 1980; Turiel, 1974)in line with the common
maxim what does not kill you will make you stronger, if you wish. However, while the
implications of the detrimental consequences of conflict to agency are rather clear (suggesting
that conflict may result in a diminished sense of agency), the positive consequences that conflict
Page 3 of 19
may have on agency are somewhat less straightforward. One way to think of conflicts as possibly
contributing to enhancement of the sense of agency is to look at conflicts through the lens of
their role in facilitating goal-directed behavior.
When ones goals are clearly prioritized, namely, when one knows what goal one wants to
achieve, conflict can actually be helpful in achieving the desired goal. From the perspective of
both cognitive neuroscience and social cognition research, goal conflicts are associated with the
activation of cognitive control. Research within these fields describes the processes underlying
goal conflicts and the subsequent effects of experiencing a conflict on decision-making
strategies and the facilitation of goal-directed behavior (Botvinick et al., 2001; Folkman &
Lazarus, 1988; Janis & Mann, 1977). Furthermore, this research suggests that goal conflicts
promote subsequent goal-directed behavior spontaneously, without necessarily involving a
conscious reminder of the intention to pursue a certain goal. Put differently, as the system
adjusts to conflict or when decisions are made in a context of conflict, an increase in goal-
directed behavior may naturally follow, regardless of ones conscious intention.
On-line Conflict Activated Control Adjustment

A compelling account of how conflict may affect subsequent goal-directed behavior comes from
Conflict Monitoring Theory (Botvinick et al., 2001). The theory depicts the process by which the
cognitive system adjusts to conflict (p.181) on-line, as events unfold in the moment. It
proposes a framework explicating how the need for control (to resolve a conflict) is first
signaled, followed by activation of control that subsequently facilitates goal-directed behavior.
Specifically, the model suggests that there are two mechanisms dedicated to the detection and
resolution of conflict. The conflict-monitoring function (in which the anterior cingulate cortex is
suggested to be central; Kerns et al., 2004) is thought to be effortless. This mechanism is
responsible for detecting the occurrence of conflict and signaling that action must be taken to
direct behavior toward meeting the desired goal. At this point, control (implemented by the
DLPFC; MacDonald, Cohen, Stenger, & Carter, 2000) kicks in, in an effortful process aimed at
ensuring goal attainment. This is achieved by directing attention toward goal-relevant stimuli
and away from distractors that might hinder goal attainment. Furthermore, and most pertinent
to the model and the current chapter, it suggests that the activated control and the goal-
directed biasing it carries with it may then transfer to the next conflict, if encountered (see
further in this chapter for discussion of the domain generality of this process). For a simple
illustration, take the Stroop task (Stroop, 1935; see MacLeod, 1991 for a thorough review) as an
example. In this task, participants goal (i.e., the assigned task goal) is to name the color a word
is written in. Trials in this task are set up such that on compatible, no conflict, trials the meaning
of the word corresponds to the ink color it is written in (e.g., the word red is written in red
ink). On incompatible, conflict trials the meaning of the word is incompatible with the ink color
it is written in (e.g., the word red is written in blue ink). Decades of research produced robust
findings from the Stroop and other Stimulus Response Compatibility tasks (see Kornblum,
Hasbroucq, & Osman, 1990, for a taxonomy) showing a congruency effect such that response
times are slower (and error rates are higher) on incompatible compared to compatible trials.
Various accounts of the type of conflict created in the Stroop and other tasks exist, but it is
generally agreed upon that cognitive control needs to be activated to assure goal attainment on
incompatible, conflict trials (but not on compatible, no conflict trials). Interestingly, it has been
found that the magnitude of the congruency effect depends on the immediate previous
Page 4 of 19
experience. When the previous experience is that of conflict (i.e., on incompatible trials), the
magnitude of the congruency effect is substantially smaller compared to a previous experience
of non-conflict. This finding, first reported by Gratton and colleagues (Gratton, Coles, &
Donchin, 1992), has since been replicated using various Stimulus Response Compatibility tasks
(e.g., Kerns et al., 2004; Sturmer, Leuthold, Soetens, Schroter, & Sommer, 2002). Taken
together, then, there is a theoretical account and empirical support for the idea that the
system can adjust to conflict online, facilitating goal-directed behavior in a rapid succession.1
(p.182) The generality of the conflict-monitoring and adaptation process has been of great
interest in research within the conflict-monitoring theoretical framework. Specifically,
researchers were interested in the question of whether adjustment to conflict can happen only
within a domain (e.g., a specific task), or whether the adaptation domain is general, namely,
once conflict is detected and control activated it can be applied to any other situation requiring
control. Utilizing various Stimulus Response Compatibility (SRC) tasks, the general conclusion
seemed to be that conflict-activated control adjustments are domain specific (i.e., that the
detection of conflict and adjustment of control cannot generalize across different types of
conflict; Egner, 2008; Funes, Lupianez, & Humphrreys, 2010a; 2010b; Hazeltine et al., 2011;
Kiesel, Kunde, & Hoffmann, 2006; Notebaert & Verguts, 2008; but see Freitas, Bahar, Yang, &
Banai, 2007; Kunde & Wuhr, 2006, for other findings). The issue of control adjustments across
domains is a highly interesting and important one, not only for basic cognitive processes, which
are presumably measured by the SRC tasks. Control is needed in numerous life domains that go
far beyond specific task goals assigned to us in the lab. To name just a few examples, we need
control to refrain from eating the brownie and instead to reach for the apple (Fishbach & Shah,
2006; Fujita & Carnevale, 2012; Myrseth & Fishbach, 2009), we need control to contain our
emotions (Ochsner & Gross, 2005, 2008), and we need control for not allowing our automatic
stereotypical associations to affect our overt behavior (Amodio et al., 2004; Amodio, Devine, &
Harmon-Jones, 2008; Payne, 2001, 2005; Sherman et al., 2008). One of the key differences
between classic SRC tasks performed in a lab and daily conflicts encountered outside the lab
rests on the personal significance of the goals one is trying to achieve. Whereas lab tasks are
composed of cold stimuli (e.g., letters) and the goals are assigned by the experimenter,
everyday conflicts are hot, as they carry greater consequences to our lives (Locke & Braver,
2010; Metcalfe & Mischell, 1999).
The question of whether personal significance may engage the system in quantitatively or
qualitatively different ways and make control adjustments across domains more likely was
recently examined (Kleiman, Hassin, & Trope, 2014). Participants performed alternating single
trials of two tasks. The first was a classic cold Flanker task (Eriksen & Eriksen, 1974) in which
response to the middle letter in an array is required. On compatible, no-conflict trials, all of the
letters in the array require the same response (e.g., SSS); on incompatible, conflict trials, the
target letter and the distractors surrounding it are mapped to competing task responses (e.g.,
SHS). The second hot task was aimed at assessing implicit bias (having to do with either
gender or race stereotypes). On compatible trials, the stimuli conveyed a common stereotype
(e.g., that women are fragile); on incompatible trials, the stimuli conveyed a counter-stereotype
(e.g., that men are fragile). Findings showed that following compatible Flanker (p.183) trials,
automatic stereotypical associations affected overt judgments (e.g., that women are weak while
Page 5 of 19
men are strong). However, following Flanker incompatible trials, automatic associations ceased
to affect overt judgments. These findings suggest that conflict-activated control (as was
activated by the Flanker task) can be adjusted across domains in an on-line setting, if it involves
personal relevance in the form of important goals that one is eager to achieve (in this case, not
being or appearing biased toward minority groups).
A Conflict Mindset
The effect of experienced conflict on subsequent goal-directed behavior is not limited to trial-to-
trial rapid modulations of the level of control. Goal conflict can also change the context in which
we reason and process information by creating a conflict mindset. A mindset is a way of
processing information that may originate from one source with a specific content, but then may
be applied to information processing in a different context (Forster, Liberman, & Kuschel, 2008;
Gollwitzer, 1990; Sassenberg & Moskowitz, 2005). Simply put, the process transfers across
content domains. We (Kleiman & Hassin, 2013) have recently proposed that (non-conscious; see
discussion of this point further in this chapter) goal conflicts create a mindset with specific
characteristics; this conflict mindset, in turn, affects reasoning in domains unrelated to the
conflict itself. The specific characteristics of a conflict mindset stem from the very basic nature
of conflicts. Imagine a situation in which one does not have any health or dieting concerns and
just wants to indulge. Eating the chocolate fudge brownie is in line with this goal. However, if
one does have health or dieting concerns, the decision of whether to eat or not to eat the
brownie is likely to be preceded by a personal variation on the following contemplation: On the
one hand, I would really like to indulge, but on the other hand, I have my weight and health to
think of. Thus, we suggested that inherent to any conflict, irrespective of the specific content, is
a process of considering multiple perspectives rather than zooming in on a sole stance (see the
concluding section of this chapter for a discussion on how this idea complements the goal
directedness idea described earlier). To examine the proposed conflict mindset, we turned to
one of the most robust biases of human thoughtthe confirmation bias (Klayman & Ha, 1987;
Koehler, 1993; Koriat, Lichtenstein, & Fischhoff, 1980; Snyder & Swann, 1978; Wason, 1960).
Judgments and decisions are often distorted due to the consideration of only one perspective,
opinion, or piece of information that concurs with ones current beliefs and attitudes. Thus,
ones judgments, decisions, and behaviors are often one sided, taking into account confirming
information while ignoring or (p.184) disregarding disconfirming or conflicting information. As
noted above, inherent to conflicts is the consideration of conflicting sides, and thus being in a
conflict mindset should significantly attenuate confirmation tendencies. In a series of studies we
found that activating a conflict mindset (by simultaneously priming two conflicting goals) indeed
subsequently led to the consideration of a broader array of informationspecifically, pieces of
information that conflicted with one another. The effect of the activated conflict mindset on
information processing was not accompanied by a conscious phenomenology of being in a
conflict state of mind (see more about explicit phenomenology further in this chapter).
This research suggested that a subtly induced goal conflict that does not reach conscious
awareness may affect information processing in unrelated decisions. An important question still
remainedwhether a conflict mindset can promote goal-directed behavior in a subsequent,
unrelated conflictual decision. When making everyday decisions, we often face several
alternatives with different features, and making a choice is often a trade-off between these
features. Our goal may be to reach a decision and choose, but this might turn out to be rather
Page 6 of 19
difficult, especially when the trade-offs represent conflicting motivations. In such cases we tend
to defer the choice, or choose a compromise option that does not involve commitment
(Anderson, 2003; Dhar, 1997; Iyengar & Lepper, 2000; Luce, 1998; Tversky & Shafir, 1992).
Echoing the ideas presented above, together with theories suggesting that experiencing
decisional or emotional conflict does not necessarily lead to an avoidant response but can rather
trigger proactive confrontation and attempts at systematic resolution of the conflict (Folkman &
Lazarus, 1988; Janis & Mann, 1977), recent research proposed that being in a conflict mindset
may facilitate conflict resolution in a decision-making setting (Savary, Kleiman, Hassin, & Dahr,
2015). Using various choice sets that comprised conflicting alternatives, results showed that
those participants for whom non-conscious goal conflict was evoked had lower deferral rates
and were more willing to commit to a decision rather than choosing the compromise option. This
conflict-induced goal-directed behavior was not accompanied by conflict-mindset conscious
phenomenology.
The Positive Effects of Conflict: Spontaneous, Unintentional Facilitation of Goal-Directed

Behavior
Utilizing different theoretical approaches, different paradigms, different timelines of events
occurring, and looking at multiple effects, the research presented in the preceding section
paints a picture in which, across these various differences, goal conflicts may facilitate
subsequent goal-directed behavior. (p.185) Moreover, it might be the case that, at least to
some extent, conflict-induced goal-directed behavior is spontaneous and unintentional. Namely,
when one is in a conflict mindset or had a preceding immediate experience of conflict,
subsequent goal-directed behavior occurs without setting anew the intention to achieve the goal.
Interestingly, personal significance may play an important role in facilitating this process. Goals
assigned by the experimenter for a specific task (e.g., naming the ink color in a Stroop task) are
fundamentally different from personally significant goals, at least to the extent that motivation
to achieve the latter should be higher. Similarly, task-created conflicts (such as the ones in
various SRC tasks) are fundamentally different from personal conflicts (such as that between the
goal to enjoy a tasty snack and that of maintaining a healthy diet). The finding that the
involvement of a more motivational conflict may alter the application of basic conflict processes
should not thus be surprising (Locke & Braver, 2010; Kleiman et al., 2014).
In summary, goal conflicts are associated with enhanced goal-directed behavior that is
spontaneous and not necessarily intentional, and that may be modulated by personal
significance (motivation). All of these factors may play an important role in the sense of ones
agency, a role that is yet to be explored.
Awareness
The issue of awareness is especially interesting in the case of agency and conflicts, because both
are traditionally viewed as a very conscious experience (but see Eitam, Kennedy, & Higgins,
2013; Karsh & Eitam, Chapter 12 of this volume; Moore, Middleton, Haggard, & Fletcher, 2012).
Can one be in a conflict without consciously experiencing it? Theoretical accounts of non-
conscious goal pursuit and recent evidence pertaining specifically to goal conflicts suggest that
this might be indeed the case.
Page 7 of 19
Non-conscious Goal Pursuit

Theories of goal pursuit have traditionally emphasized the role of conscious thought and
intention (e.g., Ajzen, 1991; Bandura, 1986; Deci & Ryan, 1985; Locke & Latham, 1990) in
setting and achieving ones goals. The intuition is clear and compelling: How can one plan, set in
motion, and achieve a desired end state without considerable conscious effort?
Some 25 years ago, this view of goal pursuit began changing as models of non-conscious goal
pursuit have been proposed (Bargh, 1990; Kruglanski, 1996). These models posit that goals are
mentally represented in memory within complex cognitive networks in which higher-order goals
are connected to lower-order goals, means for their attainment, and alternative goals. These (p.
186) networks are shaped by ones history and experience with pursuing the goals. Any
component in the network (e.g., a context in which the goal had been habitually pursued or
means for its attainment) can activate the rest of the network via activation spread. Thus, for
example, when one is exposed to the context in which the goal is habitually pursued, the goal
itself might be activated, leading to the activation of means for its attainment, which together
affect behavior and lead to goal pursuit.
Empirical investigations of these models often used the separate experiments paradigm
(Bargh, Gollwitzer, Lee-Chai, Barndollar, & Troetschel, 2001). In the first phase of a typical
experiment, a goal is primed by exposing participants (either consciously or not) to goal-related
stimuli (e.g., words like achievement and competition to prime an achievement goal). In the
second, allegedly unrelated phase, goal pursuit as manifested in priming-related changes is
examined. Participants are then asked to report whether they had a conscious experience of
pursuing the primed goal. Experiments of this sort have repeatedly demonstrated that the subtle
activation of various components in goal networks can lead to the non-conscious pursuit of
various goals. To take just a few examples, primed goals included impression formation and
memorization (Chartrand & Bargh, 1996), achievement (Bargh et al., 2001; Eitam, Hassin, &
Schul, 2008; Hassin, Bargh, & Zimerman, 2009), competition and cooperation (Bargh et al.,
2001), dieting (Fishbach, Friedman, & Kruglanski, 2003), specific task goals (Shah &
Kruglanski, 2002; 2003), sex and money (Aarts, Gollwitzer, & Hassin, 2004), solving puzzles
(Aarts, Custers, & Veltkamp, 2008; Custers & Aarts, 2005), interpersonal goals (Fitzsimons &
Bargh, 2003), attachment goals (Gillath et al., 2006), and egalitarian goals (Moskowitz,
Gollwitzer, Wasel, & Schaal, 1999; for recent reviews, see Custers & Aarts, 2010; Fishbach &
Ferguson, 2007). Taken together, these data strongly support the idea that the subtle activation
of components in goal networks can lead to goal pursuit in the absence of conscious intention
and thought (see Dijksterhuis, Chartrand, & Aarts, 2007).
Non-conscious Goal Conflicts

Goals are rarely pursued in a vacuum free of other goals, and thus if goals may be activated non-
consciously, simultaneously activated non-conscious goals may at times conflict with one
another. Even if one accepts that goals can be activated and pursued non-consciously, the notion
of non-conscious goal conflicts may pose yet another challenge. The modal view in the cognitive
sciences holds that managing conflicts in general and goal conflicts in particular is an effortful
process that requires conscious attention. This conception is only rarely acknowledged explicitly
(e.g., Norman & Shallice, 1986), but it is (p.187) often tacitconflicts are associated with
Page 8 of 19
cognitive control and the overcoming of pre-potent responses, both of which are assumed to be
predominantly conscious (Botvinick, Cohen, & Carter, 2004; Smith & Jonides, 1999). This
conception rests on two premises. The first is that one cannot really be in conflict if one does
not feel conflicted. In other words, conscious experience is inherent to conflicts. The second
premise is that non-conscious, automatic processes depend for their successful operation on pre-
existing habits, routines, or schema that smoothly channel behavior in predetermined routes (for
a review, see Wegner & Bargh, 1998). Inherent to conflicts, however, is the idea that schemas
fail to seamlessly channel behavior, thereby creating the need for a conscious intervention.
Recently, we (Kleiman & Hassin, 2011) have claimed that goal conflicts can and do occur non-
consciously. The can part of the claim is supported by two ideas: first, that goal conflicts are
pervasive in our lives on various levels and in various forms (Emmons & King, 1988; Emmons,
King, & Sheldon, 1993; Fishbach, Zhang, & Trope, 2010; Lee, Locke, & Latham, 1989); and
second, that human conscious resources are rather limited (e.g., Kahneman, 1973). Considered
together, these ideas suggest that conscious resources simply cannot handle all conflicts
consciously.2 We (Kleiman & Hassin, 2011) supported the do? part of the claim by a series of
experiments showing that goal conflicts can occur outside conscious awareness. If conflict
cannot be detected by explicit accounts of it, how would it be detected? We proposed a series of
conflict markers (which we do not deem as either necessary or exhaustive) that are likely to
arise in a goal-conflict context. The first marker is decision duration. This intuitive marker rests
on the basic assumption that decisions under conflict take longer. From the Stroop effect
(Stroop, 1935), to condom use (Abraham & Sheeran, 2003), to moral decisions (Greene,
Nystrom, Engell, Darley, & Cohen, 2004), negotiating between conflicting goals, plans, or
behaviors takes time. The second marker we proposed was behavioral variance. Goal conflicts
are created when there are multiple goals that a person finds attractive, and with them come
multiple appealing behaviors. Take donating for charity as an example. If one only cares about
accumulating wealth, then one will never donate. If one also has the conflicting goal of doing
good deeds with ones money, however, then ones behavior is likely to show more variance:
sometimes one will be tempted to donate a lot and in other times less. Hence, in conflict
situations in which a decision has to be made over and over again, behavioral variance should be
larger than in situations that are not conflictual. Physiological arousal was the third conflict
marker proposed. Decisions made in conflictual situations are characterized by difficulty and
unease, inconsistent behavioral intentions, and inconsistent affective tendencies. Decades of
research using various measures of arousal have (p.188) shown that these types of situations
are usually accompanied by higher levels of arousal (see, e.g., Allen & Crowell, 1989; Blascovich
et al., 1993; Kahneman 1973; Kelsey, 1991; Obrist, 1981). Finally, we proposed that
susceptibility to subtle, irrelevant cues may be another marker of conflict. When goals are in
active conflict they often create close-call decisions, that is, decisions in which the alternatives
seem to have very similar utilities. Metaphorically, then, the decision scales are more or less
balanced. It is exactly in these cases where minor (and possibly irrelevant) cues in the
environment have the potential of tipping the scales. Consider donating again. The behavior of
the caller who asks for the donation is unlikely to affect our decision when we are determined
and know exactly what we want to do. When we are in a conflict, however, this behavior may
indeed make a difference. If the person is nice and warm, the likelihood of a big donation may
Page 9 of 19
increase; if he is very rude, it may decrease. Environmental information, then, even when it is
irrelevant to a decision, is more likely to affect behavior in conflictual situations.
In a series of studies we used a classic social dilemma paradigm (e.g., Fehr & Fischbacher,
2003) and examined how behavior and implicit conflict markers are altered as a function of the
non-conscious goal pursuit and conflict involved. The dominant goal in social dilemma
paradigms is a selfish, self-serving goal (Brewer & Kramer, 1986; Fehr & Fischbacher, 2003).
One wants to accumulate as much resources as possible for oneself, and cooperation, or the
benefit of the community, is very often arather minor consideration. Hence, we reasoned that
non-consciously activating a cooperation goal before engagement in a social dilemma task may
create a non-conscious goal conflict between the dominant self-serving goal and the primed
cooperation goal. We sought to detect conflict implicitly by the conflict markers outlined earlier,
while at the same time assuring that conflict did not result in conscious phenomenology. Indeed,
we showed that participants in the conflict condition had longer decision durations, larger
behavioral variance, higher physiological arousal, and greater susceptibility to irrelevant
environmental cues. When probed for conflict experience, both at the end of the experiment and
on-line on a trial-by-trial basis, control and conflict participants did not differ. Hence, while
conflict markers were present, phenomenology of conflict was not.
Awareness of What?
When talking about non-conscious processes in general and non-conscious goals and conflicts in
particular, it is important to dwell on the question of what exactly participants are unaware of.
In most of the studies in which goals are non-consciously activated, priming techniques are used
in which words related (p.189) to the desired goal are primed (e.g., cooperate and share to
activate a cooperation goal). In cases in which the words are primed subliminally, it is safe to
assume that participants are unaware of them. However, often the priming is done using
supraliminal presentation of goal-related words. In such cases, participants are obviously aware
of the words, and consciously process them; so in what sense is the goal activation and pursuit
non-conscious? Research within this domain usually uses a dissociation between the behavioral
effects of goal pursuit and the explicit phenomenology of this pursuit. Thus, for example,
participants complete a word search puzzle in which words that are related to the goal of
cooperation are embedded (e.g., Bargh et al., 2001). Participants then move to perform an
ostensibly unrelated task in which their cooperation-related behavior is measured, for example,
a social dilemma task. After they complete the task, participants are explicitly probed for their
motivation to pursue the goal. A dissociation is created such that behavioral changes that are
the result of the priming manipulation are not accompanied by reports of explicit motivation to
pursue the goal. Research on non-conscious goal conflicts adopted a similar paradigm, probing
for participants explicit feelings of conflict (Kleiman & Hassin, 2011; 2013; Savary et al., 2015).
Importantly, this is the case for probing conflict-related explicit experience both at the end of
the behavioral task, and after each and every trial (Kleiman & Hassin, 2011)rendering the
explanation that by the time they reach the end of the task participants forget how conflicted
they were rather implausible.
The research on non-conscious goal pursuit and non-conscious goal conflicts thus suggests that
actions and behaviors are presumably performed without a consciously experienced intention to
perform them. It is interesting to note in this respect that not being able to ascribe intention to
Page 10 of 19
the action may lead people to misattribute their behavior to some other internal state to the
extent of confabulating self-knowledge (Bar-Anan, Wilson, & Hassin, 2010).
Conclusion
Goal conflicts are usually viewed as an aversive state with negative effects on human
functioning and well-being. From this perspective, goal conflicts should also pose a detrimental
effect on ones sense of agency, as conflicts may lead to indecision and inaction. Perhaps the
more interesting and less intuitive possibility is that goal conflicts may actually facilitate the
sense of agency because they proactively facilitate goal-directed behavior. To the extent that
goal-directed behavior is conducive to the sense of agency, goal conflicts may have positive
effects on the sense of agency.
One apparent key difference that may be essential in determining whether goal conflicts lead to
a decrease or an increase in the sense of agency is conflict (p.190) resolution. When conflict
cannot be resolved, detrimental effects for well-being, as well as for the sense of agency, may
follow. When conflict-activated cognitive control facilitates goal-directed behavior, the sense of
agency may be facilitated as well. Relatedly, it is interesting to note that conflicts inherently
involve choice, and the opportunity to choose in itself might be rewarding (Leotti & Delgado,
2011), which in turn might have implications for ones sense of agency.
Another interesting distinction is between the activation of control and the feeling of control.
Encountering a conflict leads to the activation of control. Does this activation lead to a conscious
feeling of control? Is a conscious feeling of control necessary to promote the sense of agency?
The relations between conflict-activated control, goal-directed behavior, and the role of
conscious experience, and how they all combine to affect ones sense of agency, are thus
interesting questions yet to be explored.
Notes
References
Aarts, H., Custers, R., & Veltkamp, M. (2008). Goal priming and the affective-motivational route
to nonconscious goal pursuit. Social Cognition, 26(5), 555577.
Aarts, H., Gollwitzer, P., & Hassin, R. R. (2004). Goal contagion: perceiving is for pursuing.
Abraham, C., & Sheeran, P. (2003). Implications of goal theories for the theories of reasoned
action and planned behaviour. Current Psychology, 22(3), 264.
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision
Processes, 50(2), 179211.
Allen, M. T., & Crowell, M. D. (1989). Patterns of autonomic response during laboratory
stressors. Psychophysiology, 26, 603614.
Page 11 of 19
Amodio, D. M., Devine, P. G., & Harmon-Jones, E. (2008). Individual differences in the regulation
of intergroup bias: the role of conflict monitoring and neural signals for control. Journal of
Personality and Social Psychology, 94, 6074.
Amodio, D. M., Harmon-Jones, E., Devine, P. G., Curtin, J. J., Hartley, S. L., & Covert, A. E.
(2004). Neural signals for the detection of unintentional race bias. Psychological Science, 15,
8893.
Anderson, C. (2003). The psychology of doing nothing: forms of decision avoidance result from
reason and emotion. Psychological Bulletin, 129(1), 139166.
Bandura, A. (1986). Social foundations of thought and action: a social cognitive theory. Upper
Saddle River, NJ: Prentice-Hall.
Bar Anan, Y., Wilson, T., and Hassin, R. R. (2010). Inaccurate self-knowledge formation as a
result of automatic behavior. Journal of Experimental Social Psychology, 46, 884894.
Bargh, J. A. (1990). Auto-motives: preconscious determinants of social interaction. In E. T.

Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: foundations of social
behavior (Vol. 2, pp. 93130). New York: Guilford Press.
Bargh, J. A., Gollwitzer, P., Lee-Chai, A., Barndollar, K., & Troetschel, R. (2001). The automated
will: nonconscious activation and pursuit of behavioral goals. Journal of Personality and Social
Psychology, 81(6), 10141027.
Blascovich, J., Ernst, J. M., Tomaka, J., Kelsey, R. M., Salomon, K. L., & Fazio, R. H. (1993).
Attitude accessibility as a moderator of autonomic reactivity during decision making. Journal of
Borod, J. C. (2000). The neuropsychology of emotion. New York: Oxford University Press.
Botvinick, M., Braver, T., Barch, D. Carter, C., & Cohen, J. (2001). Conflict monitoring and
cognitive control. Psychological Review, 108, 624652.
Botvinick, M. M., Cohen, J. D., & Carter, C. S. (2004). Conflict monitoring and anterior cingulate
cortex: an update. Trends in Cognitive Sciences, 8(12), 539.
Brewer, M. B., & Kramer, R. M. (1986). Cohice behavior in social dilemmas: effects of social
identity, group size, and decision framing. Journal of Personality and Social Psychology, 50, 543
549.
Brim, O., & Kagan, J. (1980). Constancy and change: A view of the issues. In O. Brim & J. Kagan
(Eds.), Constancy and change in human development (pp. 126). Cambridge, MA: Harvard
University Press.
Chambon, V., & Haggard, P. (2012). Sense of control depends on fluency of action selection, not
motor performance. Cognition, 125, 441451.
Page 12 of 19
Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of impression formation and
memorization goals: nonconscious goal priming reproduces effects of explicit task instructions.
Custers, R., & Aarts, H. (2005). Positive affect as implicit motivator: On the nonconscious
operation of behavioral goals. Journal of Personality and Social Psychology, 89, 129142.
Custers, R., & Aarts, H. (2010). The unconscious will: how the pursuit of goals operates outside
of conscious awareness. Science, 329, 4750.
Deci, E. L., & Ryan, R. M. (1985). The general causality orientations scale: self-determination in
personality. Journal of Research in Personality, 19(2), 109134.
Dhar, R. (1997). Consumer preference for a no-choice option. Journal of Consumer Research,
24(2), 215231.
Dijksterhuis, A., Chartrand, T. L., & Aarts, H. (2007). Effects of priming and perception on social
behavior and goal pursuit. In J. A. Bargh (Ed.), Social psychology and the unconscious: the
automaticity of higher mental processes (pp. 51132). Philadelphia: Psychology Press.
Egner, T. (2008). Multiple conflict-driven control mechanisms in the human brain. Trends in
Cognitive Sciences, 12, 374380.
Eitam, B., Hassin, R. R., & Schul, Y. (2008). Nonconscious goal pursuit in novel environments:
the case of implicit learning. Psychological Science, 19, 261267.
Research, 229, 475484.
Emmons, R. A. (1996). Striving and feeling: personal goals and subjective well-being. In J. Bargh
& P. Gollwitzer (Eds.), The psychology of action: linking motivation and cognition to behavior
(pp. 314337). New York: Guilford.
Emmons, R. A., & King, L. A. (1988). Conflict among personal strivings: immediate and long-
term implications for psychological and physical well-being. Journal of Personality and Social
Psychology, 54(6), 10401048.
Emmons, R. A., King, L. A., & Sheldon, K. (1993). Goal conflict and the self-regulation of action.
In D. M. Wegner & J. W. Pennebaker (Eds.), Handbook of mental control (pp. 528551). Upper
Saddle River, NJ: Prentice-Hallc.
Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of a target
letter in a nonsearch task. Perception & Psychophysics, 16, 143149.
Fehr, E., & Fischbacher, U. (2003). The nature of human altruism. Nature, 425(6960), 785791.
Page 13 of 19
Fishbach, A., & Ferguson, M. F. (2007). The goal construct in social psychology. In A. W.
Kruglanski and T. E. Higgins (Eds.), Social psychology: handbook of basic principles (pp. 490
515). New York: Guilford.
Fishbach, A., Friedman, R. S., & Kruglanski, A. W. (2003). Leading us not into temptation:
Momentary allurements elicit overriding goal activation. Journal of Personality and Social
Fishbach, A., & Shah, J. Y. (2006). Self control in action: implicit dispositions toward goals and
away from temptations. Journal of Personality and Social Psychology, 90, 820832.
Fishbach, A., Zhang, Y., & Trope, Y. (2010). Counteractive evaluation: asymmetric shifts in the
implicit value of conflicting motivations. Journal of Experimental Social Psychology, 46, 2938.
Fitzsimons, G., & Bargh, J. A. (2003). Thinking of you: nonconscious pursuit of interpersonal
goals associated with relationship partners. Journal of Personality and Social Psychology, 84,
148164.
Folkman, S., & Lazarus, R. (1988). Coping as a mediator of emotion. Journal of Personality and
Social Psychology, 54(3), 466475.
Frster, J., Liberman, N., & Kuschel, S. (2008). The effect of global versus local processing
styles on assimilation versus contrast in social judgment. Journal of Personality and Social
Freitas, A. L., Bahar, M., Yang, S., & Banai, R. (2007). Contextual adjustments in cognitive
control across tasks. Psychological Science, 18, 10401043.
Fujita, K., & Carnevale, J. J. (2012). Transcending temptation through abstraction: the role of
construal level in self-control. Current Directions in Psychological Science, 21, 248252.
Funes, M. J., Lupiez, J., & Humphreys, G. (2010a). Analyzing the generality of conflict
adaptation effects. Journal of Experimental Psychology: Human Perception and Performance, 36,
147161.
Funes, M. J., Lupiez, J., & Humphreys, G. (2010b). Sustained vs. transient cognitive control:
evidence of a behavioral dissociation. Cognition, 114, 338347.
Gillath, O., Mikulincer, M., Fitzsimons, G. M., Shaver, P. R., Schachner, D. A., & Bargh, J. A.
(2006). Automatic activation of attachment-related goals. Personality and Social Psychology
Bulletin, 32, 13751388.
Gollwitzer, P. M. (1990). Action phases and mind-sets. In E. T. Higgins & R. M. Sorrentino

(Eds.), The handbook of motivation and cognition: foundations of social behavior (Vol. 2, pp. 53
Gratton, G., Coles, M. G. H., & Donchin, E. (1992). Optimizing the use of information: strategic
control of activation of responses. Journal of Experimental Psychology: General, 121, 480506.
Page 14 of 19
Greene, J. D., Nystrom, L. E., Engell, A. D., Darley, J. M., & Cohen, J. D. (2004). The neural bases
of cognitive conflict and control in moral judgment. Neuron, 44, 389400.
Harmon-Jones, E. (2000). Cognitive dissonance and experienced negative affect: evidence that
dissonance increases experienced negative affect even in the absence of aversive consequences.
Personality and Social Psychology Bulletin, 26, 14901501.
Hassin, R. R., Bargh, J. A., & Zimerman, S. (2009). Automatic and flexible: the case of non-
conscious goal pursuit. Social Cognition, 27(1), 2036.
Hassin, R. R., Ochsner, K., & Trope, Y. (2010), Self control in society, mind, and brain. New
Hazeltine, E. Lightman, E. Schwarb, H., & Schumacher, E. H. (2011). The boundaries of

sequential modulations: evidence for set-level control. Journal of Experimental Psychology
Human Perception and Performance, 37, 18981914.
Higgins, E. T. (2005). Value from regulatory fit. Current Directions in Psychological Science, 14,
209213.
Huang, L., & Galinsky, A. D. (2011). Mindbody dissonance: conict between the senses expands
the minds horizons. Social Psychological and Personality Science, 2, 351359.
Iyengar, S., & Lepper, M. (2000). When choice is demotivating: can one desire too much of a
good thing? Journal of Personal and Social Psychology, 79(6), 9951006.
Janis, I., & Mann, L. (1977). Decision making: a psychological analysis of conflict, choice, and
commitment. New York: Free Press.
Kahneman, D. (1973). Attention and effort. Englewoods Cliff, NJ: Prentice-Hall.
Kehr, H. M. (2003). Goal conflicts, attainment of new goals, and well-being among managers.
Journal of Occupational Health Psychology, 8, 195208.
Kelsey, R. M. (1991). Electrodermal lability and myocardial reactivity to stress.

Psychophysiology, 28(6), 619631.
Kerns, J. G., Cohen, J. D., MacDonald, A. W., III, Cho, R. Y., Stenger, V. A., & Carter, C. S.
(2004). Anterior cingulate conflict monitoring and adjustments in control. Science, 303, 1023
1026.
Kiesel, A., Kunde, W., & Hoffmann, J. (2006). Evidence for task-specific resolution of response
conflict. Psychonomic Bulletin & Review, 13, 800806.
King, L. A., & Emmons, R. A. (1990). Ambivalence over expressing emotion: physical and
psychological correlates. Journal of Personality and Social Psychology, 58, 864877.
Page 15 of 19
Klayman, J., & Ha, Y. W. (1987). Confirmation, disconfirmation, and information in hypotheses
testing. Psychological Review, 94, 211228.
Kleiman, T., & Hassin, R. R. (2011). Non-conscious goal conflicts. Journal of Experimental Social
Kleiman, T., & Hassin, R. R. (2013). When conflicts are good: non-conscious goal conflicts
reduce confirmatory thinking. Journal of Personality and Social Psychology, 105, 374387.
Kleiman, T., Hassin, R. R., & Trope, Y. (2014). The control-freak mind: stereotypical biases are
eliminated following conflict activated cognitive control. Journal of Experimental Psychology:
General, 143, 498503.
Koehler, J. J. (1993). The influence of prior beliefs on scientific judgments of evidence quality.
Organizational Behavior and Human Decision Processes, 56, 2855.
Koriat, A., Lichtenstein, S., & Fischhoff, B. (1980). Reasons for confidence. Journal of
Experimental Psychology: Human Learning and Memory, 6, 107118.
Kornblum, S., Hasbroucq, T., & Osman, A. (1990). Dimensional overlap: cognitive basis for
stimulus-response compatibilitya model and taxonomy. Psychological Review, 97, 253270.
Kruglanski, A. W. (1996). Goals as knowledge structures. In P. Gollwitzer and J. A. Bargh (Eds.),

The psychology of action: linking cognition and motivation to behavior (pp. 599618). New York:
Guilford.
Kunde, W., & Whr, P. (2006). Sequential modulations of correspondence effects across spatial
dimensions and tasks. Memory & Cognition, 34, 356367.
Lee, T. W., Locke, E. A., & Latham, G. P. (1989). Goal setting theory and job performance. In L.
A. Pervin (Ed.), Goal concepts in personality and social psychology (pp. 291326). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Leotti, L. A., & Delgado, M. R. (2011). The inherent reward of choice. Psychological Science, 22,
13101318.
Locke, H. S., & Braver, T. S. (2010). Motivational influences on cognitive control: a cognitive
neuroscience perspective. In R. R. Hassin, K. N. Ochsner, and Y. Trope (Eds.), Self control in
society, mind, and brain. New York: Oxford University Press.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task performance. Upper
Saddle River, NJ: Prentice-Hall.
Luce, M. F. (1998). Choosing to avoid: coping with negatively emotion laden consumer
decisions. Journal of Consumer Research, 24, 409433.
Page 16 of 19
MacDonald, A. W., Cohen J. D., Stenger, V. A., & Carter, C. S. (2000). Dissociating the role of
dorsolateral prefrontal cortex and anterior cingulate cortex in cognitive control. Science, 288,
18351837.
MacLeod, C. M. (1991). Half a century of research on the Stroop effect: an integrative review.
Mayr, U., Awh, E., & Laurey, P. (2003). Conflict adaptation effects in the absence of executive
control. Nature Neuroscience, 6, 450452.
Metcalfe, J., & Mischel, W. (1999). A hot/cool system analysis of delay of gratification: dynamics
of willpower. Psychological Review, 106, 319.
Moore, J. W., Middleton, D., Haggard, P., & Fletcher, P. C. (2012). Exploring implicit and explicit
aspects of sense of agency. Consciousness and Cognition, 21, 17481753.
Morsella, E. (2005). The function of phenomenal states: supramodular interaction theory.

Moskowitz, G. B., Gollwitzer, P., Wasel, W., & Schaal, B. (1999). Preconscious control of
stereotype activation through chronic egalitarian goals. Journal of Personality and Social
Psychology, 77(1), 167184.
Myrseth, K. O. R., & Fishbach, A. (2009). Self-control: a function of knowing when and how to
exercise restraint. Current Directions in Psychological Science, 18, 247252.
Nieuwenhuis, S., Stins, J. F., Posthuma, D., Polderman, T. J. C., Boomsma, D. I., & de Geus, E. J.
(2006). Accounting for sequential trial effects in the flanker task: conflict adaptation or
associative priming? Memory & Cognition, 34, 12601272.
Norman, D. A., & Shallice, T. (1986). Attention to action: willed and automatic control of
behavior. In J. R. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and self
regulation: advances in research and theory (Vol. 4, pp. 118). New York: Plenum Press.
Notebaert, W., & Verguts, T. (2008). Cognitive control acts locally. Cognition, 106, 10711080.
Ochsner, K. N., & Gross, J. J. (2005). The cognitive control of emotion. Trends in Cognitive
Sciences, 9(5), 242249.
Ochsner, K. N., & Gross, J. J. (2008). Cognitive emotion regulation: insights from social cognitive
and affective neuroscience. Currents Directions in Psychological Science, 17(1), 153158.
Obrist, P. A. (1981). Cardiovascular psychophysiology: a perspective. New York: Plenum Press.
Payne, B. K. (2001). Prejudice and perception: the role of automatic and controlled processes in
misperceiving a weapon. Journal of Personality and Social Psychology, 81, 181192.
Page 17 of 19
Payne, B. K. (2005). Conceptualizing control in social cognition: how executive control

modulates the expression of automatic stereotyping. Journal of Personality and Social
Petty, R. E., Tormala, Z. L., Briol, P., & Jarvis, W. B. G. (2006). Implicit ambivalence from
attitude change: an exploration of the PAST model. Journal of Personality and Social Psychology,
90, 2141.
Priester, J. M., & Petty, R. E. (1996). The gradual threshold model of ambivalence: relating the
positive and negative bases of attitudes to subjective ambivalence. Journal of Personality and
Sassenberg, K., & Moskowitz, G. B. (2005). Dont stereotype, think different! Overcoming
automatic stereotype activation by mindset priming. Journal of Experimental Social Psychology,
41, 506514.
Savary, J., Kleiman, T. Hassin, R. R., & Dahr, R. (2015). Positive consequences of conflict on
decision making: when a conflict mindset facilitates choice. Journal of Experimental Psychology:
General, 144, 16.
Shah, J. Y., & Kruglanski, A. W. (2002). Priming against your will: how accessible alternatives
affect goal pursuit. Journal of Experimental Social Psychology, 38, 368383.
Shah, J. Y., & Kruglanski, A. W. (2003). When opportunity knocks: bottom-up priming of goals by
means and its effects on self-regulation. Journal of Personality and Social Psychology, 84, 1109
1122.
Sherman, J. W., Gawronski, B., Gonsalkorale, K., Hugenberg, K., Allen, T. J., & Groom, C. J.
(2008). The self-regulation of automatic associations and behavioral impulses. Psychological
Review, 115, 314335.
Smith, E. E., & Jonides, J. (1999). Storage and executive processes in the frontal lobes. Science,
283(5408), 16571661.
Snyder, M., & Swann, W. B. (1978). Hypotheses testing processes in social interaction. Journal
of Personality and Social Psychology, 36, 12021212.
Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental

Strmer, B., Leuthold, H., Soetens, E., Schrter, H., & Sommer, W. (2002). Control over
location-based response activation in the Simon task: behavioral and electrophysiological
evidence. Journal of Experimental Psychology: Human Perception and Performance, 28, 1345
1363.
Page 18 of 19
Tamir, M., Robinson, M. D., Clore, G. L., Martin, L. L., & Whitaker, D. (2004). Are we puppets on
a string? The contextual meaning of unconscious expressive cues. Personality and Social
Psychology Bulletin, 30, 237249.
Trope, Y., & Liberman, N. (2010). Construal-level theory of psychological distance. Psychological
Review, 117, 440463.
Turiel, E. (1974). Conflict and transition in adolescent moral development. Child Development,
45, 1429.
Tversky, A., & Shafir, E. (1992). Decision under conflict: an analysis of choice aversion.
Psychological Science, 6, 358361.
Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency: individual variation in
action identification. Journal of Personality and Social Psychology, 57, 660671.
Wason, P. C. (1960). On the failure to eliminate hypotheses in a conceptual task. Quarterly

Journal of Experimental Psychology, 12, 129140.
Wegner, D. M., & Bargh, J. A. (1998). Control and automaticity in social life. In D. Gilbert & S.
Fiske (Eds.), The handbook of social psychology (Vol. 1, pp. 446496). New York: McGraw -Hill.
Notes:
(1) It is important to note that there have been challenges to the Conflict Monitoring Model
suggesting that on-line, trial-to-trial adjustment may stem from other factors, such as response
and stimulus repetitions (Mayr, Awh, & Laurey, 2003; Nieuwenhuis et al., 2006).
(2) Which conflicts remain non-conscious and which spring to consciousness, and whether these
instances can be predicted from any specific characteristics of the conflicts (e.g., Morsella,
2005) are interesting questions in their own right.
Page 19 of 19
Inference Processes Underlying the Human Experience of Agency over
Operant Actions

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Inference Processes Underlying the Human Experience of Agency over Operant

Actions
Myrthel Dogge
Henk Aarts
DOI:10.1093/acprof:oso/9780190267278.003.0008

This chapter discusses to cognitive inferences of self-agency over operant actions and how these
inferences can draw upon unconscious (implicit) sources of information. The main processes
subsidizing the experience of self-agency are predictive motor processes based on the likelihood
that an action produces an effect, and inference processes based on the correspondence
between action outcomes and previously activated knowledge concerning these outcomes.
Recently it has been proposed that inferences processes may also produce agency experiences
when implicitly pre-activated (or primed) knowledge about action-outcomes matches with the
observation of the actual outcomes. The chapter presents studies on this unconscious authorship
ascription process by showing when and how the mere pre-activation of knowledge pertaining to
information concerning the agent, sensory effects, and socially relevant outcomes modulates
peoples feeling of agency. Recent insights into disruptions of agency inferences as well as the
neural basis of implicit and explicit inference processes are briefly discussed.
Keywords: agency inferences, goals, inference process, unconscious, action outcome
Introduction
Human beings have the capacity to experience themselves as the cause of their own behavior.
For instance, when one pushes a button on a vending machine and the intended beverage is
dispensed, one tends to readily and effortlessly ascribe this effect to oneself. This sense of
agency over operant actionthat is, when actions are followed by specific consequenceshas
initially been explained by comparator processes described in models of motor control
Page 1 of 16
Operant Actions
(Blakemore, Wolpert, & Frith, 2002; Frith, Blakemore, & Wolpert, 2000; Wolpert & Flanagan,
2001). According to these accounts, the volitional or goal-directed preparation of an action is
accompanied by the prediction of sensory action-outcomes based on efference copies generated
by the motor system. Because these internal motor predictions are proposed to be generally
very reliable, sensory outcomes are readily perceived as self-produced until this prediction no
longer corresponds with the actual outcomes following ones action (Wolpert, Ghahramani, &
Jordan, 1995).
In principle, the motor prediction model proposes that the establishment of agency experiences
depends on whether the motor system can clearly predict action outcomes. Accordingly, the
predictive ability of the motor system is decreased when moving involuntarily, when there is no
clear causal (p.200) relationship between an action and a following event, or when actions have
multiple causes and outcomes (Van der Weiden, Aarts, & Ruys, 2013; Wegner, 2002). However,
recent research suggests that the experience of agency can occur under circumstances in which
the motor prediction process cannot rely on a clear predictive signal or is overruled by beliefs
about causes of actions and outcomes (Aarts, Custers, & Wegner, 2005; Desantis, Roussel, &
Waszak, 2011; Dogge, Schaap, Custers, Wegner, & Aarts, 2012; Moore & Haggard, 2008;
Moore, Wegner, & Haggard, 2009; Wegner, Sparrow, & Winerman, 2004). This recent work thus
points to the existence of a second route to agency experiences. Specifically, it has been
proposed that agency experiences can also arise from the inferred correspondence between
actual action-effects and pre-activated knowledge of these outcomes (Moore & Haggard, 2008;
Wegner, 2002) or from other, more conceptual cues (e.g., general knowledge about causality;
Synofzik, Vosgerau & Newen, 2008).
The present chapter examines these cognitive inferences of self-agency over operant actions to
offer a better understanding of the human capacity to experience oneself as the cause of ones
own behavior. Specifically, we briefly discuss common psychological approaches to the study of
agency inferences that emphasize correspondence between goals and action consequences as an
agentive cue. Goals evoke specific control processes that deal with shielding, monitoring, and
feedback processing in the service of attaining the specific desired outcome (e.g., Carver &
Scheier, 1998)processes that focus peoples attention on the goal at hand and may be
especially important when outcomes mismatch ones goal. In addition to this explicit route to
self-agency, we discuss recent research supporting an implicit route to agency inferences, that
is, a route in which people ascribe outcomes to their actions when they correspond with
outcome information that is pre-activated by environmental (implicit) cues instead of (explicit)
goals. In principle, the priming of outcome information is proposed to merely activate the
representation of the outcome and lacks the control processes accompanying explicit goals;
hence, outcome priming might mainly cause people to increase experience agency when the
prime matches the observed outcome.
Whereas the implicit route to self-agency offers a possible account for why people experience
agency over operant actions that are triggered by environmental cues or that are selected and
executed without much conscious thought and attention, there are a few boundary conditions to
this implicit route that we will consider. Also, we briefly discuss recent research that has started
to explore differences between the implicit and explicit routes to inferences of agency. Finally,
Page 2 of 16
Operant Actions
we address some recent developments and directions for future research that may be fruitful to
further examine the role of inference processes in self-agency.
(p.201) Inferences of Agency

According to the inference account, the sense of agency results from cognitive sense-making
processes during and after observation of action-outcomes (Wegner, 2002). That is to say, upon
perceiving an event, people quickly and fluently determine whether or not it results from their
behavior. In social-psychological models of goal-directed behavior, self-agency is often seen as a
product of a comparison between an explicitly set goal or desired outcome and the actual
attainment of the outcome (e.g., Bandura, 1986; Carver & Scheier, 1998; Deci & Ryan, 1985). In
line with these models, the inference account proposes that self-agency emerges when the
perception of an outcome corresponds with the outcome that one consciously intends to attain
by performing an action, while a mismatch is generally ascribed to other causes.
However, it may be questioned whether inferences of self-agency can only originate from
intentional processes as, in daily life, one can also ascribe outcomes to oneself when no explicit
goal was formed. This experience might be explained by looking more closely at the mechanism
underlying inferences of agency. Specifically, agency is inferred whenever there is
correspondence between pre-activated representations of outcomes and actual outcomes
following from actions, which suggests that other factors that can pre-activate outcome-
representations, such as environmental primes, can instigate this process as well. Indeed,
according to the theory of mental causation (Wegner, 2002, 2003), people can infer that they
have caused an outcome whenever it matches with their prior thoughts, regardless of whether
they were its true cause or not. In this way, self-agency over an outcome can be established if
the representation of an outcome was activated by a prime before the action was performed and
the outcome was observed.
In an initial study demonstrating this notion (Wegner & Wheatley, 1999), participants and a
confederate jointly controlled the movement of a mouse cursor over objects (e.g., red car,
phone) that were presented on a computer screen. On some (forced) trials, participants received
information over their headphones about one of these objects on which the confederate would
subsequently force the participant to stop, unbeknownst to the latter. In other (unforced) trials,
the confederate allowed the participant to stop on an object. Comparisons between the distance
to an object on the screen and the primed object on forced and unforced trials showed no
differences, suggesting that auditory exposure to objects did not cause participants to stop on
the objects (and thus motor prediction processes did not produce an efference copy). After the
cursor stopped, participants were asked to indicate to what extent they felt they had caused the
cursor to stop on this position. Enhanced self-agency was reported when they received
information (p.202) concerning the stop position as compared to when they did not have this
prior knowledge. Since the experimental setup did not allow the motor prediction process to
produce reliable input for establishing a sense of agency, the agency experiences have likely
resulted from inference processes based on the congruency between the outcome primes and
the actual outcomes (Synofzik, Vogerau, & Voss, 2013).
The occurrence of prime-based self-agency inferences has been shown in both Eastern and
Western cultures (Aarts, Oikawa, & Oikawa 2010; Sato, 2009) and replicates when different kind
Page 3 of 16
Operant Actions
of tasks, primes, and outcomes are used (Aarts et al., 2005; Dannenberg, Frster, & Jostmann,
2012; Gentsch & Schtz-Bosbach, 2011; Jones, de-Wit, Fernyhough, & Meins, 2008; Linser &
Goscke, 2007; Ruys & Aarts, 2012; Van der Weiden, Aarts, & Ruys, 2010). For instance, it has
been demonstrated that the effect is not merely limited to spatial primes but also occurs for
primes that have the same shape as the outcome (i.e., left- and right-pointing arrows) and
primes that activate an abstract representation of the outcome (i.e., using color words to prime
colored circles; Linser & Goscke, 2007). Furthermore, it has been demonstrated that in addition
to non-social action-outcomes, the priming effect is also present for socially meaningful
outcomes (Ruys & Aarts, 2012). Specifically, participants report increased self-agency over the
emotional expression in another person if they are primed with this expression (compared to
when they were not primed) prior to executing an action that is likely to elicit this emotion.
These examples demonstrate that humans readily rely on implicit processes to arrive at agency
judgments.
Boundary Conditions of Prime-Based Inferences

Thus far, we have discussed that agency experiences can arise from both motor predictive and
inferential cues, the individual weight of which is determined by their reliability (e.g., Moore &
Fletcher, 2012). When the generally reliable motor predictive processes are unreliable, people
can still experience agency over outcomes if there is a match between pre-activated knowledge
regarding action-effects and the actual consequences following an action. Importantly, these
inferences of agency can even occur when implicit priming activates prior information about
specific outcomes that one knows can potentially follow from ones own action. Whereas the
implicit outcome priming effects on self-agency inferences seem to occur without much
conscious thought and thus may appear to be a rather automatic process, recent research
suggests that there are boundary conditions to these effects. In the following sections we will
discuss two of them that are relevant in the context of operant actions.
(p.203) Perceptions of Causality

Although an abundance of evidence has shown that priming an outcome before its occurrence
can enhance experienced agency over outcomes, it seems unlikely that this effect is always
present, even in the absence of a logical causal relationship between ones action and a
following effect (Van der Weiden, Aarts, & Ruys, 2011). For instance, one generally does not
believe that a press on a doorbell has caused a light to go on, even if knowledge of this event
was somehow pre-activated in ones mind. Hence, knowledge about potential causality of the
relationship between actions and following events seems to be a prerequisite for (prime-based)
inferences of agency to occur in operant actions. An initial test of this idea was provided in a
study by Sato (2009) in which participants were primed with an outcome that was congruent,
incongruent, or neutral (i.e., unrelated) with the outcome that followed their key press. In
addition, perceived causality between the action and the following effect was manipulated by
varying the contingency between actions and following effects to be either 50% or 75%. In line
with the above, priming congruent outcomes only enhanced experienced agency compared to
priming neutral outcomes when the contingency between the key press and the following effect
was high (i.e., causality was perceived), whereas this effect was absent when the contingency
was low (i.e., at the level of chance).
Page 4 of 16
Operant Actions
These results provide preliminary evidence for the idea that knowledge about potential causality
is necessary for prime-based inference to occur (consistent with models allowing for integrating
multiple cues for agency determination; e.g., Synofzik et al., 2008). However, a direct test of this
notion requires the inclusion of a condition in which agency over outcomes cannot be based on
learning and knowledge about the specific causal relation between actions and the outcomes. In
a recent experiment performing this test (Van der Weiden et al., 2011), participants were
instructed to stop rapidly presented letter strings that ostensibly covered two alternating
neutral words (i.e., glass or book), by pressing a left or right key. Subsequently one of the
words would appear on the screen and participants were asked to indicate to what extent they
believed their key press caused this outcome, which was either briefly primed or not just before
the key press. Participants who explicitly learned that there was a high contingency (80%)
between a key press and a specific outcome (e.g., left results in glass, and right in book)
reported to experience more agency over this outcome if it was primed (versus not primed)
compared to participants who had learned there was no causal relationship between the action
and the action outcome (50% contingency). Thus, when subjects learned that the outcomes
could not be causally related to the action they performed, both motor prediction and inference
processes no longer contributed to experiences of agency.
(p.204) Interestingly, the results further showed that outcome primes did enhance experienced
self-agency when participants had no knowledge about the action-outcome relations over which
self-agency was assessedthat is, when left and right key presses were learned to be unrelated
(50% co-occurrence) to soap and pen, while self-agency was assessed over book and
glass. Notably, these priming effects were as strong as when actions and outcomes were
learned to be causally related (80% co-occurrence). In other words, when individuals did not
have an appropriate causal model between action and outcome, the outcome primes were the
most reliable agentic cue in the action-outcome task at hand. This might fit well with the daily
experience of being able to ascribe outcomes to oneself that have been caused for the first time.
Together, then, these findings provide further support that cognitive inferential processes can
affect experiences of self-agency independent of motor predictive processes.
Level of Action Identification

One explanation that has been offered for the influence of perceived causality on prime-based
agency inferences is that learning about the causal relationship between actions and outcomes
changes the way in which people represent their behavior (Van der Weiden et al., 2010).
According to Action Identification Theory, people identify their actions on multiple hierarchical
levels (Vallacher & Wegner, 1987). At a high level, people identify their behavior in terms of
consequences (i.e., why they are doing things), whereas at a low level, they tend to focus on how
they are doing something (Vallacher & Wegner, 1987). To illustrate, the act of drinking water
can be seen as both renewing ones energy (i.e., high-level representation) and as grasping a
glass (i.e., low-level representation). A recent experiment has demonstrated that the levels at
which people identify their behavior affect outcome-priming effects in self-agency inferences
(Van der Weiden et al., 2010). Specifically, it was demonstrated that experienced agency was
rated higher after action outcomes were primed (vs. not primed) by participants who
represented their behavior at a high level and focused on the production of this outcome (i.e.,
stopping a stimulus on a particular position) compared to participants who represented their
behavior at a low level and focused on the execution of their action (i.e., pressing a button).
Page 5 of 16
Operant Actions
Although this study did not include a baseline condition, the pattern of data (in combination with
earlier findings with this experimental task) suggests that outcome priming reduced the
experience of agency when people represented their behavior at the action execution level, (p.
205) rather than enhancing the sense of agency as a result of representing behavior in terms of
producing outcomes.
More generally, these results might underlie the previously mentioned influence of perceived
causality on experienced agency (Van der Weiden et al., 2011). That is, when people explicitly
learn that outcomes cannot be causally related to the action they performed, and hence their
own actions do not control the occurrence of the outcomes, they might be encouraged to focus
on how they executed their action and represent their behavior at a low level. Alternatively,
when they explicitly learned that their actions and effects are causally related, this might trigger
them to focus on the implications of their actions and represent their behavior at a high level
(Van der Weiden et al., 2011; cf. Vallacher & Wegner, 1987). These changes in identification
level might in turn affect agency experiences as described earlier. When generalizing this line of
reasoning, it might be the case that other factors that influence the level of behavior
representation will affect prime-based inferences underlying experienced agency as well. For
instance, acting impulsively or failing to execute a task might cause participants to represent
their behavior at a low level (i.e., how an action should be executed), and hence is likely to
decrease the impact of outcome-priming on experienced agency (Dannenberg et al., 2012;
Vallacher & Wegner, 1987). In contrast, the strength of prime effects on agency inferences
might be enhanced in people who have a strong belief in self-causation or free will, as these
individuals are likely to identify and attend to their actions in terms of their consequences
(Baumeister, Masicampo, & DeWall, 2009; Vohs & Schooler, 2008).
Comparing Agency Inferences for Goals and Primes

So far, we have addressed the influence of primed outcomes on experienced agency (i.e.,
implicitly triggered outcome representations during action-outcome performance), and we have
shown that this influence is more predominant when individuals know that their actions can
potentially produce a specific outcome and when they think about their behavior in terms of
producing outcomes. However, whereas the (implicit) priming of outcomes clearly modulates
agency experiences, these experiences also follow from the goals (i.e., desired outcomes) that
people explicitly set and aim to attain, raising the question of whether primes and goals affect
self-agency in a similar or different way.
It is known that when people set and act on goals, they tend to focus their attention on the
specific outcome they want to achieve and, at the same time, inhibit information of other
(irrelevant and associated) outcomes (Aarts, 2012; (p.206) Fishbach & Ferguson, 2007;
Frster, Liberman, & Friedman, 2007). This way, explicitly set goals render people sensitive to
deviations from observed outcomes. In case of a match between goal and outcome, agency is
enhanced, whereas in case of a mismatch, agency experiences are more likely to be attenuated
(Van der Weiden, Ruys, & Aarts, 2013). However, the similarity between goal-based agency
inferences and prime-based agency inferences does not imply that the processes underlying
these inferences are the same. For instance, the mere priming of outcome information may
differ from the explicit setting of goals in the sense that priming does not necessarily install the
attentional control processes that are accompanied with the explicit setting of goals, but may
Page 6 of 16
Operant Actions
merely enhance the mental accessibility of the representation of the outcome and associated
outcomes by the spread of activation (Anderson, 1984; Collins & Loftus, 1975). Accordingly,
recent work has started to explore possible differences between goals and primes in inferences
of agency.
In a first attempt to examine these differences (Van der Weiden, Ruys et al., 2013; experiment
3), participants completed an adaptation of an action-outcome task employed in previous
research to assess inferences of self-agency (Aarts et al., 2005; Jones et al., 2008). Specifically,
they observed two squares moving randomly across eight locations on a grid and were made to
believe that one of these squares was under their control, whereas the computer controlled the
other square. Participants could ostensibly stop the movement of their square by pressing a
key. This action was followed by the presentation of an outcome position that either represented
the stop-position of their own square or the computers square. In the goal-condition,
participants received the explicit goal to stop their square on a particular position; in the prime-
condition, this position was briefly presented. Crucially, the degree of matching between the
pre-activated information and the actual outcomes was varied (i.e., pre-activated knowledge
could either match the outcome, or mismatch with increasing distance). Both goals and primes
that matched (versus mismatched) with actual action-effects resulted in increased agency.
However, differences between the types of pre-activation were observed when examining
experienced agency as a function of mismatch degree. That is, when a goal mismatched an
outcome, an immediate reduction in experienced agency was observed, regardless of the extent
of the mismatch. In contrast, when an outcome was primed, the decrease in self-agency
experience between matches and mismatches was smaller and was followed by a gradual
decrease that became greater with the distance between primes and outcomes. These
observations concur with the idea that goal-based inferences are specific and sensitive to
deviations, whereas outcome primes can exert an influence on self-agency inferences by
spreading activation to other associated outcome representations (Van der Weiden, Ruys, et al.,
2013).
(p.207) It is important to emphasize that the differential effects between goal-based and
primed-based agency inferences addressed above mainly follow from the assumption that
implicit outcome primes do not represent active goals that people aim to attain, and hence do
not install the attentional control processes that are accompanied with the explicit setting of
goals. However, recent work suggests that under specific circumstances goals can trigger
attentional control processes even when these goals are implicitly primed and operate without
the person being aware of the activated goal at hand (Bargh, Gollwitzer, & Oettingen, 2010;
Custers & Aarts, 2010; Moskowitz, Li, & Kirk, 2004). For example, implicitly primed goals are
more likely to control attention and behavior when the goal is represented as an action-outcome
and attached to positive affect, thus operating as a desired outcome that one is motivated to
attain (Custers & Aarts, 2010; Marien, Aarts, & Custers, 2013). Interestingly, though, the idea
that our goal pursuits also materialize non-consciously may sound counterintuitive because the
actions we conduct and the outcomes they produce are often accompanied with experiences of
agency. This raises the important and intriguing question of how one arrives at the experience
of self-agency when goal-directed behavior is the mere product of priming.
Page 7 of 16
Operant Actions
A possible answer to this question is that agency experiences not only arise from our explicitly
set goals, but accompany our implicitly primed goals as well, leading us to believe that the
outcomes of our behaviors were consciously intended, whereas in fact they were influenced by
cues in our environment outside our conscious awareness. However, we do not know yet
whether implicitly induced goals function as outcome primes or, alternatively, affect
experienced self-agency in the same way that explicitly set goals do. Future research could
explore this issue by investigating the effect of matches and mismatches on experienced self-
agency as a function of implicitly and explicitly activated goals.
Summary
Initial explanatory models of the emergence of self-agency experiences have been focused on
the role of sensorimotor predictions as agentive cues. However, under some circumstances
these cues do not have sufficient weight to explain agency experiences, such as when the causal
relationship between actions and following effects is weak. In these situations the degree of
experienced agency is likely to be determined by cognitive inferences upon observing an
outcome. Specifically, experiences of agency can arise whenever there is a match between
actual action-effects and pre-activated knowledge of these effects. This pre-activated knowledge
can be derived via an explicit route, from explicitly set goals, as well as via an implicit route (p.
208) (via outcome primes). Inferences based on the latter route are especially likely to occur
when individuals perceive their actions as a potential cause of outcomes, and represent their
behavior in terms of producing outcomes upon performing operant actions.
Recent Developments
Agency as a Product of Predictions and Inferences
Although the present chapter has been predominantly focused on agency experiences that have
been informed by inferential processes, these experiences can also emerge from the predictive
motor cues that have been discussed earlier. The idea that the sense of agency can arise from an
interaction between these types of cues has received increasing theoretical (Frith, 2013, Moore
& Fletcher, 2012; Synofzik et al., 2008; Synofzik et al., 2013) and empirical support (Moore &
Haggard, 2008; Moore et al., 2009; Sato, 2009). A question that is raised by these observations,
however, is how the different processes underlying agency experiences interact.
Recent optimal integration accounts have suggested that the extent to which each type of cue
contributes to experienced agency depends on its reliability (Moore & Fletcher, 2012; Moore et
al., 2009; Synofzik et al., 2013; see also Synofzik, Chapter 13 of this volume; Fletcher &
Fotopoulu, Chapter 16 of this volume), which in turn is determined by factors such as variance,
salience, preciseness, and so on (Synofzik et al., 2013). Since motor predictions are more
precise and are more rapidly generated than inferences, the sense of agency will usually depend
on these predictions (Synofzik et al., 2013). However, there are circumstances under which
motor predictive processes become less reliable. From this perspective, it would be interesting
and important to further explore how people rely on inferential processes to create the sense of
agency in their mind, and when and how explicit goal and outcome primes gain more weight in
this process. For example, goals facilitate attentional processes that cause people to focus on
the specific outcome and render them more sensitive to deviations of outcomes compared to
outcome primes (Van der Weiden, Ruys, et al., 2013). Thus goals might be a more reliable guide
for establishing agency compared to outcome primes that are more transient and unstable,
Page 8 of 16
Operant Actions
and hence, goals are more likely to overrule the unreliable (or even absent) input deriving from
motor prediction processes than primes.
(p.209) Disruptions of Agency

Thus far, we have elucidated processes underlying agency inferences in healthy individuals.
Research has begun to systematically explore in what way these processes are altered in
patients who suffer from disturbed agency experiences. One group of patients known to show
such disturbances are people with checking symptoms (a dimension of obsessive-compulsive
disorder; Belayachi & Van der Linden, 2010). The checking behavior of these individuals is often
accompanied by the inability to experience completion of actions (Coles, Frost, Heimberg, &
Rhaume, 2003; Tolin, Brady, & Hannan, 2008; see also Gentsch & Schtz-Bosbach, Chapter 9
of this volume). A possible mechanism underlying these feelings is a dysfunction in the implicit
process involved in the sense of agency that compares expectations and actual outcomes.
Indeed, individuals with proneness to checking have been shown to experience less agency over
a primed outcome than individuals who do not display this proneness, whereas these two groups
have equal agency experiences when the outcome is not primed (Belayachi & Van der Linden,
2010).
Another mental disorder that has been associated with disturbances in agency experiences is
schizophrenia. Patients suffering from this disorder can be characterized by both excessive
feelings of control over externally produced outcomes (i.e., megalomania) and the ascription of
self-produced outcomes to external sources (i.e., delusions of influence) (Schneider, 1955;
interpretation by Haggard, Martin, Taylor-Clarke, Jeannerod, & Franck, 2003). A previously
examined account for these diverging symptoms concerns a deficit in the motor comparison
process as specified by the comparator model (Voss et al., 2010). Specifically, the motor system
of schizophrenic patients seems less able to reliably predict the sensory consequences of their
actions, preventing sensory attenuation of reafferent sensory information (Blakemore et al.,
2002; Lindner, Thier, Kircher, Haarmeier, & Leube, 2005; Shergill, Samson, Bays, Frith, &
Wolpert, 2005; Synofzik, Thier, Leube, Schlotterbeck, & Lindner, 2010; Voss et al., 2010).
According to the optimal integration models that have been discussed earlier (e.g., Moore &
Fletcher, 2012; Synofzik et al., 2013), this reduced reliability of motor predictive signals will
cause schizophrenic patients to rely more heavily on perceptual and possibly also non-motor
inferential cues. However, non-motor inferential cues are far more fallible and less robust than
motor predictive signals and accordingly might lead to erroneous agency experiences (Synofzik
et al., 2013). In particular, participants might overestimate to what extent they caused events
due to excessive reliance on environmental cues and erroneous beliefs or, oppositely, may fail to
perceive themselves as the cause of produced outcomes if these cues are absent or remain
unnoticed due to lapses in attention (Synofzik et al., 2013).
(p.210) In line with the suggestions of the optimal integration models, a recent study has
demonstrated that agency experiences resulting from predictions are disturbed in patients
suffering from schizophrenia, whereas excessive inferences of agency are observed compared to
controls (Voss et al., 2010). It is important to note that this study only considered goal-directed
agency inferences as participants acted voluntarily to cause a certain effect. Given that goal-
directed and prime-based inference processes have been shown to be qualitatively different (Van
Page 9 of 16
Operant Actions
der Weiden, Ruys, et al., 2013), they might be differently affected in patients suffering from
schizophrenia. In a study addressing this issue it was demonstrated that schizophrenic patients
and healthy individuals showed equal increases in experienced agency when their goal to
produce a certain effect matched with the actual effect following their action (Renes,
Vermeulen, Kahn, Aarts, & Van Haren, 2013). However, only healthy controls displayed
increased agency over an outcome that was previously primed, suggesting that schizophrenic
patients rely less on primes to infer agency (Renes, Vermeulen, et al., 2013).
In short, schizophrenic patients seem to be hampered in establishing a sense of agency over

action-outcomes as result of two suboptimal or disturbed processes: (1) motor-prediction
processes seem to be unreliable to offer effective input for agency; and (2) inferences of agency
do not seem to be based on the accessibility of relevant outcome information that is implicitly
pre-activated before observing the action-outcome.
Neural Substrates of Agency

One way to gain further insight into the emergence of self-agency and its disturbance in
neurological and psychiatric disorders is to examine the neural basis of this experience. To date,
the abundance of research has focused on motor predictions as agentive cues and accordingly
has employed paradigms in which visual movement feedback is manipulated (David, 2012;
Sperduti, Delaveau, Fossati, & Nadel, 2011). Incongruent or delayed feedback about an action
and the accompanying reduced sense of agency have been associated with activity in various
areas, including the superior temporal gyrus, the inferior parietal lobe, dorsomedial prefrontal
cortex, precuneus, as well as with areas implicated in motor processes such as the pre-
supplementary motor area and the cerebellum (David, Newen, & Vogeley, 2008; Sperduti et al.,
2011). On the other hand, congruent feedback and self-agency ascriptions have been associated
with activation in the insula, primary somatosensory cortex, and premotor cortex (Sperduti et
al., 2011).
In contrast to motor predictions, research on cognitive inferences as cues of agency has

received little attention. One important exception is a recent (p.211) study demonstrating
neural activity associated with goal-based agency inferences (Renes, Van Haren, Aarts, & Vink,
2013). In this study, functional magnetic resonance imaging (fMRI) was used to measure brain
activity while participants reported their experience of self-agency from the degree of
correspondence between goals and actual action-outcomes. During high-agency experiences
over outcomes that matched prior goals (as opposed to low-agency experiences over outcomes
that mismatched prior goals), activation was observed in the inferior parietal lobe, left superior
frontal cortex, right superior frontal, and medial prefrontal cortex. Interestingly, these areas are
also implicated in a neural network for conscious awareness and the self (see also Tsakiris,
Chapter 10 of this volume), such as that expressed in the global workspace model (Baars, 2002;
Dehaene & Naccache, 2001).
An important question following from these observations is how the identified regions
communicate, especially since it is becoming increasingly clear that complex cognitive
processes, such as the experience of agency, require the integration of information that is
generated at different brain sites (Varela, Lachaux, Rodriguez, & Martinerie, 2001). To date,
only a few studies on the neural basis of agency have considered interactions between brain
Page 10 of 16
Operant Actions
regions (David, 2012). In one study, incongruent visual feedback was associated with a leading
network, consisting primarily of the left anterior inferior parietal lobe, the right supramarginal
gyrus, the right temporoparietal junction, and the anterior insula, which was sending
information to a lagging network, consisting mainly of the cingulate, posterior inferior parietal
lobe, and the prefrontal lobe (Nahab et al., 2011). The authors speculated that the leading
network is likely to be involved in mismatch detection between motor predictions and actual
action effects, whereas the lagging network translates the outcome of this comparison into a
conscious agency experience.
Although Nahab and colleagues (2011) emphasized agency experiences informed by motor
predictions, the parietal and frontal regions are also associated with cognitive inferences (Renes
Van Haren, et al., 2013) and might fulfill a similar role considering that both motor prediction
and inference processes involve a comparison between predicted outcomes and actual outcomes
that can be translated to a conscious agency experience (cf., Dehaene & Naccache, 2001).
Indeed, activation in parietal and frontal regions has been observed during mismatches between
motor predictions and action effects (i.e., ascription of outcomes to external sources; David, et
al., 2008; Sperduti et al., 2011), as well as during matches between pre-activated outcome
representations and observed outcomes (i.e., ascription of outcomes to oneself; Renes, Van
Haren, et al., 2013).
What remains unclear from these findings, however, is whether the same regions and
interactions are involved in prime-based inferences as (p.212) in goal-based inferences. In a
recent study we explored this issue by examining cortical interactions underlying both types of
inferences by means of electroencephalographic recordings (Dogge, Hofman, Boersma,
Dijkerman & Aarts, 2014). This technique allows for the assessment of phase synchronization of
neural oscillations (Sauseng & Klimesh, 2008), which has been proposed as the mechanism
underlying neural communication (Buzski & Draguhn, 2004; Fries, 2005; Sauseng & Klimesh,
2008; Varela et al., 2001). Accordingly, we were able to relatively directly assess the interaction
between brain regions during goal-directed and prime-based agency inferences in a simple
action-effect task. Inspections of connectivity patterns indicated that goal-based agency
inferences rely on frontoparietal connectivity, and that this connectivity was reduced in prime-
based agency inferences. A possible explanation for this finding is that goals facilitate
attentional processes that more strongly involve the frontoparietal network, while outcome
primes result in a noisier processing mechanism because they recruit this network to a lesser
degree. These preliminary results thus provide a first glimpse at possible differences between
prime-based and goal-driven processes during agency inferences and at the same time
demonstrate the potential of recent methodological advances in the quantification of brain
dynamics to elucidate the neural basis underlying agency processing.
Conclusion
The experience of causing external events through actions plays a pivotal part in everyday life.
Knowing whether or not one causes ones own actions and the outcomes they produce is not
only essential for self-awareness and identity, but also for understandingand interacting with
others. Whereas experienced agency over simple, unambiguous movements can generally be
informed by predictions that are part of the motor control process, people often find themselves
in more ambiguous situations in which the motor system cannot reliably predict the sensory
Page 11 of 16
Operant Actions
consequences of their actions. In these situations, self-agency experiences are informed by

cognitive inferences of congruence between prior thoughts about action-outcomes and the
actual outcomes. Importantly, the pre-activated outcome information can result from the explicit
goals that one sets and the voluntary experiences that one has when selecting and executing
ones behavior. However, whereas the goal-based agency inferences can be a reliable cue to
arrive at a sense of agency, agency may also follow from a more prime-based inference that
allows one to experience self-causation over operant actions that are triggered by environmental
cues. The present chapter has examined recent work on explicit and implicit self-agency
inferences, and we hope that the next generation of research in (p.213) this area will illuminate
how these inferences interact with other agentive cues and how they are biologically
implemented in producing the experience of self-agency in the human mind.
References
Aarts, H. (2012). Goals, motivated social cognition and behavior. In S. Fiske & C. N. Macrae
(Eds.), SAGE handbook of social cognition (pp. 7579). London: Sage.
Enhancing experienced agency by priming effect information. Consciousness and Cognition, 14,
439458.
Aarts, H., Oikawa, M., & Oikawa, H. (2010). Cultural and universal routes to authorship
ascription: effects of outcome priming on experienced self-agency in The Netherlands and Japan.
Journal of Cross-Cultural Psychology, 41, 8798.
Anderson, J. R., & Pirolli, P. L. (1984). Spread of activation. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 10, 791.
Baars, B. J. (2002). The conscious access hypothesis: origins and recent evidence. Trends in
Bandura, A. (1986). Social foundations of thought and action: a social-cognitive theory.

Englewood Cliffs, NJ: PrenticeHall.
Bargh, J. A., Gollwitzer, P. M., & Oettingen, G. (2010). Motivation. In S. T. Fiske, D. T. Gilbert, &
G. Lindzey (Eds.), Handbook of social psychology (pp. 286316). New York: Wiley.
Baumeister, R. F., Masicampo, E. J., & DeWall, C. N. (2009). Prosocial benefits of feeling free:
disbelief in free will increases aggression and reduces helpfulness. Personality and Social
Psychology Bulletin, 35, 260268.
Belayachi, S., & Van der Linden, M. (2010). Feeling of doing in obsessivecompulsive checking.
Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (2002). Abnormalities in the awareness of action.
Page 12 of 16
Operant Actions
Buzski, G., & Draguhn, A. (2004). Neuronal oscillations in cortical networks. Science, 304,
19261929.
Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. New York: Cambridge
University Press.
Coles, M. E., Frost, R. O., Heimberg, R. G., & Rhaume, J. (2003). Not just right experiences:
Perfectionism, obsessivecompulsive features and general psychopathology. Behaviour Research
and Therapy, 41, 681700.
Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing.

Custers, R., & Aarts, H. (2010). The unconscious will: how the pursuit of goals operates outside
of conscious awareness. Science, 329, 4750.
Dannenberg, L., Frster, J., & Jostmann, N. B. (2012). If only . . . : When counterfactual
thoughts can reduce illusions of personal authorship. Consciousness and Cognition, 21, 456463.
David, N. (2012). New frontiers in the neuroscience of the sense of agency. Frontiers in Human
David, N., Newen, A., & Vogeley, K. (2008). The sense of agency and its underlying cognitive
and neural mechanisms. Consciousness and Cognition, 17, 523534.
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self- determination in human
behavior. New York: Plenum Press.
Dehaene, S., & Naccache, L. (2001). Towards a cognitive neuroscience of consciousness: Basic
evidence and a workspace framework. Cognition, 79, 137.
of agency. Consciousness and Cognition, 20, 12111220.
Dogge, M., Hofman, D., Boersma, M., Dijkerman, H. C., & Aarts, H. (2014). Cortical information
flow during inferences of agency. Frontiers in Human Neuroscience, 8.
Dogge, M., Schaap, M., Custers, R., Wegner, D. M., & Aarts, H. (2012). When moving without
volition: implied self-causation enhances binding strength between involuntary actions and
effects. Consciousness and Cognition, 21, 501506.
Fishbach, A., & Ferguson, M. J. (2007). The goal construct in social psychology. In A. W.
Kruglanski & E. T. Higgins (Eds.), Social psychology: handbook of basic principles (pp. 490515).
New York: Guilford Press.
Frster, J. A., Liberman, N., & Friedman, R. (2007). Seven principles of goal activation: a
systematic approach to distinguishing goal priming from priming of non-goal constructs.
Personality and Social Psychology Review, 11, 211233.
Page 13 of 16
Operant Actions
Fries, P. (2005). A mechanism for cognitive dynamics: neuronal communication through

neuronal coherence. Trends in Cognitive Sciences, 9, 474480.
Frith, C. (2013). The psychology of volition. Experimental Brain Research, 229, 289299.
Frith, C. D., Blakemore, S. J., & Wolpert, D. M. (2000). Explaining the symptoms of
schizophrenia: abnormalities in the awareness of action. Brain Research Reviews, 31, 357363.
Gentsch, A., & Schtz-Bosbach, S. (2011). I did it: unconscious expectation of sensory
consequences modulates the experience of self-agency and its functional signature. Journal of
Cognitive Neuroscience, 23, 38173828.
Haggard, P., Martin, F., Taylor-Clarke, M., Jeannerod, M., & Franck, N. (2003). Awareness of
action in schizophrenia. Neuroreport, 14, 10811085.
Jones, S. R., de-Wit, L., Fernyhough, C., & Meins, E. (2008). A new spin on the Wheel of Fortune:
priming of action-authorship judgements and relation to psychosis-like experiences.
Lindner, A., Thier, P., Kircher, T. T., Haarmeier, T., & Leube, D. T. (2005). Disorders of agency
in schizophrenia correlate with an inability to compensate for the sensory consequences of
actions. Current Biology: CB, 15, 11191124.
Linser, K., & Goschke, T. (2007). Unconscious modulation of the conscious experience of
voluntary control. Cognition, 104, 459475.
Marien, H., Aarts, H., & Custers, R. (2013). Adaptive control of human action: the role of
outcome representations and reward signals. Frontiers in Psychology, 4(602).
integration approaches. Consciousness and Cognition, 21, 5968.
Moore, J., & Haggard, P. (2008). Awareness of action: inference and prediction. Consciousness
and Cognition, 17, 136144.
Moore, J. W, Wegner, D. M, & Haggard, P. (2009). Modulating the sense of agency with external
Moskowitz, G. B., Li, P., & Kirk, E. R. (2004). The implicit volition model: on the preconscious
regulation of temporarily adopted goals. In M. P. Zanna (Ed.), Advances in experimental social
Nahab, F. B., Kundu, P., Gallea, C., Kakareka, J., Pursley, R., Pohida, T., & Hallett, M. (2011).
The neural processes underlying self-agency. Cerebral Cortex, 21, 4855.
Renes, R. A., Van Haren, N. E. M., Aarts, H., & Vink, M. (2013). The neural correlates of self-
agency inferences. Unpublished manuscript.
Page 14 of 16
Operant Actions
Renes, R. A., Vermeulen, L., Kahn, R. S., Aarts, H., & van Haren, N. E. M. (2013). Abnormalities
in the establishment of feeling of self-agency in schizophrenia. Schizophrenia Research, 143, 50
54.
Ruys, K. I., & Aarts, H. (2012). I didnt mean to hurt you! Unconscious origins of experienced
self-agency over others emotions. Emotion, 12, 132141.
effect contribute to explicit judgment of agency. Cognition, 110, 7483.
Sauseng, P., & Klimesch, W. (2008). What does phase information of oscillatory brain activity
tell us about cognitive processes? Neuroscience and Biobehavioral Reviews, 32, 10011013.
Schneider K. Klinische Psychopathologie. Stuttgart: Thieme Verlag, 1955.
Shergill, S. S., Samson, G., Bays, P. M., Frith, C. D., & Wolpert, D. M. (2005). Evidence for
sensory prediction deficits in schizophrenia. American Journal of Psychiatry, 162, 23842386.
Sperduti, M., Delaveau, P., Fossati, P., Nadel, J. (2011). Different brain structures related to self-
and external-agency attribution: a brief review and meta-analysis. Brain Structure and Function,
216, 151157.
Synofzik, M., Thier, P., Leube, D. T., Schlotterbeck, P., & Lindner, A. (2010). Misattributions of
agency in schizophrenia are based on imprecise predictions about the sensory consequences of
ones actions. Brain, 133, 262271.
Synofzik, M., Vosgerau, G., & Voss, M. (2013). The experience of agency: an interplay between
prediction and postdiction. Frontiers in Psychology, 4, 18.
Tolin, D. F., Brady, R. E., & Hannan, S. (2008). Obsessional beliefs and symptoms of obsessive
compulsive disorder in a clinical sample. Journal of Psychopathology and Behavioral Assessment,
30, 3142.
Vallacher, R. R., & Wegner, D. M. (1987). What do people think theyre doing? Action
identification and human behavior. Psychological Review, 94, 315.
Van der Weiden, A., Aarts, H., & Ruys, K. I. (2010). Reflecting on the action or its outcome:
Behavior representation level modulates high level outcome priming effects on self- agency
experiences. Consciousness and Cognition, 19, 2132.
Van der Weiden, A., Aarts, H., & Ruys, K. I. (2011). Prime and probability: causal knowledge
affects inferential and predictive effects on self-agency experiences. Consciousness and
Cognition, 20, 18651871.
Page 15 of 16
Operant Actions
Van der Weiden, A., Aarts, H., Ruys, K. I. (2013). On the nature of experiencing self-agency: the
role of goals and primes in inferring oneself as the cause of behavior. Social and Personality
Psychology Compass, 7, 888904.
Van der Weiden, A., Ruys, K. I., & Aarts, H. (2013). A matter of matching: how goals and primes
affect self-agency experiences. Journal of Experimental Psychology: General, 142, 954966.
Varela, F., Lachaux, J. P., Rodriguez, E., & Martinerie, J. (2001). The brainweb: phase
synchronization and large-scale integration. Nature Reviews Neuroscience, 2, 229239.
Vohs, K. D., & Schooler, J. W. (2008). The value of believing in free will encouraging a belief in
determinism increases cheating. Psychological Science, 19, 4954.
Voss, M., Moore, J., Hauser, M., Gallinat, J., Heinz, A., & Haggard, P. (2010). Altered awareness
of action in schizophrenia: a specific deficit in predicting action consequences. Brain, 133, 3104
3112.
Wegner, D. M., Sparrow, B., & Winerman, L. (2004). Vicarious agency: experiencing control
over the movements of others. Journal of Personality and Social Psychology, 86, 838848.
will. American Psychologist, 54, 480.
Wolpert, D. M., & Flanagan, J. R. (2001). Motor prediction. Current Biology, 11, R729R732.
Wolpert, D. M., Ghahramani, Z., & Jordan, M. I. (1995). An internal model for sensorimotor
integration. Science-AAAS-Weekly Paper Edition, 269, 18801882.
Page 16 of 16
Agency and Outcome Prediction

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001

Antje Gentsch
Simone Schtz-Bosbach
DOI:10.1093/acprof:oso/9780190267278.003.0009

This chapter argues that the experience of authorship strongly depends on the capacity to form
predictions about the outcome of an action. It reviews cognitive psychology and neuroscience
research indicating that subjective agency experience depends on the integrity of outcome
predictions generated by internal forward models in the brain. Outcome predictions can be used
for filtering sensory input through a continuous comparison with the actual sensory feedback.
Current computational models explaining agency in terms of forward modeling and predictive
coding are discussed. Moreover, scientific work has demonstrated that dysfunctional outcome
prediction may contribute to aberrant agency experience in schizophrenia and obsessive-
compulsive disorder. The chapter concludes with a discussion about the role and nature of
internal predictions underlying functional and dysfunctional experience of agencycritically
distinguishing the psychological and neural mechanisms operating at different
phenomenological levels of agency.
Keywords: authorship, predictions, action, outcomes, sensory input, computational modeling
Agency Beyond Body and Mind

In our everyday life we deliberately or accidentally do things that change the world around us in
particular ways. We are able to determine the occurrence or course of events by our actions.
These interventions can be in the range of major life events or they can only concern the
accustomed sounds of our walking or voice. The ability to think causally and to experience
control over changes taking place not only within the boundaries of our own body but also in the
states of the external environment are integral parts of our self-experience. This sense of
Page 1 of 17
authorship appears to emerge very early in development. Long before the first signs of mirror
self-recognition, infants between 3 and 12 months of age actively explore self-produced
movements in the mirror and the contingency of mirror reflection. For example, Philippe Rochat
and colleagues (1998; also see Chapter 11 of this volume) reported observations of 3-month-old
infants actively discovering the contingency between their own leg kicks and the sounds of a
rattle. They concluded that by 23 months, infants develop a sophisticated sense of their own
body as an agent in the environment (Rochat & Morgan, 1998, p. 106). The capacity for
instrumental learning and operant action lies at the heart of this basic form of agentive self-
awareness.
(p.218) This initial sense of authorship corresponds to what has been termed minimal self-
awareness, or implicit and embodied sense of self. Philosophical theories on the sense of self
distinguish two states of self-awareness: the pre-reflective, minimal self versus the reflective,
narrative self (Gallagher, 2000). Pre-reflective experience of agency (i.e., minimal self-
awareness) does not depend on introspective capacities of the subject or explicit perceptual
monitoring of movements, but rather is generated by primary processes of perception-action
coupling, involving efferent and reafferent processes. In contrast, reflective forms of authorship
experience rely on processes of belief formation. However, empirical studies suggest that
predictive signals act as an important authorship cue at both reflective and pre-reflective levels
of self-awareness, as we will show.
Action-Effect Anticipation
The view that action-effect anticipation plays a crucial role in the control and perception of
action can be traced back to the first theoretical approach to action cognition, the ideomotor
theory (James, 1890). According to a more recent version, the common coding theory (Hommel,
Musseler, Aschersleben, & Prinz, 2001; Prinz, 1997; see also Hommel, Chapter 14 of this
volume), motor and perceptual events are linked through event codes, which form a common
representational domain of action and perception. The theory assumes that action execution
always involves activating a representation of the distal events typically following the action. In
support of this view, it has been shown that once action-outcome associations have been
learned, the compatibility between action and outcome determine agency experiences. This
compatibility effect was found to be stronger when the presence of efferent signals enables
internal models to generate predictions of the sensory action consequences (Farrer, Franck,
Paillard, & Jeannerod, 2003; Sato & Yasuda, 2005).
Intending to perform an action presumably makes the outcome more likely. However, factors
other than intention can also influence the probability of an outcome. For example, research has
assessed whether the impact of outcome expectations on the sense of agency depends on the
achievement of a particular goal or intended result through the action. Judgments of control
over an intended outcome should be higher when goal achievement was expected and more
likely (e.g., due to effort) as compared to when the achievement was unexpected (e.g., due to
good luck), even if the goal is achieved in both cases. A study by Sato and Yasuda (2005)
confirmed this hypothesis by examining the interplay between intentions and predictions in the
formation of the sense of agency. For both intended and unintended action outcomes, subjects
(p.219) experienced enhanced agency if the actual outcome was congruent with the prediction
based on their own motor response.
Page 2 of 17
At reflective levels of self-awareness, additional support for the idea that anticipatory pre-
activation of upcoming action outcomes shapes agency experience comes from research testing
a theory of apparent mental causation (Wegner, 2002; Wegner & Wheatley, 1999). This theory
suggests that we perceive ourselves as agents when our mind provides us with previews of the
action and its consequences that turn out to be accurate. Accordingly, studies have tested
whether illusions of agency can be induced by employing priming paradigms. It was found that
individuals judge uncontrollable outcomes to be more controllable when they are primed with
the outcome just before each action (Aarts, Custers, & Wegner, 2005; Linser & Goschke, 2007;
Sato, 2009; Wegner & Wheatley, 1999). This control-illusion seems to arise only in the case of a
match between the prime stimulus and the actual action-effect, and independently of whether
participants consciously perceive the prime stimulus (Linser & Goschke, 2007).
Other evidence supporting the idea that action-effect anticipation contributes to our sense of
authorship comes from neuroimaging research. Parietal cortices are believed to be crucial for
action recognition and have been implicated in the processing of agency. For example, research
has revealed that parietal lesions are associated with impaired action attribution (Sirigu,
Daprati, Pradat-Diehl, Frank, & Jeannerod, 1999), and hyperactivity of these brain areas was
observed in schizophrenia patients suffering from delusions of alien control (Spence, Brooks,
Hirsch, et al., 1997). Studies using experimental paradigms that manipulate the congruency
between predictions and action outcomes found a modulation of inferior parietal lobe activity
that was associated with the degree to which participants felt in control of movements (e.g.,
Farrer & Frith, 2002). These results, and others, have led some to conclude that the neural
response to sensorimotor incongruence and prediction error determines self-other attribution of
actions. However, additional research is needed to examine the exact mechanisms by which
these signals lead to pre-reflective feelings versus reflective judgments of agency, which is still
actively debated (Synofzik, Vosgerau, & Newen, 2008).
The Attenuated Self

One phenomenon that often accompanies agentive self-attribution of action is sensory
attenuation: individuals typically show a suppressed perceptual and physiological response
when processing sensory stimuli that result from their own action as compared to externally
produced sensory stimuli. We will refer (p.220) to this suppression response as self-
attenuation, and we argue that it provides one mechanism of self-agency registration.
One of the first experimental studies on this phenomenon was done in the animal visual system.
The results of that study indicated that efferent signals of the motor system cause a cancellation
of visual signals resulting from voluntary eye movement and thereby enable stable vision (von
Holst & Mittelstaedt, 1950). Since then, numerous studies have observed sensory attenuation
associated with self-action. Internal forward models in the motor system were proposed to filter
and attenuate the predicted sensory information (Frith, Blakemore, & Wolpert, 2000a). As a
result, self-generated sensations are thought to be associated with low levels of surprise
responses and low need for increases in cognitive control since the internal bodily state is less
affected. Thereby, it is believed that sensory attenuation implicitly contributes to the sense of
being in control. In the following, evidence from psychophysical, electrophysiological, and
neuroimaging research supporting this view will be reviewed.
Page 3 of 17
Dampened Perceptual Intensity

Psychophysical data in different sensory modalities have indicated that the subjective intensity
of a self-generated sensory event is decreased compared to stimuli produced by another person
or an environmental factor. In the sense of touch, for example, early research on the nature of
ticklish sensations showed that tickle strength is reduced for self-administered tickle compared
with tickle stimuli administered by the experimenter or by involuntary arm movements
(Weiskrantz, Elliott, & Darlington, 1971). Self-attenuation also occurs for the perception of force
and has been suggested to underlie force escalation processes through the perceived
discrepancy between self- and externally applied force (Shergill, Bays, Frith, & Wolpert,
2003). That is, subjects perceive the sensation of force applied to their finger weaker when it is
self-generated than when forces of the same magnitude are applied by another person. In the
auditory modality (see Figure 9.1), it was recently demonstrated that the loudness of self-
generated tones is reduced compared to external tones (Sato, 2008; Weiss, Herwig, & Schtz-
Bosbach, 2011). In these experiments, the auditory event is produced via button press, either by
the subject herself, by a computer, or by another person. The auditory attenuation was shown to
depend on internal motor signals, rather than mere temporal expectation, since it did not occur
even if external tones were made equally predictable (Weiss & Schtz-Bosbach, 2012).
Systematic investigations of the nature of

perceptual self-attenuation revealed that
the precision of motor predictions plays a
key role in determining (p.221) the
perceived intensity of sensory action
outcomes. For example, It was observed
that the tickle sensation diminished with
increasing congruency between the
voluntary movement and the position or
velocity of the tactile stimulus (Blakemore,
Frith, & Wolpert, 1999; Blakemore, Smith,
Steel, Johnstone, & Frith, 2000).
Furthermore, priming the action outcome
prior to action execution has been shown to
enhance self-attenuation (Sato, 2008). A
recent study in the visual modality revealed
Figure 9.1 Example of auditory self-
reduced sensitivity in detecting threshold attenuation. The intensity perception of
visual stimuli when predicted and observed tones (as evaluated by the point of subjective
action consequences matched (Cardoso- equality, dB) was lower for tones that were
self-generated in comparison to tones
Leite, Mamassian, Schtz-Bosbach, &
generated externally via an observed action
Waszak, 2010). Other research supporting of another person or a computer.
the role of motor predictions and self- Source: Weiss, Herwig, & Schtz-Bosbach
attenuation has demonstrated that (2011).
distortion of the spatio-temporal congruency
between action and effect reduces self-
attenuation (Blakemore, Frith, & Wolpert,
2001; Heinks-Maldonado, Mathalon, Gray, & Ford, 2005; Houde, Nagarajan, Sekihara, &
Merzenich, 2002).
Page 4 of 17
(p.222) Much research has examined self-attenuation in patients suffering from a disturbed
sense of agency such as schizophrenia. These studies found evidence for a lack of perceptual
attenuation specifically in schizophrenia patients showing delusions of control and passivity
experiences (Blakemore, Smith, Steel, Johnstone, & Frith, 2000; Shergill, Samson, Bays, Frith, &
Wolpert, 2005). From these results it was suggested that deficits in the prediction and
cancellation of the sensory consequences of actions, as reflected in a lack of self-attenuation,
can result in abnormal agency experience.
Reduced Neural Response

Recently, self-attenuation has also received renewed interest and attention in cognitive
neuroscience. Numerous studies have investigated neuronal discharges in primary sensory
areas following self-agency. For example, activity of the primary somatosensory cortex was
found to be diminished in response to self-produced compared to external touch (Blakemore,
Wolpert, & Frith, 1998). An MEG study reported the somatosensory cortical response to be
about one-fifth weaker when participants touched themselves with a brush than when they were
touched by another person in the same way (Hesse, Nishitani, Fink, Jousmaki, & Hari, 2009).
Similarly, in the auditory and visual modality, EEG studies revealed a reduction in the amplitude
of the N1 event-related potential (see Figure 9.2) following self-generated auditory events
(Curio, Neuloh, Numminen, Jousmaki, & Hari, 2000; Martikainen, Kaneko, & Hari, (p.223)
2005) or visual stimuli (Gentsch & Schtz-Bosbach, 2011; Schafer & Marcus, 1973). Self-
attenuation was already apparent in components occurring only 100 ms following stimulus onset
and reflecting early stages of sensorimotor processing.
It is generally assumed that predictive

forward models in the motor system
underlie self-attenuation during action. The
cerebellum plays an important role in motor
learning, and also contributes to predictive
modulation of self-produced sensation
(Blakemore, Wolpert, & Frith, 1998).
Cerebellar activity was found to depend on
the degree of discrepancy between
predicted and actual sensory consequence
and to correlate with activity in Figure 9.2 Example of self-attenuation in
somatosensory cortices (Blakemore, Frith, the visual system. Self-generated visual
stimuli (Motor-Effect Task, ME) evoked
& Wolpert, 2001). Furthermore, motor smaller N1 amplitudes as compared to
areas of the frontal cortex involved in motor externally generated visual stimuli (Effect-
preparation, such as the supplementary only Task, E). Motor potentials were
subtracted (Motor-only Task, M; Difference
motor area and primary motor cortex, have
ME-M), to allow for an interpretation of the
been implicated in predictive processes direct contrast of conditions without motor
underlying sensory gating (Haggard & confound.
Whitford, 2004; Voss, Ingram, Haggard, & Source: Gentsch & Schtz-Bosbach (2011).
Wolpert, 2006). For example, theta burst
stimulation of the primary motor cortex was
shown to reduce self-attenuation, possibly
Page 5 of 17
by transiently interfering with the predictive process that accompanies the generation of
efferent signals (Voss, Bays, Rothwell, & Wolpert, 2007).
Despite the abundant evidence that sensory action outcomes are associated with self-
attenuation, how this phenomenon is related to the explicit experience of authorship remains
obscure. Will the same cues that enhance explicit authorship also modulate self-attenuation at
the neural level? To investigate this question, we conducted an EEG study using a typical
priming paradigm to induce control illusions in healthy subjects. Participants were primed with
a visual action-effect immediately before each action. EEG data revealed that congruent primes
were associated with enhanced self-attenuation of the visual N1 component (Gentsch & Schtz-
Bosbach, 2011). Moreover, although unaware of the prime, participants judged their own
agency to be higher when they were congruently primed with the upcoming action effect.
In sum, these findings are suggestive of a top-down modulation of the perceptual system by
motor predictions, resulting in attenuation of the sensory consequences of self-action.
Nonetheless, there are still unresolved issues regarding the exact mechanisms and neural
events responsible for this modulation (see also Hughes, Desantis, & Waszak, 2013). The
following section of this chapter highlights two influential computational theories, the
comparator theory and the predictive coding theory, which provide somewhat different accounts
of the sense of agency as an emergent property of brain states involved in top-down prediction
of bottom-up sensory input.
(p.224) Predictive Models of Agency

The notion of motor prediction in the understanding of perceptual capacities was already
considered by the first theories of visual perception (Helmholtz, 1866). More recently, the
development of forward models in motor control (Wolpert, Ghahramani, & Jordan, 1995) and
Bayesian inference in perception (Friston, 2005) has added a computational basis to the
neurocognitive mechanisms underlying predictive processes. Originally, these theoretical
frameworks were developed in order to describe motor control and general perceptual
cognition, respectively. However, they have also been used to explain the emergence of specific
phenomenological experiences associated with self-awareness, such as of having a sense of
agency. The idea common to both is that the sense of agency arises from successful top-down
predictions of the sensory consequences of motor behaviors.
Motor Control and Comparator Theory

Based on computational theories of motor control (Wolpert, Ghahramani, & Jordan, 1995), the
well-established comparator theory (see Figure 9.3A) is still the most prominent theory of
agency and its disturbances (Blakemore, Frith, & Wolpert, 2001; Frith, 1992). According to the
theory, internal forward models attempt to predict the likely sensory consequences of motor
actions on the basis of efferent signals. Besides their primary role in on-line performance
monitoring and motor control, these internal motor models are considered to have an important
role in action awareness and perception. Specifically, sensory prediction signals generated by
forward models are thought to be canceled out of the afferent signal (cf. reafferent principle,
Helmholtz, 1866; von Holst & Mittelstaedt, 1950), and thereby contribute to self-other
distinction and sensory gating of the afferent input (Blakemore, Frith, & Wolpert, 1999). Hence,
Page 6 of 17
successful prediction, that is, congruence of predicted and actual outcome, is thought to lead to
self-registration and a subjective sense of agency.
Evidence for the comparator model has come from both explicit judgments of agency, but also
from self-attenuation, as reviewed earlier. Additional evidence was derived from studies on
psychiatric patients with abnormal agency experiences. Specifically, delusions of control in
schizophrenia, it is argued, can be explained by a dysfunction of the comparator mechanism
involving prediction and afferent signals (Frith, Blakemore, & Wolpert, 2000a). Indicative data
showed higher thresholds in these patients for detecting experimental distortions of action
feedback (Daprati et al., 1997; Franck et al., 2001), possibly due to imprecise internal
predictions preventing reliable perception of self-action. A study by Synofzik and colleagues
(2010) confirmed this finding using spatial (p.225) distortions of pointing movements and
demonstrated an over-reliance on external sensory information in these patients, which
correlated with the presence of delusions of influence.
Page 7 of 17
Recently, there have been critical views on

the use of the comparator model as a sole
explanatory mechanism underlying the
conceptual capacity or incapacity of correct
attribution of agency (Synofzik, Vosgerau, &
Newen, 2008). These are based on findings
that both self- and external attribution can
occur for the same comparator mismatch
(e.g., Farrer et al., 2003) and on research
revealing sources of action-related
sensations that are independent of motor Figure 9.3 Predictive models of agency. (a)
According to the comparator model, the
mechanisms (for a review, see Wegner,
sense of agency is related to an internal
2003). Consequently, an important forward mechanism, which uses an efference
extension of the comparator theory has copy of the motor command predicting the
sensory consequence of an action. If the
been proposed (Moore & Fletcher, 2012;
comparison between predicted and actual
Synofzik, Vosgerau, & Lindner, 2009; see sensory input results in a match, this would
also Synofzik, Chapter 13 of this volume; indicate self-agency. (b) A schematic of
active inference. Prediction errors (dashed
Fletcher & Fotopoulou, Chapter 16 of this
lines) about the consequences of actions are
volume). The new model emphasizes (Bayes- sent forward from error units (E) to state
optimal; precision-weighted) integration of units (S) to update top-down predictions of
internal (p.226) and external cues proprioceptive signals (solid lines) at various
levels of the neurocognitive hierarchy. Motor
operating differently at conceptual and non- action is driven by proprioceptive prediction
conceptual levels of agency registration. errors and directs the sampling of predicted
Accordingly, both sensorimotor signals and sensory input in order to minimize surprise.
conceptual information serve as authorship The sense of agency is considered an
emergent property of successful suppression
cues and are integrated and re-weighted of prediction error. Both models propose
continuously, depending on their availability that prediction mechanisms may underpin
and reliability, in a dynamically changing the distinction between self and non-self and
the experience of agency. A disruption in
environment. The integration of these cues
this system is thought to play a crucial role
is thought to differ between contexts, in the development of positive symptoms in
individuals, and levels of cognitive schizophrenia (Friston, 2012b; Frith,
processing. For example, a basic non- Blakemore, & Wolpert, 2000b).
conceptual feeling of agency is thought to
depend on sensorimotor systems integrating
internal motor and afferent representations
with top-down predictive and emotional signals. In contrast, a higher-level interaction of these
sensorimotor signals with information about context, general beliefs, and social norms is
assumed to establish a conceptual judgment of agency.
Predictive Coding and Free Energy Principle

Most recent computational modeling accounts of action and cognition go beyond the purely
efference-based mechanisms suggested in the comparator account. One theory of brain function
that is currently influential in cognitive neuroscience is the free energy principle associated
with the notion of predictive coding (Friston, 2005). Compared to the forward model view, this
theory suggests a more general and multilevel account of perceptual abilities beyond motor
Page 8 of 17
cognition. Accordingly, perceptual content is determined by top-down modulation of sensory

processing within a hierarchically organized architecture and reflects the attempt to minimize
the free energy induced by a stimulus at each processing level. It is thought that at each state of
the sensory processing hierarchy, the conditional probability of a stimulus is represented and
used for top-down predictions to the next lower level. These predictions are compared with the
sensory evidence, and mismatches between both are communicated via prediction error signals
from lower to higher levels.
A second important feature of this theoretical framework is the radical idea that an action just is
an attempt to minimize somatosensory prediction error. Although the theory was originally
developed to explain sensory perception, the notion of predictive coding has been extended to
describe motor behavior and sense of agency (Friston, 2012b). Perception is understood to be
the process of minimizing prediction error at all levels of the cortical hierarchy (cf., perceptual
inference), and action is considered a means to minimize prediction error by guiding sensory
sampling to fulfill prior expectations (cf., active inference, Friston, Daunizeau, Kilner, & Kiebel,
2010). This so-called active inference account of agency (see Figure 9.3B) equates sensory and
motor representations by relating them formally to the principle of minimization of prediction
error. Action representation and sense of agency are implicit (p.227) in the cycle of active
sampling of sensations (Friston, 2012a). That is, agency beliefs are thought to emerge from
successful sampling of sensory information, which minimizes uncertainty and prediction error,
and which increases confidence in predictions. This is assumed to occur at various levels of the
neurocognitive hierarchy, thereby challenging current frameworks that distinguish only between
two levels of agency registration (Synofzik, Vosgerau, & Newen, 2008).
In current applications of the model, it is argued that agency experience is based not only on
prediction and suppression of sensorimotor signals but also on interoceptive predictions about
the autonomic consequences of motor behavior (Seth, Suzuki, & Critchley, 2012). It is thought
that the interoceptive self in general, as a core sense of self-hood, can be based not only on
sensorimotor self-registration but also on predictive components involving, for example,
predictive interoceptive signals concerning autonomic regulation in a given state. In other
words, according to this view, both interoceptive and exteroceptive processes are thought to
inform parallel predictive coding schemes that are assumed to contribute to an integrated
representation of the self.
Deficient Prediction Error and Disturbed Agency

Existing predictive models of agency, as discussed above, aim at providing functional accounts
of disturbances of sensed agency, those characterized by delusional beliefs and failures in
inferencelike schizophreniain particular. Observations of a positive relation between deficits
in sensory attenuation and high-level misattribution of agency in schizophrenia have been taken
as evidence for correctness of these models. Abnormalities of active inference, as an embodied
form of predictive coding, could potentially account for abnormal repetitive behavior and
aberrant agency perception in obsessive-compulsive disorder.
Schizophrenia
Pathological agency experiences in schizophrenia patients have been investigated with respect
to the role of outcome predictions in the formation of agency experiences. In particular,
Page 9 of 17
symptoms such as delusions of control and auditory hallucinations have often been taken as a
paradigmadic case of agentive misattributions. Patients suffering from these symptoms feel that
their actions or thoughts are being controlled by an external force and not by them. This has
been quantified by studies using action recognition tasks, in which patients with delusions of
control show a difficulty in self-attribution of actions (Daprati et al., 1997; Franck et al.,
2001). As already noted when (p.228) discussing the comparator theory, the pathological
mechanism has been hypothesized to arise from a specific impairment in generating precise
predictions of the sensory effects of movements (Feinberg, 1978; Frith, 1992; Frith, Blakemore,
& Wolpert, 2000b). This may explain why these patients have been found to assign more weight
to external cues, with this weighting found to correlate with the strength of the patients
delusions of influence (Synofzik, Thier, Leube, Schlotterbeck, & Lindner, 2010).
Another line of research has examined the ability of schizophrenia patients to attenuate self-
induced sensory events. These studies revealed that patients suffering from agentive
misattributions do not exhibit the normal reduction of neuronal responses to self-generated, as
compared to externally generated, stimuli (Ford & Mathalon, 2004; Heinks-Maldonado,
Mathalon, Gray, & Ford, 2005). Other work involving perceptual decision tasks confirmed this
lack of self-attenuation in schizophrenia patients by showing that perceptual ratings were not
decreased for self-induced sensory stimulation (Blakemore, Smith, Steel, Johnstone, & Frith,
2000; Shergill, Samson, Bays, Frith, & Wolpert, 2005). The underlying pathophysiological
mechanism was hypothesized to involve dopamine-transmitted prediction errors, leading to a
disturbance in updating and generating precise predictions (Fletcher & Frith, 2009). In other
words, from a computational modeling perspective, predictive coding has been used to
formulate disturbances of agency and sensory attenuation deficits in schizophrenia as a
reduction in the precision that is ascribed to sensory predictions relative to sensory input (for a
recent review, see Adams, Stephan, Brown, Frith, & Friston, 2013).
Obsessive-Compulsive Disorder
A neglected disorder in the field of agency research is obsessive-compulsive disorder (OCD),
even though abnormalities in the awareness and control of motor actions are at the core of the
phenomenological expression of this disorder. OCD patients have a severely impaired agency
experience. This becomes particularly apparent in the presentation of not just right
phenomena or patients incompleteness experiences related to self-action. For example, an OCD
patient may be unable to achieve the feeling of having closed the door despite an effective
action and despite being aware of his exaggerated need for certainty. The experience of
incompleteness regarding actions and their outcomes has been identified as a core motivational
dimension underlying symmetry/ordering and checking subtypes of OCD (Ecker & Gnner,
2008; Pitman, 1987). Interestingly, OCD has been long neglected in the field of agency research
despite additional indicators for dysfunctional agency processes such as abnormalities in
memory for self-action (McNally & Kohlbeck, (p.229) 1993) or a lacking sense of self and
partial depersonalization experiences during action (Hoffmann & Hofmann, 2010).
Recent experimental approaches started to systematically address aberrant agency experience

and action-effect anticipation processes in OCD. A study by Belayachi and Van der Linden (2010)
tested whether prime-induced illusory agency experience for action outcomes can be induced in
Page 10 of 17
a sample of subjects with subclinical levels of OCD. Participants without checking symptoms
experienced illusory agency in conditions of congruent priming consistent with findings of prior
studies (Aarts, Custers, & Wegner, 2005; Wegner & Wheatley, 1999). In contrast, judgments of
individuals with checking symptoms remained unaffected by prime stimuli. This seems to
suggest that anticipatory information about an action outcome is not used for establishing a
subjective sense of agency in obsessive-compulsive checking. In order to test this hypothesis
more directly, we measured neuronal responses to sensory action outcomes in conditions of high
and low predictability of the outcome. In a clinical sample of OCD patients, we observed a lack
of predictive self-attenuation (Gentsch, Schutz-Bosbach, Endrass, & Kathmann, 2012).
Specifically, the typical reduction in N1 amplitude following self-generated sensory outcomes
was absent in these patients. This finding is in line with work indicating hypofunctioning of
sensory gating in OCD (Rossi et al., 2005) and extends this research by suggesting a central role
of internal forward models in the pathophysiology of aberrant agency feelings in OCD.
This research seems to suggest a shared deficit between OCD and schizophrenia in the top-down
modulation of basic sensorimotor processing during self-action. However, the extent to which
this is driven by the same or different underlying mechanisms, such as aberrant cue weighting
or prediction error signaling, remains to be addressed by future research. At the conceptual
level of belief formation, both disorders differ crucially, since delusional external misattribution
of agency to alien sources is present only in schizophrenia. In fact, OCD is better characterized
by internal misattributions of agency as reflected in inflated beliefs of personal responsibility,
which is considered a core motivational component underlying compulsive behaviors by
cognitive theories (Rachman, 2002). Together, research on psychiatric disorders indicates that
abnormal agency experience depends on the integrity of sensorimotor prediction processes.
However, the comparison of findings in OCD and schizophrenia patients suggests that at the
level of conceptual action representation, agency attribution is critically determined by the
cognitive system into which abnormal reafferent or prediction error signals are integrated (in
line with Synofzik, Vosgerau, & Newen, 2008).
With respect to the constellation of motor symptoms in OCD, repetitive behaviors in these
patients could be understood in the framework of active (p.230) inference: behavior emerges
through the active sampling of the environment that serves to change and optimize the precision
or uncertainty of sensory expectations. In the presence of impaired sensorimotor gating and
hypersensitivity to deficient prediction error signals, processes of updating and optimization
may remain incomplete. This may lead to performing an action repeatedly without being able to
achieve an inner sense of completion and satisfaction.
Summary and Conclusion: Toward a Multifactorial Model of Agency

One core aspect of the sense of agency refers to the human ability to experience self-causality,
not only for ones own bodily movements, but also for changes in the external world, including
interpersonal situations. In this chapter, we have reviewed evidence from behavioral,
neuroscientific, computational, and patient studies, including our own data, which together
point toward an important role of outcome prediction in the generation of subjective agency
experience. Research over the past decade has begun to investigate the nature and neural
implementation of the predictive coding architecture in the human brain. Outcome predictions
can be used for filtering sensory input through a continuous comparison with the actual sensory
Page 11 of 17
feedback. Therefore, the functioning of forward models has been explored by measuring
predictive inhibitory gating during sensory processing. Typically, self-generated sensory events
are suppressed, as compared to externally generated events, due to a precise cancellation of
afferent input by forward predictions. This mechanism of sensorimotor gating seemingly
underlies the capacity to accurately discriminate between self-produced and external
stimulation and thereby is thought to contribute to the sense of agency. Moreover, scientific
work has demonstrated that dysfunctional outcome prediction may contribute to aberrant
agency experience in schizophrenia and obsessive-compulsive disorder. Current computational
models explaining agency in terms of forward modeling and predictive coding have been
discussed in this chapter, and a critical distinction has been highlighted between neural and
cognitive mechanisms operating at different phenomenological levels of agency.
We now know that several factors affect neural activations in response to action outcomes,
including attention, emotion, and interoception; however, only recently has work begun to
explore the interplay between prediction mechanisms and other major determinants of
perception. One example for the impact of emotion and outcome valence on sense of agency is
the well-studied positivity bias (Miller & Ross, 1975), with people making more self-attributions
for favorable than unfavorable outcomes. By considering how different cognitive processes
interact with top-down predictions and influence the sensory (p.231) processing of action
outcomes, researchers will be able to refine current multicomponent frameworks of sense of
agency (Synofzik, Vosgerau, & Newen, 2008). This increasing knowledge of basic neurocognitive
mechanisms will eventually enable us also to develop and test more precise potential
neurocognitive endophenotype models of complex psychiatric disorders such as OCD and
schizophrenia, and will help to clarify both etiological understanding and diagnostic
classification.
References
enhancing experienced agency by priming effect information. Consciousness and Cognition,
14(3), 439458.
Adams, R. A., Stephan, K. E., Brown, H. R., Frith, C. D., & Friston, K. J. (2013). The
computational anatomy of psychosis. Frontiers in Psychiatry, 4, 47.
Belayachi, S., & Van der Linden, M. (2010). Feeling of doing in obsessive-compulsive checking.
Consciousness and Cognition, 19(2), 534546.
Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (1999). Spatio-temporal prediction modulates the
perception of self-produced stimuli. Journal of Cognitive Neuroscience, 11(5), 551559.
Blakemore, S. J., Frith, C. D., & Wolpert, D. M. (2001). The cerebellum is involved in predicting
the sensory consequences of action. Neuroreport, 12(9), 18791884.
Page 12 of 17
Blakemore, S. J., Smith, J., Steel, R., Johnstone, C. E., & Frith, C. D. (2000). The perception of
self-produced sensory stimuli in patients with auditory hallucinations and passivity experiences:
evidence for a breakdown in self-monitoring. Psychological Medicine, 30(5), 11311139.
Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self-produced
tickle sensation. Nature Neuroscience, 1(7), 635640.
Cardoso-Leite, P., Mamassian, P., Schtz-Bosbach, S., & Waszak, F. (2010). A new look at
sensory attenuation. Action-effect anticipation affects sensitivity, not response bias.
Curio, G., Neuloh, G., Numminen, J., Jousmaki, V., & Hari, R. (2000). Speaking modifies voice-
evoked activity in the human auditory cortex. Human Brain Mapping, 9(4), 183191.
Daprati, E., Franck, N., Georgieff, N., Proust, J., Pacherie, E., Dalery, J., et al. (1997). Looking
for the agent: an investigation into consciousness of action and self-consciousness in
schizophrenic patients. Cognition, 65(1), 7186.
Ecker, W., & Gnner, S. (2008). Incompleteness and harm avoidance in OCD symptom
dimensions. Behaviour Research and Therapy, 46, 895904.
the experience of agency: a positron emission tomography study. Neuroimage, 18(2), 324333.
Farrer, C., Franck, N., Paillard, J., & Jeannerod, M. (2003). The role of proprioception in action
recognition. Consciousness and Cognition, 12(4), 609619.
Farrer, C., & Frith, C. D. (2002). Experiencing oneself vs another person as being the cause of
an action: the neural correlates of the experience of agency. Neuroimage, 15(3), 596603.
Feinberg, I. (1978). Efference copy and corollary discharge: implications for thinking and its
disorders. Schizophrenia Bulletin, 4(4), 636640.
Fletcher, P. C., & Frith, C. D. (2009). Perceiving is believing: a Bayesian approach to explaining
the positive symptoms of schizophrenia. Nature Reviews Neuroscience, 10, 4858.
Ford, J. M., & Mathalon, D. H. (2004). Electrophysiological evidence of corollary discharge

dysfunction in schizophrenia during talking and thinking. Journal of Psychiatric Research, 38(1),
3746.
Franck, N., Farrer, C., Georgieff, N., Marie-Cardine, M., Dalery, J., dAmato, T., et al. (2001).
Defective recognition of ones own actions in patients with schizophrenia. American Journal of
Psychiatry, 158(3), 454459.
Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal

Society of London Series B Biological Sciences, 360(1456), 815836.
Page 13 of 17
Friston, K. (2012a). Embodied inference and spatial cognition. Cognitive Processing, 13(Suppl
1), S171177.
Friston, K. (2012b). Prediction, perception and agency. International Journal of

Psychophysiology, 83(2), 248252.
Friston, K., Daunizeau, J., Kilner, J., & Kiebel, S. J. (2010). Action and behavior: a free-energy
formulation. Biological Cybernetics, 102(3), 227260.
Frith, C. D. (1992). The cognitive neuropsychology of schizophrenia. Hillsdale, NJ: Erlbaum

Associates.
Frith, C. D., Blakemore, S. J., & Wolpert, D. M. (2000a). Abnormalities in the awareness and
control of action. Philosophical Transactions of the Royal Society of London Series B Biological
Sciences, 355(1404), 17711788.
Frith, C. D., Blakemore, S. J., & Wolpert, D. M. (2000b). Explaining the symptoms of
schizophrenia: abnormalities in the awareness of action. Brain Research Reviews, 31(23), 357
363.
Gallagher, I. I. (2000). Philosophical conceptions of the self: implications for cognitive science.
Gentsch, A., & Schtz-Bosbach, S. (2011). I did it: unconscious expectation of sensory
consequences modulates the experience of self-agency and its functional signature. Journal of
Cognitive Neuroscience, 23(12), 38173828.
Gentsch, A., Schutz-Bosbach, S., Endrass, T., & Kathmann, N. (2012). Dysfunctional forward
model mechanisms and aberrant sense of agency in obsessive-compulsive disorder. Biological
Psychiatry, 71(7), 652659.
Haggard, P., & Whitford, B. (2004). Supplementary motor area provides an efferent signal for
sensory suppression. Cognitive Brain Research, 19(1), 5258.
Heinks-Maldonado, T. H., Mathalon, D. H., Gray, M., & Ford, J. M. (2005). Fine-tuning of
auditory cortex during speech production. Psychophysiology, 42(2), 180190.
Helmholtz, H. (1866). Handbuch der Physiologischen Optik. Leipzig: Voss.
Hesse, M. D., Nishitani, N., Fink, G. R., Jousmaki, V., & Hari, R. (2009). Attenuation of
somatosensory responses to self-produced tactile stimulation. Cerebral Cortex, 20(2), 425432.
Hoffmann, N., & Hofmann, B. (2010). Zwanghafte Persnlichkeitsstrung und

Zwangserkrankungen. Therapie und Selbsthilfe (1st ed.). Berlin: Springer.
Hommel, B., Musseler, J., Aschersleben, G., & Prinz, W. (2001). The Theory of Event Coding
(TEC): a framework for perception and action planning. Behavioral and Brain Sciences, 24(5),
849878; discussion 878937.
Page 14 of 17
Houde, J. F., Nagarajan, S. S., Sekihara, K., & Merzenich, M. M. (2002). Modulation of the
auditory cortex during speech: an MEG study. Journal of Cognitive Neuroscience, 14(8), 1125
1138.
prediction. Psychological Bulletin, 139(1), 133151.
James, W. (1890). The principles of psychology. Cambridge, MA: Harvard University Press.
(Original work published 1890).
Linser, K., & Goschke, T. (2007). Unconscious modulation of the conscious experience of
voluntary control. Cognition, 104(3), 459475.
Martikainen, M. H., Kaneko, K., & Hari, R. (2005). Suppressed responses to self-triggered
sounds in the human auditory cortex. Cerebral Cortex, 15(3), 299302.
McNally, R. J., & Kohlbeck, P. A. (1993). Reality monitoring in obsessive-compulsive disorder.

Behaviour Research and Therapy, 31(3), 249253.
Miller, D. T., & Ross, M. (1975). Self-serving biases in the attribution of causality: fact or fiction?
Pitman, R. K. (1987). A cybernetic model of obsessive-compulsive psychopathology.

Comprehensive Psychiatry, 28(4), 334343.
Prinz, W. (1997). Perception and action planning. European Journal of Cognitive Psychology,
9(2), 129154.
Rachman, S. (2002). A cognitive theory of compulsive checking. Behavioural Research and

Therapy, 40, 625639.
Rochat, P., & Morgan, R. (1998). Two functional orientations of self-exploration in infancy.
British Journal of Developmental Psychology, 16, 139154.
Rossi, S., Bartalini, S., Ulivelli, M., Mantovani, A., Di Muro, A., Goracci, A., et al. (2005).
Hypofunctioning of sensory gating mechanisms in patients with obsessive-compulsive disorder.
Biological Psychiatry, 57(1), 1620.
Sato, A. (2008). Action observation modulates auditory perception of the consequence of others
actions. Consciousness and Cognition, 17(4), 12191227.
effect contribute to explicit judgment of agency. Cognition, 110(1), 7483.
Page 15 of 17
Sato, A., & Yasuda, A. (2005). Illusion of sense of self-agency: discrepancy between the
predicted and actual sensory consequences of actions modulates the sense of self-agency, but
not the sense of self-ownership. Cognition, 94(3), 241255.
Schafer, E. W., & Marcus, M. M. (1973). Self-stimulation alters human sensory brain responses.
Science, 181(95), 175177.
Seth, A. K., Suzuki, K., & Critchley, H. D. (2012). An interoceptive predictive coding model of
conscious presence. Frontiers in Psychology, 2, 395.
Shergill, S. S., Bays, P. M., Frith, C. D., & Wolpert, D. M. (2003). Two eyes for an eye: the
neuroscience of force escalation. Science, 301(5630), 187.
Shergill, S. S., Samson, G., Bays, P. M., Frith, C. D., & Wolpert, D. M. (2005). Evidence for
sensory prediction deficits in schizophrenia. American Journal of Psychiatry, 162(12), 2384
2386.
Sirigu, A., Daprati, E., Pradat-Diehl, P, Frank, N., Jeannerod, M. (1999). Perception of self-
generated movement following left parietal lesion. Brain, 122, 18671874.
Spence, S. A., Brooks, D. J., Hirsch, S. R., et al. (1997). A PET study of voluntary movement in
schizophrenic patients experiencing passivity phenomena (delusions of alien control). Brain, 120,
19972013.
ones actions. Brain, 133(Pt 1), 262271.
Synofzik, M., Vosgerau, G., & Lindner, A. (2009). Me or not mean optimal integration of
agency cues? Consciousness and Cognition, 18(4), 10651068.
von Holst, E., & Mittelstaedt, H. (1950). Das Reafferenzprinzip. Naturwissenschaften, 37, 464
476.
Voss, M., Bays, P. M., Rothwell, J. C., & Wolpert, D. M. (2007). An improvement in perception of
self-generated tactile stimuli following theta-burst stimulation of primary motor cortex.
Neuropsychologia, 45(12), 27122717.
Voss, M., Ingram, J. N., Haggard, P., & Wolpert, D. M. (2006). Sensorimotor attenuation by
central motor command signals in the absence of movement. Nature Neuroscience, 9(1), 2627.
Cognitive Sciences, 7(2), 6569.
Page 16 of 17
Wegner, D. M., & Wheatley, T. (1999). Apparent mental causation. Sources of the experience of
will. American Psychology, 54(7), 480492.
Weiskrantz, L., Elliott, J., & Darlington, C. (1971). Preliminary observations on tickling oneself.
Nature, 230(5296), 598599.
Weiss, C., Herwig, A., & Schtz-Bosbach, S. (2011). The self in action effects: selective
attenuation of self-generated sounds. Cognition, 121(2), 207218.
Weiss, C., & Schtz-Bosbach, S. (2012). Vicarious action preparation does not result in sensory
attenuation of auditory action effects. Consciousness and Cognition, 21(4), 16541661.
integration. Science, 269(5232), 18801882.
Page 17 of 17
The Relations Between Agency and Body Ownership

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Additive or Independent?
Manos Tsakiris
DOI:10.1093/acprof:oso/9780190267278.003.0010

The exact relation between the sense that ones body is ones own (body ownership) and the
sense that one controls ones body (agency) has been the focus of much speculation, but remains
unclear. A series of behavioral experiments has suggested that control over bodily movements
and ones sense of control over them produce a more global awareness of ones body in space
and time, lending support to an additive model whereby agency and body ownership are
strongly related; the ability to control actions is a powerful cue to body ownership. However,
functional neuroimaging findings support an alternative independence model whereby agency
and body ownership are qualitatively different experiences, triggered by different inputs, and
recruiting distinct brain networks. The apparently inconsistent results between psychophysical
and neuroimaging experiments are interpreted in the context of the recent understanding of
neuropsychological syndromes that affect action- and/or body-awareness.
Keywords: agency, body ownership, action, control, neuroimaging
Introduction: The Sense of the BodyBody Ownership and/or Agency?

Embodied cognition operationalizes the self as a physical entity characterized by the processing
of multisensory and motor signals that generate the experiences of ownership over ones body
and agency over ones actions. Recent approaches stemming from this emphasis on
sensorimotor processing and its relevance for higher cognition have attempted to explain these
two basic senses of ones body: the sense of agency and the sense of body ownership. Body
ownership refers to the special perceptual status of ones own body, which makes bodily
Page 1 of 13
sensations seem unique to oneselfthat is, the feeling that my body belongs to me, and is ever
present in my mental life (Gallagher, 2000). Agency refers to a persons ability to control his or
her actions, and through them, events in the external world. Sense of agency refers to the
experience of being in control of ones own actions. We experience agency throughout our
waking lives to the extent that we control the movements of our body in walking, talking, and
other voluntary actions, and we also feel and know that we control them.
A phenomenological analysis of bodily experiences suggests that body ownership and agency
reflect two distinct ways in which we experience our (p.236) body. For example, in the case of
involuntary movement, I experience that I am moving, and therefore that it is my hand moving. I
thus have a sense of ownership for the movement and the body part that is moved passively. At
the same time, I normally do not have a sense of agency for such movement (since it is not I who
caused it). Thus, we experience body ownership not only during voluntary actions, but also
during passive movement and at rest (see also Longo & Haggard, 2009; van den Bos &
Jeannerod, 2002). In contrast, only voluntary actions should produce a sense of agency. Several
studies confirm that agency is closely linked to the generation of efferent motor signals and the
monitoring of their effects (e.g., Blakemore, Wolpert, & Frith, 2002). In contrast, the sense of
body ownership can be induced only by afferent sensory signals (Botvinick & Cohen, 1998).
Evidence from neuropsychological syndromes also supports the distinction between these two
senses. Somatoparaphrenia is a neurological condition, which is usually related to anosognosia
for hemiplegia (AHP) and occurs after predominantly right hemispheric lesions. Patients with
somatoparaphrenia believe that their limbs contralateral to the side of the lesion belong to
someone else, and the disorder is often accompanied by the inability to feel tactile sensations in
the non-belonging part of the body (for a review, see Vallar & Ronchi, 2009). In contrast, other
syndromes suggest that one can have an abnormal sense of agency without loss of body
ownership. For example, schizophrenic patients with delusions of control demonstrate a striking
failure to experience their own agency over the actions that they execute. The delusion of
control is an example of a passivity experience in which a patient feels that his own actions are
being created, not by himself, but by an outside force, usually an external agent (for a review,
see Blakemore, Wolpert, & Frith, 2002, and Chapter 16 by Fletcher & Fotopoulou in this
volume). The main feature of this symptom is that the intention to act is misattributed to another
agent, whereas the ownership of the body part that executes the action is not. Though intuitively
the two senses of body ownership and agency seem to jointly constitute a coherent integrated
body awareness, their exact relation remains unknown. In this chapter, we review behavioral
and functional neuroimaging experiments that have tried to elucidate the relation between these
two fundamental senses of our body.
Modeling the Relation Between Body Ownership and Agency

On one view, the relation between agency and body ownership is additive, meaning that agency
entails body ownership (Tsakiris & Haggard, 2005). This view follows from the observation that
one can control the movements of ones (p.237) own body, but not other objects, at will
(Descartes, [1637] 2006). Thus, agency offers a strong cue to body ownership that might also be
used to produce a more integrated coherent body ownership. On this view, the sense of agency
should involve the sense of body ownership, plus a possible additional experience of voluntary
control. According to the additive model, agency entails body ownership. As will be shown
below, support from this model has come mainly from behavioral studies that highlighted the
Page 2 of 13
effects of agency on body awareness over and above the sense of body ownership that we have
during non-agentic, such as passive, movement. In terms of the underpinning neural
mechanisms that would account for such additive relationship, this model predicts some
common activations in agency and ownership conditions, and precludes the prediction that of
body-ownership-specific activations that are not also present in conditions that produce agency.
An alternative independence model suggests that sense of agency and sense of body
ownership rely on qualitatively different brain mechanisms, without any common components.
Such a model would be consistent with the literature on tool use, whereby we experience control
of extra-corporeal objects without necessarily experiencing ownership over them. But would that
be true for ones body? In other words, would the agency over ones body be independent of its
ownership? According to the independence model, the brain could have distinct networks for
sense of body ownership and sense of agency, such that we would not expect common
activations between conditions that produce agency and ownership, and would also predict
agency- and ownership-specific brain areas.
The Additive Model and Supporting Behavioral Evidence

Given the operationalization of agency and body ownership presented earlier (see also Tsakiris
& Haggard, 2005), the study of their relation should involve the systematic manipulation of the
own-action component itself, by comparing voluntary action and passive movement conditions.
We assume that a feeling of agency, that is, the feeling of moving my body voluntarily (as
opposed to the judgment of agency; see Synofzik et al., 2008, and Chapter 13 by Synofzik in this
volume) cannot be experimentally manipulated in a consistent way unless the action component
itself is systematically manipulated. A voluntary movement and a passive displacement applied
externally may be physically identical, but are psychologically different: the voluntary movement
supports a sense of agency, while the passive movement does not (Gallagher, 2000). Implicit in
this experimental approach to agency is the assumption that a sense of body ownership (i.e., the
sense that this is my body, independent of (p.238) whether it is moving or not) is present both
during active and passive movement. What therefore distinguishes the two conditions is the
critical addition of agency: only during an active voluntary movement do I have a sense of
agency, and body ownership, over my moving hand, whereas during a passive movement or a
purely sensory situation (e.g., see the Rubber Hand Illusion; Botvinick & Cohen, 1998), I have
only a sense of body ownership (e.g., that my hand is moving or I experience touch on my hand).
This approach recalls Wittgensteins (1953) question, What is left over if I subtract the fact that
my arm goes up from the fact that I raise my arm? Recent experimental studies have inverted
the philosophical question, asking, What is added when I raise my arm over and above the fact
that my arm goes up? This view treats agency as an addition to the somatic experience of
ownership.
Studies of this kind have manipulated the intention/preparation stage of the motor sequence.
However, since the experience of intention itself is thin and elusive, most studies have measured
the experience of later stages, such as body movement and its external effects. Their aim was to
understand how agentic actions (i.e., intentions) structure the perception of events that relate to
ones own moving body and/or the effects of such movements in the external world. The next
step was to use this indirect or implicit evidence to inform psychological theories about agency
and its relation to body ownership.
Page 3 of 13
Such an approach has been adopted in recent studies that focus on time awareness,
somatosensory perception, and proprioceptive awareness during voluntary action. Importantly,
a significant methodological advantage of studying these domains is that one can directly
compare how the agentic nature of movement affects these three domains over and above the
mere presence of movement cues; that is, one can directly compare voluntary to passive
movements.
Consistent results have shown how the presence of agency changes the experience of the body
and the outside world, measured using dependent variables such as temporal awareness and
spatial representation of the body. They thus provide indirect or implicit evidence about agency.
Three fundamental and robust outcomes of agency volition emerge: a temporal attraction effect,
a sensory attenuation effect, and a change in the spatial representation of the body itself.
Agency and Temporal Attraction Effects

Action fundamentally changes the experience of time. Both actions and their effects occur at
specific measurable points in time, making correlation between subjective and objective time
possible. Therefore, time perception has been one of the most important indirect methods for
studying agency. In one approach, participants are asked to judge the perceived onset of
voluntary actions, and (p.239) of a sensory event (a tone) occurring shortly afterward. The
perceived time of the action was shifted later in time, toward the ensuing tone, compared with a
baseline condition where no tone occurred. The perceived time of tones, in contrast, was shifted
earlier in time, back toward the action that caused them, relative to a baseline condition in
which no action was made. This intentional binding effect (Haggard et al., 2002) suggests that
the experience of agency reflects a specific cognitive function that links actions and effects
across time, producing a temporal attraction between them. Crucially, no such effects were
found when passive, involuntary movements were applied, suggesting that intentional binding is
a specific marker of the sense of agency.
Agency and Sensory Attenuation Effects

Sensory stimulation of equal magnitude is perceived as being less intense when it is self-
generated than when it is externally or passively generated (see Gentsch & Schtz-Bosbach,
Chapter 9 of this volume). This phenomenon of sensory attenuation is a robust feature of
voluntary motor control. Computational models suggest that it reduces the possibility of
computational overload by reafferent signals reflecting self-generated outcomes. Since the
sensory consequences of such actions can be predicted internally, there is no need to sense
them, and they are accordingly attenuated. This prediction is thought to involve efference copies
of the motor command, and internal neural models of the motor system (Frith, Blakemore, &
Wolpert, 2000). This concept has been extended from computational motor control to the
experience of agency. On this view, the experience of ones own actions depends on the outcome
of the comparison between the predicted and the actual state of our bodies. Sensory stimulation
generated by ones voluntary actions is predicted and attenuated. Therefore, when there is little
or no discrepancy between the predicted and actual state of the body, an individual can be
reassured that he or she was the agent. This approach can correctly discriminate between
internally generated and external sensory events, and can therefore ascribe agency. However,
since it suppresses perception of self-generated information, it cannot explain why there is a
positive experience of agency at all. Models based on attenuation treat agency as the absence of
Page 4 of 13
exteroceptive perceptual experience, not as a positive experience in itself. However, the

phenomenon of sensory attenuation may be a reliable functional signature of agency, which can
be used as an implicit measure in experimental studies.
Agency and Spatial Body-Representation Effects

We previously defined agency as an addition to normal experience of the body. Recent evidence
suggests that agency transforms the experience of the (p.240) body, as well as adding to it. A
number of studies have compared the effects of voluntary action and passive movement on
proprioceptive awareness of ones body. Agency generally enhances both spatial and temporal
(Tsakiris et al., 2005) processing of proprioceptive information. Tsakiris, Prabhu, and Haggard
(2006) used the Rubber Hand Illusion (RHI; Botvinick & Cohen, 1998) to show that voluntary
actions produce a more coherent and global proprioceptive representation of the body than do
passive movements. In the RHI, synchronous stimulation of both a rubber hand, or a video
image of the hand, and the participants unseen hand produces a strong illusion that the rubber
hand is part of ones own body. A reliable behavioral proxy of the illusion is a shift in the
perceived location of the participants hand toward the rubber hand. When the stimulation
involved passively displacing the participants hand, and monitoring the movement via a video
image of the hand, the effect was confined to the individual finger that was passively displaced.
In contrast, when the participant actively moved the same finger, the proprioceptive drift that
measures the illusion transferred to other fingers also. Voluntary action appeared to integrate
distinct body parts into a coherent, unified awareness of the body, while equivalent passive
stimulation produced local and fragmented effects on proprioceptive awareness.
Taken together, the review of the behavioral studies above suggests that the unity of bodily self-
consciousness may be an important additive product of agency given to the omnipresent sense
of body ownership.
The Search for the Neural Correlates of Agency: Does the Neural Evidence Support
the Independence Model?
The framework of comparing active to passive movements to study agency implies that an
experience, or the fact of agency, is added to the normally continuous and omnipresent sense of
body ownership. Previous accounts based on behavioral (see above) and introspective evidence
favor the additive model, since they identify a common sense of body ownership, plus an
additional component unique to action control (Longo & Haggard, 2009). Behavioral and
neuroimaging studies have also focused on the neurocognitive processes that underpin body
ownership and agency (Ehrsson, Spence, & Passingham 2004; Farrer & Frith, 2002; Farrer et
al., 2003; Fink et al., 1999; Tsakiris et al., 2007), but the exact neural bases of these two aspects
of body awareness remain unclear. For example, neuroimaging studies that investigated the
sense of body ownership using the RHI (see Botvinick & Cohen, 1998) report that activations in
the bilateral premotor cortex and the right posterior insula are associated with the illusion of
ownership of the rubber hand, (p.241) and are present only when visual and tactile
stimulations are synchronized (Ehrsson et al., 2004; Tsakiris et al., 2007). Studies investigating
the neural signatures of the sense of agency have used similar methods, such as the systematic
manipulation of visual feedback to alter the experience of ones body in action. Activity in the
right posterior insula was correlated with the degree of match between the performed and
viewed movement, and thus with judgments of self-attribution (Farrer et al., 2003). Conversely,
Page 5 of 13
activity in the right dorsolateral prefrontal cortex (Fink et al., 1999; Leube et al., 2003), right
inferior parietal lobe, and temporo-parietal junction (Farrer et al., 2003, 2008) was associated
with the degree of disparity between the performed and the viewed movement, and thus with
actions not attributed to the self, according to the participants judgments.
These studies were largely based on manipulating visual feedback to either match or mismatch
the participants manual action, similar to the behavioral experiments on agency described
earlier. However, such manipulations cannot separate the contributions of efferent and afferent
signals that are both inevitably present in manual action. These imaging data may therefore
confound the neural correlates of agency and body ownership. For example, with undistorted
visual feedback of an action, there is a three-way match between efferent motor commands,
afferent proprioceptive signals, and vision. Thus, any effects seen in such conditions could be
due to congruence between (a) efferent and proprioceptive signals, (b) efferent signals and
visual feedback, (c) proprioceptive signals and visual feedback, or (d) some complex interaction
of all three signals. Conversely, when visual feedback is distorted (spatially or temporally), there
is sensorimotor conflict between efferent signals and vision, but also inter-sensory conflict
between proprioceptive and vision. As a result, any differences between match and mismatch
conditions could reflect sensorimotor comparisons (relating to sense of agency) or
proprioceptive-visual comparisons (relating to sense of body ownership). As a result, such
experimental designs cannot distinguish between the additive and the independence model of
agency and body ownership.
However, as suggested earlier, the senses of agency and body ownership can be disentangled
experimentally, by comparing voluntary action with passive movement, as shown above.
Tsakiris, Longo, and Haggard (2010) implemented this experimental design in a neuroimaging
study to disentangle the neural basis of the relation between the sense of body ownership and
agency using functional magnetic resonance imaging (fMRI). Body ownership was manipulated
by presenting real-time or delayed visual feedback of movements, and agency, by comparing
voluntary and passive movements. Synchronous visual feedback causes body parts and bodily
events to be attributed to ones own self (Longo & Haggard, 2009). The experiment aimed at
testing the additive and the (p.242) independence models of the agency and body-ownership
relations. The first, additive model, holds that agency entails body ownership. On this view,
active movements of the body should produce both a sense of body ownership and a sense of
agency. The feeling of being in control of a body should involve the sense of that bodys
ownership. This produces three concrete predictions about brain activations in agency and
ownership conditions. First, there should be some activations common to agency and body-
ownership conditions. Second, there should be an additional activation in agency, which is
absent from body ownership. Third, there should be no activation in the body-ownership
condition that is not also present in the agency. A second model, the independence model, holds
that sense of agency and sense of body ownership are qualitatively different experiences,
without any common component. On this view, the brain could contain distinct networks for
sense of body ownership and sense of agency. The independence model produces three concrete
predictions. First, there should be no common activations between agency and ownership.
Second, there should be a specific activation in agency conditions that is absent from ownership.
Third, there should be a specific activation in ownership that is absent from agency. In addition
to the collection and analysis of fMRI data, participants were asked to answer a series of
Page 6 of 13
questions referring to their experience of agency and/or body ownership during the various
experimental conditions.
Overall, the introspective evidence from Tsakiris, Longo, and Haggard (2010) broadly supported
the additive model of agency. According to the additive model, a similar sense of body
ownership would be present both for active and passive movement conditions with synchronous
visual feedback, but the sense of agency would additionally be present following voluntary
movements. Indeed, participants reported significantly more agreement with questionnaire
items reflecting agency in the active/synchronous condition compared to the other three
conditions. In particular, body-ownership questions were also more highly rated in the active/
synchronous condition as compared to the passive/synchronous condition, suggesting that
agency strengthens the experience of body ownership. In terms of expected brain activations, if
the addition of agency to body ownership enhances the same kind of experience, then we would
expect to find at least some shared activations between agency and body ownership. The other
(independent) hypothesis suggests that agency is not simply an addition to body ownership,
but a qualitatively different process. This independence model would predict different patterns
of brain activity in the two cases.
To distinguish between the neural predictions of the additive and independence models, the first
analysis focused on brain areas that are commonly activated by agency (induced via active
movement) and sensory-driven body (p.243) ownership (induced via passive movement). This
analysis revealed no suprathreshold activations common to the two conditions, inconsistent with
the additive model that predicted at least some common activations. A second hypothesis
derived from the additive models is that there should be no activations for body ownership that
are not also present for agency. However, both body ownership and agency were associated
with distinct and exclusive patterns of activation, providing direct evidence that their neural
substrates differ. In particular, agency was specifically associated with activations in the pre-
supplementary motor area, the superior parietal lobe, the extrastriate body area, and the dorsal
premotor cortex bilaterally (BA6). In relation to a purely sensory-driven body ownership,
suprathreshold activations were observed in a network of midline cortical structures, including
the precuneus, the superior frontal gyrus, and the posterior cingulate. Notably, these midline
cortical activations recall recent suggestions of a dedicated self-referential processing network
(Northoff & Bermpohl, 2004; Northoff et al., 2006) in the default mode network (Gusnard et al.,
2001; Schneider et al., 2008).
Thus, neuroimaging data supported an independence model, while questionnaire data supported
an additive model. This somewhat surprising inconsistency may be explained in at least two
distinct ways. First, the questionnaire data may reflect a limitation of the folk psychological
concepts used to describe our embodied experience during sensation and movement. Folk
psychology suggests that agency is a very strong cue for ownership, so that I experience
ownership over more or less any events or objects that I control. However, the experience of
ownership of action during agency may represent a distinctive type of ownership that should not
be necessarily conflated with ownership of sensations or body parts.1 Second, the apparent
dissociation between neural activity and introspective reports may suggest that there is not a
one-to-one mapping between brain activity and conscious experience. Qualitatively similar
subjective experiences of ownership appear to be generated by quite different brain processes
Page 7 of 13
in the passive/synchronous and active/synchronous condition. Models involving a single neural

correlate of each specific consciousness experience have been highly successful in the study of
individual sensory percepts, particularly in vision (Haynes & Rees, 2006). However, the aspects
of self-consciousness that we call sense of body ownership and sense of agency are not unique
elemental percepts or qualia in the same way. Rather, they may be a cluster of subjective
experiences, feelings, and attitudes (Synofzik, Vosgerau, & Newen, 2008).
Suprathreshold activations unique to the experience of agency were observed in the pre-
supplementary motor area (pre-SMA), the superior parietal lobe, the extrastriate body area, and
the dorsal premotor cortex bilaterally (BA6). The pre-SMA is strongly involved in the voluntary
control of action (Goldberg, 1985). (p.244) Neurosurgical stimulation studies further suggest
that it contributes to the experience of volition itself: stimulation of pre-SMA can produce an
urge to move, at stimulation levels below threshold for evoking physical movement (Fried et
al., 1991; also see Chapter 4 by Gilron, Simon, & Mukamel in this volume). Voluntary action was
present in both the active/synchronous and the active/asynchronous conditions (Tsakiris, Longo,
& Haggard, 2010): these differed only in the timing of visual feedback, and the resulting sense
of agency. However, the pre-SMA activation was greater in the synchronous agency condition,
where visual feedback confirms that the observed movement is temporally related to the
voluntary motor command, suggesting that the pre-SMA plays an important role not only in
conscious intention (Lau, Rogers, Haggard, & Passingham, 2004), but also in the sense of
agency.
The observed premotor activation (BA6) is also of relevance to a different type of action-
awareness deficit. Anosognosia for hemiplegia (AHP; see also Chapter 17 by Berti, Garbarini, &
Pia in this volume) involves denial of motor deficits after right hemisphere stroke. It arises, in
part, by a failure to monitor signals related to ones own movement, and is associated with
lesions in right BA44 and BA6 (Berti et al., 2005), Interestingly, anosognosic patients seem to
ignore the conflict between their own intention to move, and the manifest lack of movement of
the left hand. They appear to perceive their intention, but not the failure of their intention to
trigger appropriate proprioceptive and visual feedback (Fotopoulou et al., 2008). These findings
are consistent, therefore, with an involvement of this area in the sense of agency, based on
conflicts between sensory and motor signals. The roles of pre-SMA and BA6 in this experiment
could either reflect an advance intention-based prediction of the sensory feedback of action, or a
delayed post-dictive attribution of sensory feedback to the self. However, recent experimental
studies on AHP patients have revealed that AHP deficits arise from an impairment of an advance
intention-based predictive model of agency. Fotopoulou et al. (2008) tested four hemiplegic
patients with AHP and four without anosognosia (non-AHP) who were provided with false visual
feedback of movement in their left paralyzed arm using a prosthetic rubber hand. This allowed
for realistic, three-dimensional visual feedback of movement, and deceived patients into
believing the rubber hand was their own. Crucially, in some conditions, visual feedback that was
incompatible with the patients intentions was given. For instance, in a critical condition,
patients were instructed to move their left hand, but the prosthetic hand remained still. This
condition essentially mirrored the classic anosognosic scenario within an experimentally
controlled procedure. In this way the study was able to examine whether the ability to detect the
presence or absence of movement, based on visual evidence, varied according to whether the
patient had planned to move his or her limb or not. (p.245) The key measure of interest was the
Page 8 of 13
patients response to a movement detection question (i.e., Did your left hand move?), which
required a simple yes/no response. The results revealed a selective effect of motor intention in
patients with AHP; they were more likely than non-AHP controls to ignore the visual feedback of
a motionless hand and to claim that they had moved it when they had the intention to do so (self-
generated movement) than when they expected an experimenter to move their own hand
(externally generated movement), or when there was no expectation of movement. In other
words, patients with AHP only believed that they had moved their hand when they had intended
to move it themselves, while they were not impaired in admitting that the hand did not move
when they had expected someone else to move it. By contrast, the performance of non-AHP
patients was not influenced by these manipulations of intention, and they did not claim they
moved their hand when the hand remained still. These results confirm that AHP is influenced by
motor planning, and in particular that motor awareness in AHP derives from the processing of
motor intentions. A recent lesion-mapping study suggested that premotor areas BA6 and 44,
which are implicated in action monitoring, are the most frequently damaged areas in patients
with AHP (Berti et al., 2005). This finding may explain why these patients fail to register their
inability to move, but it does not address the functional mechanism that underpins their illusory
awareness of action per se. The experimental study of Fotopoulou et al. (2008) provided direct
evidence for the hypothesis that awareness of action is based on the stream of motor commands
and not on sensory inflow, supporting to a certain extent the independence hypothesis between
agency and ownership.
Conclusion
One important implication of the experiments described in this chapter is that the sense of
agency seems to be closely linked to the appropriate processing of efferent information within
the motor system. For example, the experiments on intentional binding and sensory attenuation
suggest that efferent signals are necessary for eliciting these effects, and support the
conceptualization of the sense of agency as an efferent-driven predictive process. Interestingly,
the processing of efferent information is important not only for the sense of agency, but also for
the coherent temporal spatial and sensory awareness of ones own body (see the section The
Additive Model and Supporting Behavioral Evidence), suggesting that the phenomenological
coherence of the bodily self is primarily linked to the neurophysiological processing of agentic
actions, rather than mere sensations. However, different neural networks appear to underlie our
experience of embodiment in sensation and in action, even though the experiences themselves
have common elements.
(p.246) To conclude, the present chapter has contrasted two alternative models of the relation
between body ownership and agency. While the analysis of behavioral results and introspective
reports lends support to the additive model, the analysis of the fMRI data shows support for the
independence model. Activity in premotor areas (pre-SMA and BA6) was associated with the
sense of agency, while activity in midline cortical structures was associated with a purely
sensory-driven sense of body ownership. In addition, the analysis showed no shared activations
between the two. This apparent dissociation between agency and body ownership is further
supported by the literature on the anarchic hand syndrome (Marcel, 2003). Such patients report
a lack of sense of agency over the anarchic hand, while they do retain a sense of body
ownership. Interestingly, the reverse dissociation, whereby people would experience agency, but
not body ownership, is harder to envisage. However, cases of patients with AHP who also
Page 9 of 13
display somatoparaphrenic delusions may represent one such case. When the examiner asks the
patient to look at her arm and report whose hand this is, the patient would say that this arm
belongs to someone else (Vallar & Ronchi, 2009). However, if the patient is asked whether she
can move her left arm, she would deny paralysis and report her ability to move voluntarily,
displaying a non-veridical awareness of her agency (Fotopoulou et al., 2008). Note, however that
AHP can also dissociate from somatoparaphrenia (Cutting, 1978). Interestingly, recent lesion-
mapping studies show that the critical lesion site for AHP is the right premotor cortex (BA6 and
BA44; see Berti et al., 2005), while the critical lesion sites for somatoparaphrenia symptoms may
involve the temporoparietal cortex and the posterior insula (Baier & Karnath, 2008; Vallar &
Ronchi, 2009). Studies of deafferentation also support the dissociation between agency and body
ownership. IWs performance in agency tasks (Haggard & Cole, 2007) suggests that his sense of
agency is normal, while his sense of his own body is clearly dramatically affected, as illustrated
both by his subjective reports (Cole, 1995) and by his difficulty in using his own somatosensory
experiences as a reference to understand others (Bosbach, Cole, Prinz, & Knoblich, 2005).
Taken together, the available evidence suggests a qualitative distinction between the brain
bases of sense of agency and sense of body ownership, consistent with the neuropsychological
literature. Different neural networks appear to underlie our experience of embodiment in
sensation and in action, even though the experiences themselves have common elements. There
are many cases in psychology where quite different mechanisms can be enlisted for a common
behavioral or perceptual goal: reading by words and reading by letters are the best-known
examples. Our findings suggest that the unified experience of ones own body may similarly
depend on two quite different neural mechanisms. How the various networks reported in the
literature interact to produce the unity of (p.247) bodily self-consciousness that characterizes
everyday life, and that appeared in our participants subjective reports, remains a key topic for
future research.
Acknowledgments
The European Platform for Life Sciences, Mind Sciences, and the Humanities grant by the
Volkswagen Stiftung for the Body-Project: interdisciplinary investigations on bodily
experiences.
Note
References
Baier, B., & Karnath, H. O. (2008). Tightlink between our sense of limb ownership and
selfawareness of actions. Stroke, 39, 486488.
Berti, A., Bottini, G., Gandola, M., Pia, L., Smania, N., Stracciari, A., Castiglioni, I., Vallar, G., &
Paulesu, E. (2005). Shared cortical anatomy for motor awareness and motor control. Science,
309(5733), 488491.
Page 10 of 13
Bosbach, S., Cole, J., Prinz, W., & Knoblich, G. (2005). Inferring anothers expectation from
action: The role of peripheral sensation. Nature Neuroscience, 8, 12951297.
Botvinick, M., & Cohen J. (1998). Rubber hands feel touch that eyes see. Nature, 391(6669),
756.
Cole JD (1995). Pride and a daily marathon. Cambridge, Massachusetts: MIT Press; originally
London: Duckworth.
Cutting, J. (1978). Study of anosognosia. Journal of Neurological Neurosurgican Psychiatry, 41,

54855.
Descartes, R. [1637] (2006). A discourse on the method (I. Maclean, Trans.). Oxford: Oxford
University Press.
Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). Thats my hand! Activity in premotor
cortex reflects feeling of ownership of a limb. Science, 305, 875877.
Farrer, C., Franck, N., Georgieff, N., Frith, C. D., Decety, J., & Jeannerod, M. (2003) Modulating
the experience of agency: a positron emission tomography study. Neuroimage, 18(2), 324333.
Farrer, C., Frey, S. H., Van Horn, J. D., Tunik, E., Turk, D., Inati, S., & Grafton, S. T. (2008). The
angular gyrus computes action awareness representations. Cerebral Cortex, 18, 254261.
Farrer, C., & Frith, C. D. (2002). Experiencing oneself vs another person as being the cause of
an action: the neural correlates of the experience of agency. Neuroimage, 15, 596603.
Fink, G. R., Marshall, J. C., Halligan, P. W., Frith, C. D., Driver, J., Frackowiak, R. S., & Dolan, R.
J. (1999). The neural consequences of conflict between intention and the senses. Brain, 122,
497512.
Fotopoulou, A., Tsakiris, M., Haggard, P., Vagopoulou, A., Rudd, A., & Kopelman, M. (2008). The
role of motor intention in motor awareness: an experimental study on anosognosia for
hemiplegia. Brain, 131, 34323442.
Fried, I., Katz, A., McCarthy, G., Sass, K. J., Williamson, P., & Spencer, D. D. (1991). Functional
organization of human supplementary motor cortex studies by electrical stimulation. Journal of
Frith, C. D., Blakemore, S. J., & Wolpert, D. M. (2000). Abnormalities in the awareness and
control of action. Philosophical Transactions Royal Society London Series B Biological Sciences,
355(1404), 17711788.
Gallagher, S. (2000). Philosophical conceptions of the self: implications for cognitive science.
Goldberg, G. (1985). Supplementary motor area structure and function: review and hypotheses.
Behavioural and Brain Sciences, 8, 567616.
Page 11 of 13
Gusnard, D. A., Akbudak, E., Shulman, G. L., & Raichle, M. E. (2001). Medial prefrontal cortex
and self-referential mental activity: relation to a default mode of brain function. Proceedings of
National Academy of Sciences, USA, 98, 42594264.
Haggard, P., & Cole, J. (2007). Intention, attention and the temporal experience of action.
Consciousness &Cognition, 16, 211220.
Haynes, J., & Rees, G. (2006). Decoding mental states from brain activity in humans. Nature
Reviews Neuroscience, 7, 523534.
Lau, H. C., Rogers, R. D., Haggard, P., & Passingham, R. E. (2004). Attention to Intention.
Science, 303, 12081210.
Leube, D. T., Knoblich, G., Erb, M., & Kircher, T. T. (2003). Observing ones hand become
anarchic: an fMRI study of action identification. Consciousness & Cognition, 12, 597608.
Longo, M. R., & Haggard, P. (2009). Sense of agency primes manual motor responses.
Perception, 38, 6978.
Marcel, A. J. (2003). The sense of agency: awareness and ownership of actions and intentions. In
J. Roessler & N. Eilan (Eds.), Agency and self-awareness (pp. 4893). Oxford: Oxford University
Press.
Northoff, G., & Bermpohl, F. (2004). Cortical midline structures and the self. Trends in Cognitive
Sciences, 8, 102107.
Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H., & Panksepp, J. (2006).
Self-referential processing in our braina meta-analysis of imaging studies on the self.
Neuroimage, 31, 440457.
Schneider, F., Bermpohl, F., Heinzel, A., Rotte, M., Walter, M., Tempelmann, C., Wiebking, C.,
Dobrowolny, H., Heinze, H. J., & Northoff, G. (2008). The resting brain and our self: self-
relatedness modulates resting state neural activity in cortical midline structures. Neuroscience,
157, 120131.
Synofzik, M., Vosgerau, G., & Newen, A. (2008) Beyond the comparator model: a multifactorial
two-step account of agency. Conscious and Cognition, 17(1), 219239.
Tsakiris, M., & Haggard, P. (2005). Experimenting with the acting self. Cognitive
Neuropsychology, 22, 387407.
Tsakiris, M., Haggard, P., Franck, N., Mainy, N., & Sirigu, A. (2005). A specific role for efferent
information in self-recognition. Cognition, 96(3), 215231.
Page 12 of 13
Tsakiris, M., Hesse, M., Boy, C., Haggard, P., & Fink, G. R. (2007). Neural correlates of body-
ownership: a sensory network for bodily self-consciousness. Cerebral Cortex, 17, 22352244.
Tsakiris, M., Longo, M. R., & Haggard, P. (2010). Having a body versus moving your body:
neural signatures of agency and body-ownership. Neuropsychologia, 48(9), 27402749.
Tsakiris, M., Prabhu, G., & Haggard, P. (2006). Having a body versus moving your body: how
agency structures body-ownership. Consciousness and Cognition, 15(2), 423432.
Vallar, G., & Ronchi, R. (2009). Somatoparaphrenia: a body delusion. A review of the
neuropsychological literature. Experimental Brain Research, 192, 533551.
van den Bos, E., & Jeannerod, M. (2002). Sense of body and sense of action both contribute to
self-recognition. Cognition, 85, 177187.
Wittgenstein, L. [1953] (1998). Philosophical investigations. London: Blackwell.
Notes:
(1) For example, Marcel distinguished between attributing an action to ones self and attributing
the intentional source of the action to ones self. Patients with anarchic hand have a clear sense
that their involuntary movements are their own, but they strongly deny intending them (Marcel,
2003). Since the patients often themselves report this dissociation as surprising, folk psychology
may not adequately capture the difference between ownership of intentional action and
ownership of bodily sensation.
Page 13 of 13
The Innate Experience of Self-Agency

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001

Philippe Rochat
DOI:10.1093/acprof:oso/9780190267278.003.0011

Research is presented showing that an implicit sense of self-agency is probably developing from
birth, and possibly even before that. These time points are months before the child begins to
manifest explicit (conceptual) self-knowledge, an event that is reported to occur typically in the
second year. Implicit self-knowledge in infancy, including the sense of self-agency, is rooted in
intermodal perception and action. This chapter proposes a particular role for embodied
synesthesia involving proprioceptionthe modality of self-perception par excellence. From
around 6 weeks postpartum, infants show signs of increasingly systematic and deliberate
exploration of their own body, and in particular the perceptual consequences of self-produced
actions. Infants have expectations regarding such consequences, getting much pleasure when
they are met, experiencing displeasure and frustration when they are not. The pleasures of
control associated with self-agency would form the affective core and driving motivational force
behind mental development from the outset.
Keywords: self-agency, self-knowledge, embodied synesthesia, proprioception, motivation, mental development
Introduction
There is deep, primordial pleasure in controlling and sensing the impact that one has on people
and things. The perception of self-agency and the sense of ones own body in relation to the
environment is what psychology is all about, what the life of the mind rests upon. In this chapter
I will propose that the experience of having control and associated pleasures are at the core of
psychic life from birth (and probably also in the womb during the last trimester of gestation).
Argumentum a contrario, I also want to insist that from the outset, a lack of perceived self-
agency and control can be the source of deep mental confusion, a hindering force of optimum
Page 1 of 13
(healthy) development. In general, I want to promote the idea, like many others long before me,
that the experience of having control is arguably what drives psychological development
(Watson, 1995) and that the lack of such control can deeply disrupt development and can be the
source of devastating helplessness (Seligman, 1975).
The general argument is that the experience of self-agency, a perceptual experience that
evolved to become associated with an innate mood-boosting and reinforcing affectivity (i.e.,
the pleasures of control), is the necessary prerequisite for the blossoming of consciousness in
child development. This innate affectivity is encapsulated in the reinforcing effects of connecting
actions (p.252) of the body and their effects on the environment, both social and physical.
Accordingly, the basic idea is that consciousness, in the broad sense of having knowledge and
being aware of such knowledge (i.e., not being asleep or comatose) is inseparable from the
capacity to perceive and predict consequences of ones own actions, and includes being
irresistibly drawn toward the highly reinforcing pleasures of enacting such capacity.
There are necessary pre-requisites for perceived self-agency, all expressed in early infancy. They
correspond to what can be seen as five developmental pillars of self-awareness and
consciousness in general. These pre-requisites are the following:
1. The capacity to have emotions and experience feelings, unlike robots, machines, or
other zombie-like entities. This is what defines a sentient creature in contrast to
thermostat-like machines, for example. Indeed, to perceive self-agency, one must be
endowed with the capacity to experience reinforcing pleasures by getting feedback
regarding the impact of ones own actions on things and people.
2. The ability to perceive ones own body as an active entity differentiated from other
entities in the world, not in confusion with them. This is the basic capacity for a
distinction between self and world, unlike some sort of inherent blooming, buzzing,
confusion (James, 1890).
3. The capacity to perceive ones own body as organized and coordinated, with parts that
are not experienced as moving independently but rather always in unison. That would
also include some sort of unified processing of simultaneous input from the various
perceptual systems.
4. The capacity to perceive the situation of ones own body in relation to other entities in
the world and toward which actions can be oriented.
5. The capacity to perceive ones body as substantial and occupying space, hence as
being potentially an obstacle and offering physical resistance to other entities in the
world, a source of force and physical impact.
Here I want to suggest that these necessary pre-requisites of self-agency exist from the outset of
child development, with manifestations even in utero. Perceived self-agency would be part of our
innate mental architecture and the necessary foundation of consciousness in development.
First Experiences of Feelings

Experiencing self-agency entails minimal phenomenal awareness, the necessary experience of
what it is like to have embodied effect and control over things. But when does it all begin? When
Page 2 of 13
can we safely ascribe affective mental (p.253) states (i.e., experiencing the feelings of pain and
pleasure) to the developing child, the newborn, and possibly even the fetus?
For a long time, infants were seen essentially as sensing but not feeling pain (Rochat, 2011;
Rochat, 2014). As a case in point, until fairly recently (the 1940s and 1950s), infants and young
children were routinely operated on without any anesthesia. Medical teams would paralyze
squirming infants by the injection of Curare or similar paralytic agents. Even today, local
anesthetics are not routine in painful procedures on newborns (heel prick and circumcision),
even in ultramodern, state-of-the-art maternity hospitals. There is a continuing overtone in the
adult mind, around the world, and particularly in certain cultures, that infants have either no
feelings, less feelings, or that feelings experienced at this early stage might not be as
consequential for lack of memory (infantile amnesia). Such intuitions, however, defy current
physiological as well as behavioral evidence.
What is typically recognized as the necessary neuroanatomical underpinning of experiencing

feelings like pain is already in place by the third trimester of gestation, even possibly much
earlier (8 weeks). Behaviorally, it is also by the third trimester of pregnancy that fetuses
demonstrate unmistakable facial expressions of pain or joy (smiling) with corresponding heart
rate and other expressions of overall behavioral states like wakefulness or sleep (Hata, Dai, &
Marumo, 2010; Hepper, 2002; Prechtl, 1977). In human ontogeny, it is thus reasonable to think
that by at least 31 weeks of gestation, almost 3 months before term birth and 6 months after
conception, subjectivity and a minimal ability to experience feelings are de facto in place. From
this point in prenatal development, and probably even prior, it is not farfetched to be ethically
concerned about the comfort and potential suffering of fetuses (see Rochat, 2011, 2014 for more
detailed discussion).
Recent progress in fetal psychology research suggests that there are prenatal signs of
experiencing feelings. The well-organized emotional expressions, combined with the remarkable
continuity of prenatal and postnatal development, supports the idea that first experiences of
feelings, and therefore the potential for minimal self-awareness, might emerge 810 weeks
before birth (3032 weeks gestational age). Keeping in mind the striking continuity of behaviors
observed during the last 10 weeks of gestation and what can be readily observed and tested
after birth (Prechtl, 1984), what can be seen in the newborn could stand also for what is not
readily testable in the womb, from at least 32 weeks, when fetal behaviors show all the aspects
of what is observed after birth.
Another demonstration that, at least from birth, infants feel and are not just sensing the world,
is the fact their behavior involves much more than simple automatic responses triggered by
particular stimulations. Infants manifest (p.254) much more than simple reflexes from birth.
Rather, they show complex systems of action oriented toward particular resources with affective
values that they perceive and feel, rather than just sense in some sort of on/off switch or trigger
mechanism. From birth, infants are actors and feelers, rather than responders or just sensors.
Newborns are best described as active explorers of a world that has values they feel and
experience in reference to particular mental states: satiety, hunger, pain, comfort, pleasure,
fear, surprise, or curiosity.
Page 3 of 13
Acting and Feeling by Neonates

Behavior at birth and in the first months of life can be divided in two basic kinds: automatic
reflex systems and action systems. Reflex systems correspond to what pediatricians probe
immediately after birth to assess the neurophysiological health of the infant, when they perform
typical neurobehavioral tests to see if there are any neurological red flags. Action systems, in
contrast, correspond to behavioral patterns present at birth that are much more than triggered
responses to specific stimulations. They correspond to actions that are oriented toward
functional goals and particular resources in the environment: food, protection, comfort, and so
on. Reflex systems entail mere sensing and responding, whereas action systems involve feeling
and meaningful perception. Both systems are expressed immediately after birth, yet they involve
drastically different capacities.
The beating of the heart, the movements of the lungs in breathing, the shaking of the whole
body under a cold spell, the knee jerk response, or the blinking of eyelids in response to an air
puff all belong to the first kind of triggered (sensed) bodily movements. They are automatic and
reflex responses of the organism. The control of such movements is endogenous and self-
contained. These movements consist in highly predictable stimulus-response loops or tight
circuits. They are in essence automatic, triggered by particular stimulations.
Following the physiologist Sherringtons first account of this kind of bodily movement, the
control is encapsulated and rigidly prescribed within the organism as reflex arcs. It typically
involves low-level, sub-cortical neural networks: surgically decorticated animals continue to
express such movements (Sherrington, 1906). What controls such movements are closed
feedback systems that are similar to thermostats controlling for constant temperature inside a
house. Those mechanisms are simple when considered in isolation, but they are complex when
considered in interaction with each other, each calibrated to respond to particular ranges of
stimulation from the environment that are internal as well as external to the body. Each system,
which controls a particular autonomic/reflex response, is also adapted to interact with myriad
other similar systems that, in concert, maintain the integrity of the organism (p.255) as a whole
living and adapting system. But they do so in a very mechanistic way, with no reference to any
sorts of mental states, controlled by closed-loop feedback systems (i.e., autonomic reflexes)
ensuring basic physiological functioning (like breathing, digesting, or swallowing). They keep
the individual organism alive, but such movements do not involve any perception, or any
particular higher-order treatment of basic physiological signals or sensations that are sensed
but not felt in reference to mental states. The infant regurgitating an object that obstructs the
wind pipe does it automatically, not feeling the danger but sensing the obstruction, in the same
way that a thermostat senses automatically the crossing of a temperature threshold, without
feeling temperature change. In a way, this kind of bodily movement is sensitive but
psychologically blind to the environment to which it responds. Such movements are triggered by
nonspecified circumstances.
Action systems, in contrast, correspond to bodily movements that are more than autonomic or
reflex responses. Also expressed from birth and prominent during the first 6 weeks of life, they
are distinct on two basic grounds. First, they are movement systems consisting of actions that
are oriented toward specific functional goals. These systems are by definition adapted to tap into
available resources that exist outside the individual organism, in the surrounding environment:
Page 4 of 13
food, surfaces, objects, or people. Second, these movements are organized into systems that are
flexible, capable of changing based on previous experiences, and adjusting to novel
circumstances. They allow room for learning, controlled by open-feedback loop systems. Action
systems do entail perception (feelings) and learning. Contrary to reflexes, they entail some
psychology and meaning making (perception), presumably reflecting higher-order cortical
involvement. In support of this second kind, numerous research studies indicate that infants at
birth show more than autonomic/reflex arcs.1
For a long time, newborn sucking, grasping, stepping, rooting, and head turning were merely
construed as reflexes or automatic responses triggered by nonspecific stimulations. Multiple
studies now exist showing in fact that such movements need to be construed as actions rather
than reflexesactions that are already oriented toward particular features and resources in the
environment, such as faces or objects with a certain shape, texture, consistency, or smell.
Newborns, immediately after birth, track with their eyes and even show attempts at reaching
toward objects that move close by in their field of view. More impressive is the fact that they do
so preferentially when the object consists, for example, of a schematic face-like display. Old
research, validated multiple times since then, indicates that newborns tend to track more
canonical face-like displays (two adjacent dots for eyes above vertically aligned two dots for
nose and mouth) compared to non-canonical face-like displays with the same features but
scrambled (Morton & Johnson, 1991). Similarly, newborns (p.256) tend to suck differentially on
pacifiers that are more or less mimicking the biological nipple of the mother. They suck less and
tend significantly to increase oral exploration as a function of the eccentricity of a pacifier
compared to the biological nipple in terms of texture and consistency (Rochat, 1983).
We found the same kind of results when recording newborns grasping of objects varying in
texture and consistency that are placed in one of their palms (Rochat, 1987). In more recent
years, researchers have even established that newborn infants are significantly more inclined to
orient their face toward gauze impregnated with their own mothers amniotic fluid or breast
milk as compared to gauze impregnated with the amniotic fluid or the breast milk of another
woman who just gave birth (Marlier, Schaal, & Soussignan, 1998). If newborns orient and root
to smells or face-like displays, if they suck and grasp at objects introduced in their mouth or in
their hands, they do so with discrimination and preference.
This kind of movement is not made of autonomic, reflex responses triggered by nonspecific
stimulation. It is under the control of previous experiences (learning) and is intrinsically
oriented toward particular environmental resources. It calls for some psychology that engages
more than sub-cortical structures. Even if we cannot qualify these movements as being
intentional (another loaded term), it is reasonable to qualify them, not as automatic response-
like reflexes, but rather as adaptive actions generated in relation to objects that infants
perceive, and hence feel as agents, rather than simply sense as a physiological machine. Such
experience of feeling is the personal possession of the child by which minimal, implicit self-
awareness is expressed, based on embodied perceptual competencies including innate
synesthesia.
Page 5 of 13
Newborns Objective Perception

Infancy research of the past four decades has changed our views on the starting state of mental
life, namely what is it like to be a newborn. Until then, developmental theorists tended to
endorse the view, in their own ways, of an initial state of confusion with the environment, the
famous initial blooming, buzzing, confusion proposed by William James (1890). Neonates were
essentially construed as stimulus-bound, sensing but not feeling the world, their behavior
primitively reduced to ready-made, evolved automatisms (reflexes or pulsions). The newborns
world was envisioned as not yet objectified or differentiated, the child born in a state of
confusion with the environment. At the beginning, subjectivity and objectivity were seen as
confounded and in need of progressive integration through experience (e.g., cognitive distancing
and construction in the case of Piaget (Piaget, [1936] 1952), ego development in the case of
Freuds pulsion theory (Freud, [1905] 2000).
(p.257) More recent research shows that, in fact, healthy newborns do perceive the world
objectively and are not in a state of subject-object confusion. From birth they express a
difference between what pertains to their own body and what pertains to the world out there.
Although babies are born with poor contrast sensitivity and grating acuity (Banks & Shannon,
1993), infancy researchers investigating newborn vision demonstrate that despite the obvious
developmental lag of the modality, active perceptual processing does take place at birth. For
example, and relevant to our discussion, by using habituation and novelty preference paradigms,
researchers have established that newborn infants, only a few hours old, when awake and alert,
perceive the real (distal) size of objects, not the varying (proximal) sizes projected onto the
retina. Newborns perceive size constancy of objects (Granrud, 1987), most likely via visuo-
proprioceptive convergence cues from both eyes as they line their gazes and focus on the distal
object (Kellman & Aterberry, 2006). In all, this kind of empirical evidence suggests that newborn
infants have feeling experience, and are not just limited to sensing what is recorded at the
proximal level of the receptors (i.e., the retina). Early perceptual competency of perceiving a
world that is distal and objectified in relation to the self forms the necessary core for the
perception of self-agency: the perception and control of self-produced action in relation to non-
self things in the world.
Perceived Self-Unity at Birth

Synesthesia corresponds to the spontaneous, implicit metaphorical experience of a sensation
or percept in one modality that is simultaneously experienced in another. For example, one
might experience the particular timbre or pitch of a sound with the vivid experience of a specific
color, the experience of time duration corresponding to the obligatory experience of a particular
spatial layout or form (Simner et al., 2006). Neuroscientists have now established the embodied
(neurobiological) reality of such synesthesic experiences that, according to existing surveys,
are part of the life of approximately 5% of all adults (Hubbard, Arman, Ramachandran &
Boynton, 2005; Spector & Maurer, 2009).
An intriguing idea is that adult cases of synesthesia might in fact be remnant and magnifying
cases of inter-sensory connections that are present at birth, pruned and somehow inhibited in
the course of typical perceptual development (Spector & Maurer, 2009). Accordingly, these
connections would be expressed in muted forms in all adults (but see Deroy & Spence, 2013,
Page 6 of 13
for a critique and alternative account). In this view, synesthesia could be the natural starting
state of all subjective sensory experience. We would indeed (p.258) start off with a conflation
of all sensory modalities, as suggested by William James in his statement about blooming,
buzzing, confusion. However, this experiential conflation (what James refers to as the pure
sensory experience of newborns), rather than being the symptom of a major incompetence, as it
has been taken by most infancy researchers over the past 30 years, would be rather the sign of a
competence to reinforce a basic, innate potential to experience self-unity.
An abundance of research shows that infants are born with the ready-made opportunity to link
experiences from the various sense modalities, experiences that co-occur and tend to be
qualitatively linked, corresponding to particular feeling tones and profiles. From the start,
intermodal systems might exist that allow these sensory experiences to coalesce into the
affective core of subjective experience that ultimately gives it values: values in rudimentary
polarized terms such as pleasure or displeasure, stress or calm, soothing or enhancing,
attunement or disharmony, bonding or estrangement. All these represent affective meanings
(good or bad feelings) that are at the core of what would be a unified subjective experience at
birth. But what kind of empirical evidence is there that supports the assertion of a rich primitive
sensory conflation, a conflation that would harmonize, rather than confuse, early experience?
In relation to synesthesia, there is an abundance of empirical evidence showing that infants from
birth are readily able to process information across sensory modalities. One-month-old infants
are reported to discriminate an object they see projected on a screen based on the previous
experience of an analogous object explored with their mouth only (i.e., a smooth spherical
pacifier or a bumpy spherical pacifier texture; Meltzoff & Borton, 1979). In another series of
highly controlled, careful psychophysical studies on newborns in the early 1980s, Lewkowicz
and Turkewitz (1980) demonstrated that neonates transfer learning from the auditory to the
visual modality. Following visual habituation to either a bright or a dimmed light, they
responded differently to corresponding soft or intense sounds in the auditory domain.
In support of such unitary or common functioning of the senses at the outset, an even older
neurobehavioral study by Wolff and collaborators (1974) showed that the tactile stimulation of
the newborns wrist evokes activation of the somatosensory cortex. Moreover, this activity is
significantly enhanced when the infant hears also a white noise. Such auditory-tactile interaction
is not found in adults, a phenomenon that appears to be specific to the perceptual experience of
newborns. As additional neuro-developmental evidence on an early unitary functioning of the
senses, Neville and collaborators show that if infants respond to spoken language with, as
expected, enhanced activity in the auditory cortex, unlike adults and children, they also respond
with enhanced activity in the visual cortex (Neville, 1995).
(p.259) In further support of the natural primacy of synesthetic experience, Mondloch and
Maurer (2004) show in a series of studies that 23-year-old children tend to be naturally inclined
to perceive the same pitch-lightness, color-letters, or sound-shape correspondences typically
expressed by synesthetic adults (but also, to some extent, by non-synesthetic adults). Young
toddlers, for example, perceive that a higher pitch sound goes with a brighter color; a nonsense
word made of rounded vowels goes with a jagged shape (e.g., te-ta-ke goes with a sharp edged
form, or that the letter A goes with the color red).
Page 7 of 13
In all, these findings, among many others, support the idea of a highly organized intermodal and
resonating embodied experience at birth. Early perceptual experience is made of rich sensory
correspondences and implicit a-modal representations that can be said to be metaphorical
because they transcend the particularities of the sense modalities as singular perceptual
systems. It is an experience that carries rich conflation and correspondences, not the cognitive
confusion that has been assumed by many infancy researchers, including myself, since Jamess
misconstrued blooming, buzzing confusion.
Perceived Self-Agency in Infancy

From birth, infants are capable of perceiving their own body as an entity among other entities
an entity that has unity, is differentiated, occupies space, and is substantial. In addition, from at
least 2 months of age, there is good evidence that infants have a sense of their own agency on
objects; they are aware of themselves as embodied agents in the world (Rochat, 2001).
For example, we were able to show that newborn infants do discriminate between self-
stimulation and stimulations coming from the outside world, suggesting that they are not in a
state of confusion with the world outside. They root (i.e., orient head and mouth) significantly
more toward the finger of an experimenter touching their cheek than their own hand
spontaneously brought in contact with the peri-oral region of the face (Rochat & Hespos, 1997).
We also showed that 2-month-olds are attentive and systematically explore the auditory
consequences of their own action while sucking on a sound-producing pacifier (Rochat &
Striano, 1999). They differentiate between sounds that are perfectly contingent but that are or
are not linked to the physical pressures they apply on the pacifier. In the context of our
research, from 2 months of age (though not at birth) infants show clear signs that they perceive
themselves as an agent of what they hear.
Other empirical observations demonstrate further the minimal subjectivity of neonates, who
seem to experience the world with an implicit differentiated sense of themselves as embodied
perceivers. For example, there is some (p.260) evidence that from birth, infants differentiate
movements of the own body (ego motions) from movements of objects and things in the world
that occur independently of the self (allo motions). Newborns pick up visual information that
specifies ego-motion or movements of their own body while they, in fact, remain stationary
(Jouen & Gapenne, 1995). This kind of observation points to the fact that from birth, infants are
endowed with the perceptual, qua inter-modal, capacity to pick up and process meaningfully
self-specifying information. It includes the early experience of a body that is substantial and
occupies space. For example, 2-month-old infants are documented showing protective and
avoidant behavior by raising their hands in front of their face as they perceive a solid object
(e.g., a ball) looming toward them (Ball & Tronick, 1971). Numerous studies show that from at
least 5 months of age, infants perceive their own embodied self-unity. For example, they react
and pay more attention to images of themselves that violate the calibrated sense of their own
body, the canonical way one limb typically looks and feels as it moves in relation of another (e.g.,
one leg configuration and movement in relation to the other). If this canonical calibration of the
body is optically tricked via inverted or delayed video feedback, the infant tends to show
surprise and explore the novelty of the embodied experience, moving and looking at their limbs
significantly more and longer (Bahrick & Watson, 1985; Rochat & Morgan, 1995; Schmuckler,
1996).
Page 8 of 13
Overall, all these observations suggest that from birth, infants possess minimal, implicit self-
awareness, hence subjectivity in terms of first-person perspective. Once again, we are not born
in a state of confusion with the environment, but rather possess a sense of ourselves as
differentiated and situated entities among other entities in the world. Newborns perceive
objects as distal and distinct from their own body, thus showing the potential for engaging in
object relations, and as a byproduct the capacity to exert control over them. It is thus reasonable
to posit that from the outset, all conditions are in place for the development of first feelings of
self-agency.
Innate Pleasures and Frustrations of Self-Agency

From an affective and motivational perspective, among the most robust findings of infancy
research of these past 40 years is the emotionally disruptive effect of an interrupted loop
between perception and action. This is true in the context of both physical and social
interactions. From the age of 23 months, infants are capable of instrumenting their body to
produce effects on objects or people: pulling a string attached to one of their wrists or ankles to
activate a music box or move a mobile above their crib (Watson, 1995; Rovee-Collier & Hayne,
2000). This kind of instrumental learning is (p.261) not only highly reinforcing for the young
infant but also a source of frustration and anger during an extinction phase. A series of infant
studies using a hand-pulling and audio-visual consequence paradigm document that from 2
months of age, children express interest and joy while learning about their own agency.
Inversely, they express a frustration-like emotion corresponding to anger based on the
Maximally Discriminative Facial Movement Coding System, or MAX (Alessandri, Sullivan &
Lewis, 1990; Sullivan & Lewis, 1989). Infants from the outset expect consequential events to
happen following self-generated actions. If these events do not follow, frustration and negative
affects emerge. This is particularly evident in the social realm when children begin to manifest
primary intersubjectivity (Trevarthen, 1980). This corresponds to the first explicit signs of a
shared experience in the context of dyadic face-to-face exchanges. By 6 weeks, infants manifest
socially elicited smiling in such context and enjoy the give-and-take pragmatics of proto-
conversation (Bruner, 1983), quickly expecting particular responses from engaging social
partners. If the latter become non-responsive by displaying a sudden still face (Tronick et al.,
1978) or show slight lack of timing in their affective bids and attunement, the infants social
engagement is shown to systematically drop. This is often associated with unmistakable signs of
displeasure like fussing and even crying. Such robust findings of displeasures linked to the
sudden violation of social as well as physical expectations are unmistakable signs of the innate
affectivity in terms of a basic polarized pleasure-displeasure dichotomy. This is revealing of the
underlying drive of what we can reasonably construe as innate signs of the human embodied
experience of self-agency.
It is from this experiential capacity and the dynamics of the affectivity associated with it that
children learn and grow consciousness of their relation to the world. As many pioneer
developmental psychologists proposed, the sense of self-agency is at the core of mental
development. It is, for example, the cornerstone of Jean Piagets seminal infancy works (Piaget,
[1936] 1952, [1938] 1955). The subjects sense of efficacy and self-generated exploration would
indeed be the main engine behind the ontogenetic growth of consciousness.
Note
Page 9 of 13
References
Alessandri, S. M., Sullivan, M. W., & Lewis, M. (1990). Violation of expectancy and frustration in
early infancy. Developmental Psychology, 26(5), 738744.
Bahrick, L. E., & Watson, J. S. (1985). Detection of intermodal proprioceptive-visual contingency

as a basis of self perception in infancy. Developmental Psychology, 21, 963973.
Ball, W., & Tronick, E. (1971). Infant Responses to Impending Collision: Optical and Real.
Science, 171, 818820.
Banks, M. S., & Shannon, E. S. (1993). Spatial and chromatic visual efficiency in human
neonates. In C. E. Granrud (Ed.), Carnegie-Mellon Symposium on cognitive psychology (pp. 1
46). Hillsdale, NJ: Erlbaum Associates.
Bruner, J. S. (1983). Childs talk. New York: W. W. Norton.
Deroy, O., & Spence, C. (2013). Why we are not all synesthetes (not even weakly so).
Psychonomic Bulletin and Review, 20(4), 643664.
Freud, S. ([1905] 2000). Three essays on the theory of sexuality. New York: Basic Books.
Granrud, C. E. (1987). Size constancy in newborn human infants. Investigative Ophtamology and
Visual Science, 28(Supplement), 5.
Hata, T., Dai, S. Y., & Marumo, G. (2010). Ultrasound for evaluation of fetal neurobehavioural
development: from 2-d to 4-d ultrasound. Infant and Child Development. Special Issue: Towards
a Fetal Psychology, 19(1), 99118.
Hepper, P. G. (2002). Prenatal development. In A. Slater & M. Lewis (Eds.), Introduction to

infant development. New York: Oxford University Press.
Hubbard, E. M., Arman, A. C., Ramachandran, V. S., & Boynton, G. M. (2005). Individual
differences among graphemecolor synesthetes: brainbehavior correlations. Neuron, 45, 975
985.
James, W. (1890). The principles of psychology. New York: Henry Holt.
Jouen, F., & Gapenne, O. (1995). Interactions between the vestibular and visual systems in the
neonate. In P. Rochat (Ed.), The self in infancy: theory and research (pp. 277302). Amsterdam:
North-Holland, Elsevier Publishers.
Kellman, P. J., & Arterberry, M. E. (2006). Infant visual perception. In D. Kuhn, R. S. Siegler, W.
Damon, & R. M. Lerner (Eds.), Handbook of child psychology: Vol. 2, Cognition, perception, and
language (6th ed., pp. 109160). Hoboken, NJ: John Wiley & Sons.
Page 10 of 13
Lewkowicz, D. J., & Turkewitz, G. (1980). Cross-modal equivalence in early infancy: auditory
visual intensity matching. Developmental Psychology, 16, 597607.
Marlier, L., Schaal, B., & Soussignan, R. (1998). Neonatal responsiveness to the odor of amniotic
and lacteal fluids: a test of perinatal chemosensory continuity. Child Development, 69(3), 611
623.
Meltzoff, A. N., & Borton, R. W. (1979). Intermodal matching by human neonates. Nature, 282,
403404.
Mondloch, C., & Maurer, D. (2004). Do small white balls squeak? Pitchobject correspondences
in young children. Cognitive, Affective, and Behavioral Neuroscience, 4, 133136.
Morton, J., & Johnson, M. H. (1991). CONSPEC and CONLERN: a two-process theory of infant
face recognition. Psychological Review, 98(2), 164181.
Neville, H. J. (1995). Developmental specificity in neurocognitive development in humans. In M.

Gazzaniga (Ed.), The cognitive neurosciences (pp. 219231). Cambridge MA: MIT Press.
Piaget, J. ([1936] 1952). The origins of intelligence in children. New York: International
Universities Press.
Piaget, J. ([1938] 1955). The construction of reality in the child. New York: Routledge and Kegan
Paul.
Prechtl, H. F. R. (1977). Assessment and significance of behavioural states. In S. R. Berenberg

(Ed.), Brain, fetal and infant: current research on normal and abnormal development (pp. 79
90). Den Haag: Martinus Nijhoff.
Prechtl, H. F. R. (Ed.) (1984). Continuity of neural functions: from prenatal to postnatal life.
Oxford: Blackwell Scientific Publications.
Rochat, P. (1983). Oral touch in young infants: response to variations of nipple characteristics in
the first months of life. International Journal of Behavioral Development, 6, 123133.
Rochat, P. (1987). Mouthing and grasping in neonates: evidence for the early detection of what
hard or soft substance afford for action. Infant Behavior and Development, 10, 435449.
Rochat, P., & Hespos, S. J. (1997). Differential rooting response by neonates: evidence of an
early sense of self. Early Development and Parenting, 6(34), 105112.
Rochat, P., & Striano, T. (1999). Emerging self-exploration by two-month-old infants.

Developmental Science, 2, 206218.
Rochat, P. (2001). The infants world. Cambridge, MA: Harvard University Press.
Rochat, P., & Morgan, R. (1995). Spatial determinants in the perception of self-produced leg
movements by 3- to 5-month-old infants. Developmental Psychology, 31, 626636.
Page 11 of 13
Rochat, P. (2011). The self as phenotype. Cognition and Consciousness, 20(1), 109119.
Rochat, P. (2014). Origins of possession: owning and sharing in development. Cambridge:

Cambridge University Press.
Rovee-Collier, C. K., & Hayne, H. (2000). Memory in infancy and early childhood. In E. Tulving &
F. Craik (Eds.), Handbook of memory (pp. 267374). New York: Oxford University Press.
Schmuckler, M. A. (1996). Visual-proprioceptive intermodal perception in infancy. Infant

Behavior and Development, 19, 221232.
Seligman, M. E. P. (1975). Helplessness: on depression, development, and death. San Francisco:

W. H. Freeman.
Sherrington, C. S. (1906). The integrative action of the nervous system. New Haven, CT: Yale
University Press.
Simner, J., Sagiv, N., Mulvenna, C., Tsakanikos, E., Witherby, S., Fraser, C., et al. (2006).
Synesthesia: the prevalence of atypical cross-modal experiences. Perception, 35, 10241033.
Spector, F., & Maurer, D. (2009). Synesthesia: a new approach to understanding the
development of perception. Developmental Psychology, 45(1):175189.
Sullivan, M., & Lewis, M. (1989). Emotion and cognition in infancy: facial expressions during
contingency learning. International Journal of Behavioral Development, 12(2), 221237.
Trevarthen, C. (1980). The foundations of intersubjectivity: developments of interpersonal and

cooperative understanding in infants. In D. R. Olson (Ed.), The social foundations of language
and thought: essays in honor of Jerome S. Bruner. New York: W. W. Norton.
Tronick, E., Als, H., Adamson, L. B., Wise, S., & Brazelton, T. B. (1978). The infants response to
entrapment between contradictory messages in face to face interaction. Journal of the American
Academy of Child Psychiatry, 17, 113.
Watson, J. S. (1995). Self-orientation in early infancy: the general role of contingency and the
specific case of reaching to the mouth. In P. Rochat (Ed.), The self in infancy: theory and
research. Advances in Psychology, 112 (pp. 375394). Amsterdam: North-Holland/Elsevier
Science.
Wolff, P., Matsumiya, Y., Abrohms, I. F., van Velzer, C., & Lombroso, C. T. (1974). The effect of
white noise on the somatosensory evoked responses in sleeping newborn infants.
Electroencephalography and Clinical Neurophysiology, 37, 269274.
Notes:
(1) For a general review, see Rochat (2001).
Page 12 of 13
Page 13 of 13
Motivation from Control

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
A Response Selection Framework
Noam Karsh
Baruch Eitam
DOI:10.1093/acprof:oso/9780190267278.003.0012

This chapter discusses motivation stemming from being a causal agent and its underlying
mechanisms. It first reviews theories and findings from experimental psychology and
neuroscience with the goal of forming an integrative conceptualization of motivation from
control. It then differentiates between outcome and control feedback and defines control
feedback as the information the system uses for agency decision (versus testing for the presence
and value of an outcome). The chapter continues by explaining motivation from control using a
response selection process that is sensitive to the amount of control afforded by each response
of the response set, above and beyond the outcomes it is associated with. It ends by speculating
how the proposed explanation may contribute to the understanding of currently unexplained
behavioral phenomena such as stereotypical behaviors and self-mutilation.
Keywords: control, agency, response selection, motivation, self-mutilation
Introduction
Speculating on what is essential for the survival of an organism is a fruitful strategy for
indentifying the information that its brain-mind has adapted to promote (Eitam & Higgins, 2010,
2014). Once the information considered important for an organism has been identified, we can
proceed to ask how the search for and availability of such information influences the processes
of action generation (among others).
Page 1 of 22
Whatever the organism desires or needs to obtain, one constant challenge it faces is learning
cause-and-effect relations in a dynamic environment. Such learning enables preparation for
exploiting or escaping imminent changes in the environment. Moreover, information about
cause and effect is necessary for selecting appropriate actions for attaining desired outcomes
(Hauf, Elsner, & Aschersleben, 2004). In line with the motivational relevance framework
mentioned above (Eitam & Higgins, 2010), the importance of such information may explain why
controlthe registration that something has happened due to ones action (whether falsely or
correctly)has been postulated to be a motivator in both humans and lower animals (Burger &
Cooper, 1979; DeCharms, 1968; Higgins, 2012; Kelley, 1971; White, 1959).
(p.266) Recently, Higgins (2012) proposed the term control effectiveness as a motivational
force that can operate independently of the pursuit of valued outcomes (which he termed
outcome effectiveness). From this perspective, people may act for success in control
effectiveness (e.g., feeling efficacious or having autonomy or how you got there), regardless of
the actions expected outcome value (or what you end up with).
Recasting this perspective in information-processing terms (see also Eitam & Higgins, 2010,
2014; Eitam, Kennedy, & Higgins, 2013; Eitam, Miele, & Higgins, 2013), and differentiating
between these two types of information, we henceforth refer to information that is relevant for
decisions of agency as control feedback and to information relevant to ones degree of
obtaining desired outcomes as outcome feedback.
Distinct Types of Information from Action Effects: Control Versus Outcome

Feedback
Central accounts of human motivation emphasize that actions positive outcomes are reinforcing
(Thorndike, 1927; Skinner, 1953; Sutton & Barton, 1998; ODoherty, 2004). Correspondingly,
evidence for the influence of outcome feedback on motivation is abundant and has been
extensively described and reviewed (e.g., Carver & Scheier, 1990; Fishbach, Eyal & Finkelstein,
2010; Locke, 1982; Niv, Joel & Dayan, 2006). Hence, our focus is on reviewing evidence that
control feedback affects motivation when it is redundant or even uninformative as outcome
feedback.
One ubiquitous form in which humans and other animals enjoy control is by choosing. Leotti and
Delgado (2011; see also Leotti & Delgado, 2014) showed that people liked a cue that signaled an
upcoming free-choice trial more than a cue that signaled an upcoming no-choice trial, even
though choice did not have any actual impact on the outcome and presumably costs mental
resources (Baumeister, Bratslavsky, Muraven, & Tice, 1998). More broadly, research on choice
suggests that the act of choosing is rewarding at some level, regardless of the valence of choice
outcome (for a review, see Leotti, Iyengar, & Ochsner, 2010). Other studies in which
participants were given the opportunity to choose found that so-called intrinsic motivation
(measured by reported interest, enjoyment, and repeated choice of the task) is enhanced
presumably because of the increase in a feeling of autonomy (for a meta-analysis, see Patall,
Cooper, & Robinson, 2008).
A primary route for exercising control is through affecting the environment. In a largely
overlooked paper, Stephens (1934) suggested that in order to specify the effect of punishment
Page 2 of 22
on ones task performance, the physical medium of the feedback (i.e., the non-outcome
feedback-related part of the (p.267) effect) should be experimentally controlled. In several
experiments, he found that actions tend to be repeated after participants were given wrong as
feedback compared to when participants were given no feedback at all. Stephens suggested that
the mere fact that something happened that was contingent on ones action strengthened the
preceding action or, as he described it, had a stamp-in effect.
An action that was followed by a mundane and trivial perceptual change (e.g., a brief change in
luminance) is one of the simplest instances of (distal) control feedback (Thirkettle, Walton, Shah,
Gurney, Redgrave, & Stafford, 2013). Such extremely basic control feedback is easier to
manipulate and thus it is easier to control its informativeness regarding ones performance
(outcome feedback). Recently, Eitam, Kennedy, and Higgins (2013) used exactly this type of
primitive control feedback to show that control, through affecting the environment, motivates
behavior. In three experiments, participants responded to colored circles rapidly descending in
one of four vertical paths (Figure 12.1). Participants were to press the key that corresponded to
the spatial location of the falling circle, which was randomly determined. For half of the
participants, the circles changed their color and disappeared immediately after the key press
(positive control feedback), while for the other half the circle simply continued in its downward
path (no/negative control feedback). The (p.268) results showed that participants responses
were reliably faster when receiving control feedback as compared to not receiving it.
Figure 12.1 The Stamp tasks two basic

conditions (Eitam et al., 2013). In the
Control Feedback condition, the red
response cue changes its color to white and
disappears immediately after pressing the
relevant key; in the No-Effect condition, the
circle cue continues to descend with no
effect following a key press.
Page 3 of 22
One could reasonably argue that in Eitam

and colleagues (2013) experiments the
action effects carried some outcome
feedback as having an effect was always
tied to task success (a correct response).
Recently, Karsh and Eitam (2015) modified
the original paradigm to clearly
differentiate between the contribution of
outcome and control feedback to
motivation. Participants were asked to
freely and randomly select and press one of Figure 12.1 The Stamp tasks two basic
four response keys on the appearance of a conditions (Eitam et al., 2013). In the
Control Feedback condition, the red
cue. This setup minimized the potential role
response cue changes its color to white and
of effects as outcome feedback, as no disappears immediately after pressing the
performance proper (apart from being relevant key; in the No-Effect condition, the
random) was actually required from the circle cue continues to descend with no
effect following a key press.
participants. To manipulate control
feedback, participants were assigned to one
of three conditions (Figure 12.2): a High
Probability Effect condition in which any key press would reliably deliver an effect (at a 90%
chance); a No Effect condition in which no key press would (p.269) deliver an effect; and a Key
Specific Probability Effect condition in which response keys varied in their reliability of
delivering effects (90%, 60%, 30%, 0).
Page 4 of 22
Figure 12.2 This illustration schematically

describes a single trial of the three
experimental conditions of Karsh and
Eitams (2015) study. After the appearance
of a cue (a red circle), participants had to
freely and randomly respond with one of four
relevant response keys. Immediately
following their response, the red cue could
change its color to white for 100 ms and
disappear (effect trials) or no perceptual
change would appear (no effect trials). As
described in text, the chance of such an
effect occurring after a response was
manipulated according to experimental
conditions.
Page 5 of 22
Karsh and Eitam (2015) found that although

participants in the Key Specific Probability
Effect condition seemingly attempted to
respond randomlythe proportions of the
four responses were very close to one
anotherthey tended to select the keys
associated with a higher chance of causing
an effect more frequently than those
associated with no chance of causing an
effect; that is, control feedback
(operationalized here as having an effect)
affected participants task-relevant attempts
to respond randomly by acting as a classic
(and competitive) reinforcer. A second
important result of these experiments was
the replication of Eitam et al. (2013) of
faster reaction time for the Highly
Probability Effect condition (vs. No Effect
condition). This replication enabled the
generalization of the motivating effect that
control feedbackhas over two, very
different, tasks.
Figure 12.2 This illustration schematically
Most important for our current argument is describes a single trial of the three
that the pattern of increasing the experimental conditions of Karsh and
probability and speed of selection of a Eitams (2015) study. After the appearance
response due to its reliable association with of a cue (a red circle), participants had to
freely and randomly respond with one of four
control feedback resembles the one found
relevant response keys. Immediately
for outcome feedback (e.g., food or money) following their response, the red cue could
and suggests that a similar mechanism is change its color to white for 100 ms and
behind both. In short, to exploit the disappear (effect trials) or no perceptual
change would appear (no effect trials). As
environment, a response represented with described in text, the chance of such an
better outcome is likely to be selected over effect occurring after a response was
other responses from the response set manipulated according to experimental
(Sutton & Barton, 1998) and will also tend conditions.
to be selected faster. We elaborate further
in the section on the effect control feedback
has on response selection.
Thus far, we have argued that action effects hold at least two distinct types of information, both
of which can motivate action: information about the progress toward attaining a desired
outcome (outcome feedback), and information that is relevant for decisions of agency (control
feedback). We have demonstrated that control feedback in the form of having an effect
influences further action in a manner that is similar to outcome feedback. To further develop the
dissociation between outcome and control feedback, in the following section we attempt to
specify the conditions that lead the mind to identify an effect as control feedback.
Page 6 of 22
What Information Does the Mind-Brain Count as Control Feedback?

Control Feedback-Based Sensory Prediction
Action effects are encoded with the actions they relate to (Hommel, Musseler, Aschersleben, &
Prinz, 2001; James, 1890; Prinz, 1987), seemingly even when they are goal (i.e., outcome)
irrelevant (Band, Steenbergen, Ridderinkhof, Falkenstein, & Hommel, 2009). Given the
informational framework proposed (p.270) earlier, one can speculate that the brain invests
resources to monitor action effects because control feedback is valuable to the organism in the
long run. In order to perform a goal-directed action, the infant must first learn contingencies
between her actions and their effects in the current environment; only then can she voluntarily
reproduce these effects to attain desired outcomes (Elsner & Aschersleben, 2003). For example,
according to the theory of event coding (TEC), only after an actor associates the code of a
particular action with the code of the actions effects are the representations of the intended
action effects considered causally responsible for the selection of appropriate actions (Elsner &
Hommel, 2001).
To register an effect as control feedback (i.e., ones own action effect) the mind-brain needs to
determine that it indeed is responsible for its occurrence. In an influential model of action
control and decisions of agency (the comparator model), Blakemore, Wolpert, and Frith (1998)
mustered evidence that the motor system uses information about ones motor prediction to
identify the effects agent (me versus other; Blakemore, Frith, & Wolpert, 1999; Blakemore &
Frith, 2003). Specifically, the motor control system generates an internal forward model that is
based on an efference copy of the motor command and compares it with the action effect (i.e.,
the actual sensation). Accordingly, a match between the sensory prediction and the actual
sensation increases the subjective or explicit decision of agency, defined as the ability to refer to
oneself as the author of ones own actions (Vignemont & Fourneret, 2004). However, when there
is a significant discrepancy between the sensory prediction and the actual sensation, confidence
about oneself being the agent decreases. Thus, by this model, decision of agency follows from a
comparison between predicted sensory consequences that is based on an efference copy of the
motor command (which is also in charge of passing this information to a different comparator in
order to modify the action, if necessary, to achieve the desired goal).
Importantly, this model also implies that the relevant comparator for agency decisions is
orthogonal to direct information about the identity of the desired goal and relies solely on the
sensory prediction from the efference copy of the motor command and the actual sensory
feedback. To clarify further, we can dissociate between goal-relevant information (i.e., outcome)
and agency-relevant information. To determine whether the desired goal has been achieved, one
can only use information about the represented desired goal and the actual goal state and can
remain blind to sensory-motor information regarding the causal chain that led the process.
However, for decisions of agency, one needs to compare the sensory prediction from the motor
command to the actual sensory feedback regardless of any information about the identity of (or
the progress toward) the desired goal.
(p.271) Haggard, Clark, and Kalogeras (2002) showed that the temporal binding effect (TB)
the subjective closeness in time of ones own actions and their effectsoccurs only when an
action is voluntary. Given that TB is sensitive to many manipulations related to agency (Moore &
Obhi, 2012) and is described by some as an implicit measure of self-agency (Moore & Haggard,
Page 7 of 22
2008), the TB effect gives some support to the notion that implicit decision of agency also
depends on the existence of predicted sensation from the motor command.
We propose that one path for an effect to earn its relevance as control feedback (i.e., ones
own action effect) is the degree of its congruence with ones sensory prediction that is based on
an efference copy of the motor command. Hence, in one category of control feedback, events
indicate that the environment (including oneself) was affected in line with ones sensory
prediction. Importantly, this definition does not exclude previous conceptual terms that were
used in relation to the notion of human agency, such as effectance (White, 1959),
competence (White, 1959), efficacy (Bandura, 1982), or autonomy (Deci & Ryan, 1987) as
explanatory concepts, but merely serves to explain motivation from control using current models
of action control.
Fixed Parameters in Determination of Control Feedback

We have indicated that predicted sensory consequence holds information crucial for decisions of
agency and thus earns its status as control feedback. However, although motor predictions
seem to be an important parameter in the generation of motivation from control, it can be
generated only after some contingency between action and effect has been encoded (Bednark,
Reynolds, Stafford, Redgrave & Franz, 2013; Elsner & Aschersleben, 2003) and enables relevant
sensory predictions. The motor control system must be capable of generating and updating its
motor plane in accordance with novel self-caused changes in the environment, tooand, through
that process, learning causal relationships between actions and novel effects (Bednark et al.,
2013).
Given the above, we propose that the mechanisms involved in decisions of agency may also,
orthogonal to the motor prediction, be sensitive to a set of fixed parameters of action effects that
are independent of prior knowledge. We suggest that a leading candidate for such a fixed
parameter in the determination of control is the temporal contingency between action and
effect.
Recently, Walton, Thirkettle, Redgrave, Gurney, and Stafford (2013) showed that when inserting
a minimum of 75 ms temporal delay between an agents action and novel action effect (without a
relevant sensory prediction), action-effect learning is impaired. The author suggested that low
latency (p.272) dopamine signal is in charge of stamping-in the action that preceded the
unexpected event. However, when the effect is temporally delayed, the preceded action
identification is decreased, possibility due to contamination of the motor records. As for agency
decisions, when action effect is temporally delayed, it is more difficult for the agent to learn
causal relations between ones motor command and its sensory consequences.
Interestingly, temporal contingency was affecting agency determination also when motor
predictions were enabled. For instance, temporal discrepancy between an action and a
proceeding effect decreases the temporal binding effect, imposed as an implicit measure of
sense of agency, when participants motor system was able to generate sensory predictions
(fixed blocks) and when it was unable (randomized blocks) (Haggard, Clark, & Kalogeras, 2002).
Although the effects mentioned above could result from different mechanisms, based on these
studies and our own, we suggest that temporal contingency may function as a fixed control
Page 8 of 22
parameter that is independent of the specific content of the motor prediction or of its mere
existence and that affects control-based response selection.
Recently Eitam et al. (2013) studied the influence of the temporal delay on motivation from
control. They showed that although participants were able to predict action-effect delay (using a
between-subject design with multiple trials), the agents own action effects (as a private case of
control feedback) enhance performance only when they appear less than 300 ms after the
action. Definitely, much more research is required to establish the degree of which the temporal
parameter is indeed fixed (i.e., prior knowledge independent) and to discover other fixed
parameters that affect implicit decisions of agency. Future research should also consider a
biological explanation for why temporal delay is a crucial parameter in agency determination.
For instance, as suggested earlier, the sensory consequences from action effect is compared to
the sensory prediction for agency determination. Even a few hundred ms delay between action
and effect could be more than the life span of the neural network representing the sensory
prediction. Considering this alternative explanation, the sensory prediction could be condemned
due to delay, and without sensory prediction, no comparison can be made and hence positive
agency cannot be determined.
Thus far, we have argued that a positive decision of agency is motivating in and of itself, and we
have proceeded to outline two categories of events that would be seen by the mind as control
feedback and which are markedly different from parameters that would define outcome
feedback. In the next section, we link motivation from control feedback to relevant
neuroscientific models and present the mechanism through which we propose that a positive
decision of agency motivates action.
(p.273) Control as Reward: A Neuroscientific Perspective on Motivation from

Control
In the second section, we mentioned the similarity between the processes through which control
(positive decision of agency) and outcome (desired goal) feedbacks motivate behavior. A positive
outcome following ones action (i.e., outcome feedback) leads to a phasic increase in dopamine
secretion in the striatum (Delgado, Locke, Stenger & Fiez, 2003; Fiorillo, Tobler & Schultz,
2003; Schultz, 1998); this, in turn, influences preferences (Morris, Nevet, Arkadir, Vaadia, &
Bergman, 2006) and increases goal (i.e., outcome) directed behavior (Palmiter, 2008).
According to Redgrave, Gurney, and Reynolds (2008), an unexpected sensory event that follows
an action generates a phasic dopamine burst to reinforce the re-selection of the previous motor
command in the same context in order to confirm oneself as the agent who caused the event. If
no relevant motor copy preceded the event, the event is likely to be caused by an external
source. Thus, if the event was externally caused, repetition of the preceding but irrelevant motor
command would fail to cause the event.
Dopamine plays another role. Once the causal relation between action and effect is learned, the
transient dopaminergic response to a stimulus (the cue) is considered to carry information about
the magnitude and probability of reward given that cue (Tobler, Fiorillo, & Schultz, 2005). This
represented reward value is considered to reflect both subjective and objective reward values
(for a review, see Schultz, 2012) and to affect decision-making. In a trial by trial free-choice
Page 9 of 22
task, Samejima, Ueda, Doya, and Kimura (2005) showed that the striatum represents an actions
reward value (crucially, before the execution of that action) and in that manner determines
response selection.
Crucially and in line with the behavioral evidence reviewed before, recent neuroscientific
findings suggest that control feedback activates the reward system regardless of the outcome
feedback it carries. While most studies on action effect focused mostly on positive versus
negative outcome feedback, Behne, Scheich, and Brechmann (2008) studied the reward
properties of an effect that was invalid as outcome feedback but was valid as control feedback.
In this study, OK appeared immediately after the participants pressed a button (while
performing a tone discrimination task), regardless of the accuracy of the response (hence invalid
as outcome feedback). The study also included two other conditions, a pseudo-feedback
condition (in which OK appeared at random timesthat is, not contingent on the action) and a
no-feedback condition. Behne et al. (2008) found that the left dorsal striatum, which is strongly
implied in representing the motivational value of actions (ODoherty, 2004), was activated more
after outcome uninformative but temporally contingent (p.274) (i.e., control informative)
feedback compared to the action non-contingent (temporally uncorrelated with an action) and
no-feedback conditions. Although one may argue that OK is a positive rather than a neutral
feedback, even though it appeared regardless the accuracy of the responses, only the
contingency between action and feedback was associated with strong activation in the dorsal
striatum.
Similarly, Tricomi, Delgado, and Fiez (2004) showed that the caudate nucleus is activated for
valued outcome feedback, but only when the participants perceived a connection between their
actions and these outcomes. In one of their experiments (Experiment 3), pseudo choice-
dependent outcome activated the caudate nucleus more than when participants had no
(apparent) choice. In light of these findings, Tricomi et al. (2004) proposed that the perception of
an action-outcome relationship (i.e., control feedback), not the mere presence of a positive
outcome, is necessary for caudate activation. Their findings also uncovered a correlation
between caudate activation and subjective rating of control that was stronger in the choice
condition than for the no-choice condition (see also Leotti and Delgado, 2011). In the language
of the proposed framework, only information categorized as control feedback is followed by
reward (because of the information that control feedback carries). In the next section, we will
focus on response selection as a mechanism underlying observed motivation and specifically on
how control feedback affects response selection.
The Effects of Control Feedback on Response Selection

Action production involves selection among many potential actions. The optional actions are
represented with their predicted reward values, and the action with the highest predicted
reward is selected (Redgrave, Prescott, & Gurney, 1999). Additionally, representations of
possible actions compete for resource allocation in winner takes all manner (Redgrave et al.,
1999). Redgrave et al. (1999) assign the role of selecting between optional actions to the basal
ganglia, again implicating the neurotransmitter dopamine in selection between competing
options for action.
Page 10 of 22
The striatum is involved in action control through representing the predicted reward values of
specific responses, thus enabling value-relevant selection (for reviews, see Balleine, Delgado &
Hikosaka, 2007; ODoherty, 2004). If a given response is represented with a higher reward value
compared to other response options, it is selected more rapidly (Brown & Bowman, 1995) and
more frequently (Samejima et al., 2005).
(p.275) As noted, the striatum is also activated following an action contingent event (control
feedback) that is uninformative regarding outcomes (Behne et al., 2008). Analogously, we
propose that a response is also encoded and represented with its control predicted reward value
in order to successfully compete with other optional responses for future selection.
The problem of selection is posed as the resolution of the conflict between competing courses of
action that are afforded by an object (Cisek, 2007; Cisek & Kalaska, 2010). Given a response
selection mechanism that is affected by control feedback, a response represented with a greater
control reward value will be preferred among other response options from the response set.
This proposition was recently supported in Karsh and Eitams (2015) experiment (described
previously in this chapter), which showed an increase in both speed and frequency of selecting a
response associated with high probability to deliver an effect. Thus, predicted control biased
response selection even though it somewhat damaged task performance (outcome).
A two-step account model of sense of agency had described different processing systems that
generate conceptual (explicit high-level judgment of agency) and non-conceptual (implicit low-
level feeling of agency) accounts of sense of agency (Synofzik, Vosgerau, & Newen, 2008).
Recently, Moore, Middleton, Haggard, and Fletcher (2012) provided an empirical support for the
existence of different processing behind implicit and explicit sense of agency. Consistently, and
due to our interest in the effect that positive decision of agency has on response selection, we
suggest that to understand how control information (input) influences the generation of
further action (output), it is beneficial to think of implicit and explicit decisions of agency in
order to understand the nature of their influence on different levels of action selection.
Following this conceptualization, Karsh and Eitam (2015) explored whether implicit (actual
number of control feedback) and explicit (control available knowledge) decisions of agency
affect different levels of response selection. Their findings indicated that both the sum of actual
control feedback and control available knowledge affect action selection (actually, reaction
times). However, when conscious knowledge of effects and actual control feedback were used
simultaneously to predict speed of action selection, only actual control feedback emerged as a
significant predictor.
The authors suggested that explicit decision of agency (explicit knowledge of effectors degree
of control) influenced the choice of effectorwhich finger to press (which is a relatively abstract
and consciously available representation)but only indirectly affected the selection of motor
parameters (a lower-level action representation, which is considered to be inaccessible to
consciousness). Conversely, the selection of low-level motor parameters was directly affected by
implicit decisions of agency (for a simple illustration, see Figure 12.3). These (p.276) findings
are consistent with current views on how the abstractness of the representation of an action
guides both further action selection (Badre, Kayser, & DEsposito, 2010) and the degree to
Page 11 of 22
which it is accessible to consciousness (Carota, Desmurget & Sirigu, 2009; Haggard, Clark, &
Klogeras, 2002).
In this section, we both suggested and

presented supporting findings that the
selection of actions is affected by their
predicted control value. In the final section
of this chapter, we further develop control
feedback-based response selection to
propose that as outcome-based response
selection, these two can vary on an
optimality dimension. Specifically, given
that decision of agency can result from non-
conceptual sensory effects coming from
ones own body (e.g., haptic and
proprioceptive feedback) and from more
complex social effects (p.277) from the
outside environment, we refer to
circumstances as non-optimal when
control-based action selection is largely Figure 12.3 A schematic illustration of our
control-based response selection framework.
based on sensory or perceptual feedback
The illustrated path begins with any event,
rather than on a more central and external or internal, which elicits a
potentially broader influence (effect), such conscious representation of the intended
as an effect of a social nature (i.e., socially action effect (marked by the letter S). The
upper squares represent higher-level
informative).
response options (R1Rn; e.g., optional
fingers) and the rotated lower squares
Non-Optimal Control Feedback-Based represent lower-level (e.g., motor) response
Response Selection parameters (r1rn). The width and size of the
A Shared Mechanism for Stereotypy and squares depicts the magnitude of a
Social Interactions responses predicted control value. The
response option with the highest predicted
Although any deliberate action demands control value is selected. The dashed arrows
selection at all levels of the action (see depict control information that is relevant
Cisek, 2007), at least when outcomes are for explicit decision of agency (i.e., a match
between intended and actual effect, and
concerned individuals are seemingly tuned information out of the motor system such as
to high-level outcome feedback (e.g., the desires and beliefs). The dotted arrows
why of an action, such as construing the depict control value from explicit (right) and
implicit (left) control decisions, which is fed
end state of a hands movement toward the back to the action control system, updating
switch on the wall as turning the lights on, action parameters control predicted value,
rather than of pressing the light switch; which will in turn, together with outcome
predicted value (described elsewhere),
Vallacher & Wegner, 1987). We propose determine further response selection.
that the same principle may hold for
control-based action selection in that action
will tend to be selected based on broader
context control feedabck.
Page 12 of 22
As described above, the motor control system uses an internal prediction model to predict the
sensory consequences of a given motor command on ones body. However, actions often have
extended effects beyond ones body that the mind can model as a control feedback. For instance,
in the case of affecting the external environment, as in social interactions (e.g., communication),
the motor control system can model other peoples response to ones motor command (Wolpert,
Doya, & Kawato, 2003). We suggest that since affecting the social environment extends ones
influence more than affecting ones body, in a social context, selection would tend to be based
on social control feedback more than on the narrow influence of affecting ones body.
At the very early stages of a childs development, the social contexts of action effects are
probably not even registered as control feedback (as they require substantial understanding of
the social environment to generate an internal forward model that considers an environment-
specific context). In these developmental phases then, it seems reasonable to argue that the
sensory feedback of ones action from controlling ones body dominates response selection (for a
similar argument, see Oztop, Bradley, & Arbib, 2004). Response selection that is solely
determined by proximal control feedback from controlling ones body may manifest itself by
continuously repeating responses that produce such feedback, as in stereotypy.
Indeed, normally developing infants often engage in repetitive behaviors producing more
proximal control feedback (e.g., kicking, rocking and waving; Thelen, 1981). These behaviors do
not serve a distant goal and it seems that (p.278) they are maintained for the sake of the
movement itself (Thelen, 1980). After the first year of life, such behaviors gradually cease, and
the infant engages more frequently in actions that can potentially control the outside
environment (e.g., influencing the social environment; Carpenter, Nagell, Tomasello,
Butterworth, & Moore, 1998).
Stereotypies are also common in individuals with autism. According to Baron-Cohen, Leslie, and
Frith (1985), the child suffering from autism finds the social environment unpredictable,
reflecting and crystallizing her poor social skills. Recently it was suggested that impaired social
skills in individuals with autism could result from a poor internal forward model for social effects
(Blackemore & Decety, 2001). In one study (Stoit et al., 2011), autistic dyads and normal dyads
conducted a joint action task that required predicting the consequences of the action of the
partner and coordinating to succeed in the task. Their results showed that the autistic dyads
performed poorer on this task than the normal dyads and showed an impaired sense of agency
relative to the control group, but were no different from the normal participants when the same
task was conducted individually.
Our proposed framework suggests one route through which a lack of ability to exert control (by
generating motor predictions) on the social environment could lead to stereotypy. Specifically,
the lack of the ability to generate motor predictions for social effects leads to action selection
based on sensory (proximal) control feedback from controlling ones body (possibly it is the
minds only option to exercise control by having an effect). In a vicious cycle, this, in turn,
maintains and reinforces stereotypy and social isolation.
Some support for this proposal relating social effects and stereotypy comes from studies on the
role of oxytocin (OT) in both repetitive behaviors and the processing of social information. One
Page 13 of 22
hypothesized effect of OT is to increase the processing of social information (Bartz, Zaki, Bolger,
& Ochsner, 2011); intriguingly, Hollander et al. (2003) showed that the administration of OT to
older individuals suffering from autism significantly decreased their stereotypical behaviors.
Although further research is required to determine whether OT induction reduced stereotypy
through increasing decisions of agency over social effects, this interpretation lends some
support to the proposed link between lack of social effects and stereotypy.
Our proposal that stereotypy can result from a deficit in agency decisions for social effects has
some biological support. The striatum, as mentioned above, is thought to represent predicted
reward value (from both outcome and control feedback) and in that manner determines
response selection, and is found to function abnormally in the autistic population (Chevallier,
Kohls, Troiani, Brodkin, & Schultz, 2012). Specifically, Dawson et al. (2005) hypothesized that
individuals with autism may have impairment in representing reward value (p.279) from social
stimuli. Based on the finding presented above, we suggest that both stereotypy and maladaptive
social interactions may reflect non-optimal response selection based on control-predicted value
from own-body effects, perhaps because of impairment in representing reward-predicted value
from social effects. Our proposal fits well with recent arguments that the impairment in social
cognition in autism is a consequence of a social motivation deficit rather than its cause
(Chevallier et al., 2012).
In some extreme cases of deprivation of control feedback, humans and animals have no
opportunity for generating effects over the external environment, simply because it is not
afforded by the environment; in other words, they find themselves helpless (Seligman, 1972). In
line with our hypothesis, research on socially deprived children and animals under
environmental restrictions shows an increase in stereotypy in these populations (Maclean, 2003;
Powell, Newman, Pendergast, & Lewis, 1999). One longitudinal study explored the prevalence of
behaviors that are associated with deprivation in institutionalized children at three time points
(Beckett et al., 2002). One common behavior that decreased significantly after adoption was
rocking behavior (47% to 19%), a stereotypy, which was significantly higher among children
who experienced the longest deprivation.
In summary, when control feedback from affecting the social environment is either not
registered as control feedback at all (as in the example above involving young babies or autistic
children) or simply not (or less) afforded by the environment (e.g., the case of institutionalized
children), positive decisions of agency from affecting ones own body (such as sensory effects
from repeated movements) gains prevalence, and hence the actions associated with it are
rewarded, selected, and rewarded again, further strengthening their probability to be selected
yet again. This general sketch of the shared mechanism for stereotypy and social interaction
provides a motivational and mechanistic framework for stereotypy and its correlation with
maladaptive social interactions.
Control Deprivation and Control Compensation

When experiencing high degrees of control, people are likely to be more confident in their ability
to perform certain actions or attain desired outcomes (Taylor & Brown, 1988). However, when
experiencing or perceiving low degrees of control feedback, people compensate for their lack of
perceived control by regaining the feeling of control indirectly (e.g., trusting powerful others;
Page 14 of 22
see Kay, Gaucher, Napier, Callan, & Laurin 2008; see also Skinner, 2007) or by engaging in
maladaptive behaviors such as eating restrictions or hostility (Shapiro, Schwartz, & Astin,
1996). In extreme cases of control deprivation (p.280) (e.g., captivity), people and animals are
more likely to engage in stereotypical repetitive behaviors (e.g., Garner, Meehan, & Mench,
2003) or to develop learned helplessness (Abramson, Seligman, & Teasdale, 1978).
As is the case with desired outcomes, the proposed mechanism underlying motivation from
control is consistent with the possibility that control deprivation may change the reward value of
otherwise identical control feedback events. For example, neuroscientific studies showed that
the level of dopamine secretion followed by a positive outcome is affected by ones current state
(Epstein et al., 2003; Nader et al., 2012). Epstein et al. (2003), for example, showed that food
deprivation increased the reinforcing value of food but did not significantly affect its hedonic
preference. Accordingly, an action that precedes control feedback may be rewarded with
greater control value after a prolonged state of control feedback deprivation. Hence, an action
that receives positive control feedback after a prolonged state of control deprivation may be
even more strongly preferred over other response options.
Using our proposed control motivation framework, control compensation can help us
understand various seemingly persistent or motivated behaviors that cannot be explained
through outcome feedback (i.e., eyelash pulling, nail biting). While the control compensation
explanation for these behaviors is similar to that of stereotypy in the previous section,
stereotypy and social interaction are, at least seemingly, related because they reflect (for typical
individuals) the normal developmental trajectory of causal learning (from the ones own body to
controlling the social environment). The behavioral phenomena mentioned in the current section
are a more diverse phenotype of compensatory control and hence require additional factors for
determining the exact form that compensatory control would take. Consider the compensatory
control analysis of engagement in self-mutilation such as cutting behavior: the subjective
experience of an uncontrollable prolonged state of negative affect leads to, for example, cutting
behavior, which leads to the very salient effects (albeit own-body effects) of a painful sensation,
bleeding, and other physical marks (i.e., control feedback), and is rewarded with increased
control reward value (given the constant state of control deprivation). Hence, re-experiencing a
similar trigger (e.g., negative affect) leads to an increasingly growing probability of re-selecting
the same action (cutting behavior), reflecting the increased expected reward value of positive
decisions of agency (for various other reasons for cutting behavior, see Kumar, Pepe, & Steer,
2004).
Conclusion
In this chapter we have brought together findings and related terminologies from psychology
and neuroscience to suggest a coherent framework for (p.281) motivation from control. We
presented various findings that support the argument that action effects that pass as
information about ones agency (control feedback) are themselves motivating. We continued by
asking what kind of information influences the mind-brain decisions of agency. Building on our
and others work on explicit and implicit decisions of agency, we suggested that one kind of
control information the mind-brain is tuned to are effects that are highly congruent with ones
recent motor predictions. In addition, we presented reasons to believe that the mind-brain is
also sensitive to additional (fixed) parameters for classifying events as positive control
Page 15 of 22
feedback, such as very close temporal contingency between actions and effects. Further
research is clearly needed to reveal which parameters are crucial for reaping motivation from
control.
Next, we introduced our main argument by which motivation from control plays out as
response selection as a function of the degree of control feedback-based reward. This framework
received empirical support from a series of experiments showing that reliable control feedback
affected both speed (RT) and frequency of responding. Here, too, much further research is
needed to see how outcome (e.g., goal-related) and control feedback combine (or compete) to
affect action selection.
Finally, we sketched how the proposed framework would explain negative outcome-related
behaviors such as stereotypy and self-mutilation behavior. We speculated on the malfunction of
a natural hierarchy of control feedbacks in some cases (e.g., stereotypy) and a possible
compensatory relationship between different courses of actions that is based on the subjective
need for control feedback (self-mutilation).
Acknowledgments
This work was supported by a research grant by the Israeli Science Foundation to B.E. (277/12).
References
Abramson, L. Y., Seligman, M. E. P., & Teasdale, J. (1978). Learned helplessness in humans:
Critique and reformulation. Journal of Abnormal Psychology, 87, 4974.
Badre, D., Kayser, A. S., & DEsposito, M. (2010). Frontal cortex and the discovery of abstract
action rules. Neuron, 66, 315326.
Balleine, B. W., Delgado, M. R., & Hikosaka, O. (2007). The role of the dorsal striatum in reward
and decision-making. The Journal of Neuroscience, 27, 81618165.
Band, G. P. H., van Steenbergen, H., Ridderinkhof, K. R., Falkenstein, M., & Hommel, B. (2009).
Action-effect negativity: irrelevant action effects are monitored like relevant feedback. Biological
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122
147.
Baron-Cohen, S., Leslie, A. M., & Frith, U. (1985). Does the autistic child have a theory of
mind? Cognition, 21, 3746.
Bartz, J. A., Zaki, J., Bolger, N., & Ochsner, K. (2011). Social effects of oxytocin in humans:
context and person matter. Trends in Cognitive Sciences, 15, 301309.
Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: is the
active self a limited resource? Journal of Personality and Social Psychology, 74, 12521265.
Page 16 of 22
Bednark, J. G., Reynolds, J. N. J., Stafford, T., Redgrave, P., & Franz, E. A. (2013). Creating a
movement heuristic for voluntary action: electrophysiological correlates of movement-outcome
learning. Cortex, 49, 771780.
Beckett, C., Bredenkamp, D., Castle, J., Groothues, C., OConnor, T. G., Rutter, M., & the English
and Romanian Adoptees Study Team. (2002). Behavior patterns associated with institutional
deprivation: a study of children adopted from Romania. Journal of Developmental and Behavioral
Pediatrics, 23, 297303.
Behne, N., Scheich, H., & Brechmann, A. (2008). The left dorsal striatum is involved in the
processing of neutral feedback. Neuroreport, 19, 14971500.
Blakemore, S. J., & Decety, J. (2001). From the perception of action to the understanding of
intention. Nature Reviews: Neuroscience, 2, 561567.
Blakemore, S. J., & Frith, C. D. (2003). Self-awareness and action. Current Opinion in
Neurobiology, 13, 219224.
Blakemore, S. J., Frith, C. D., & Wolpert, D. W. (1999). Spatiotemporal prediction modulates the
perception of self-produced stimuli. Journal of Cognitive Neuroscience, 11, 551559.
Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self produced
tickle sensation. Nature Neuroscience, 1, 635640.
Brown, V. J., & Bowman, E. M. (1995). Discriminative cues indicating reward magnitude
continue to determine reaction time of rats following lesions of the nucleus accumbens.
European Journal of Neuroscience, 7, 24792485.
Burger, J. M., & Cooper, H. M. (1979). The desirability of control. Motivation and Emotion, 3,
381393.
Carota, F., Desmurget, M., & Sirigu, A. (2009). Forward modeling mediates motor awareness. In
L. Nadel and W. Sinnott-Armstrong (Eds.), Conscious will and responsibility: a tribute to
Benjamin Libet (pp. 97108). Oxford: Oxford University Press.
Carpenter, M., Nagell, K. & Tomasello, M. (1998) Social cognition, joint attention, and
communicative competencies from 9 to 15 months of age. Monographs of the Society of
Research in Child Development, 63, 1143.
Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and negative affect: a
control-process view. Psychological Review, 97, 1935.
Chevallier, C., Kohls, G., Troiani, V., Brodkin, E. S., & Schultz, R. T. (2012). The social
motivation theory of autism. Trends in Cognitive Sciences, 16, 231239,
Cisek, P. (2007). Cortical mechanisms of action selection: the affordance competition hypothesis.
Philosophical Transactions of the Royal Society B: Biological Sciences, 362, 15851599.
Page 17 of 22
Cisek, P., & Kalaska, J. F. (2010). Neural mechanisms for interacting with a world full of action
choices. Annual Review of Neuroscience, 33, 269298.
Dawson, G., Webb, S. J., Wijsman, E., Schellenberg, G., Estes, A., Munson, J., et al. (2005).
Neurocognitive and electrophysiological evidence of altered face processing in parents of
children with autism: implications for a model of abnormal development of social brain circuitry
in autism. Developmental Psychopathology, 17, 679697.
DeCharms, R. (1968). Personal causation. New York: Academic Press.
Deci, E. L., & Ryan, R. M. (1987). The support of autonomy and the control of behavior. Journal
Delgado, M. R., Locke, H. M., Stenger, V. A., & Fiez, J. A. (2003). Dorsal striatum responses to
reward and punishment: effects of valence and magnitude manipulations. Cognitive, Affective
and Behavioral Neuroscience, 3, 2738.
de Vignemont, F., & Fourneret, P. (2004). The sense of agency: a philosophical and empirical
review of the who system. Consciousness and Cognition, 13, 1119.
Eitam, B., & Higgins, E. T. (2010). Motivation in mental accessibility: relevance of a

representation (ROAR) as a new framework. Social and Personality Psychology Compass, 4, 951
967.
Eitam, B., & Higgins, E. T. (2014). Whats in a goal? The role of motivational relevance in
cognition and action. Brain and Behavior Science, 37, 141142.
Eitam, B., Kennedy, P. M., & Higgins, E. (2013). Motivation from control. Experimental Brain
Research, 229, 475484.
Eitam, B., Miele, D. B., & Higgins, E. T. (2013). Motivated remembering: remembering as
accessibility and accessibility as motivational relevance. In D. E. Carlston (Ed.), The Oxford
handbook of social cognition (pp. 463475). New York: Oxford University Press.
Elsner, B., & Aschersleben, G. (2003). Do I get what you get? Learning about the effects of self-
performed and observed actions in infancy. Consciousness and Cognition, 12, 732751.
Psychology: Human Perception and Performance, 27, 229240.
Epstein, L. H., Truesdale, R., Wojcik, A., Paluch, R. A., & Raynor, H. A. (2003). Effects of
deprivation on hedonics and reinforcing value of food. Physiology and Behavior, 78, 212227.
Fiorillo, C. D., Tobler, P. N., & Schultz, W. (2003). Discrete coding of reward probability and
uncertainty by dopamine neurons. Science, 299, 18981902.
Fishbach, A., Eyal, T., & Finkelstein, S. R. (2010). How positive and negative feedback motivate
goal pursuit. Social and Personality Psychology Compass, 4, 517530.
Page 18 of 22
Garner, J. P., Meehan, C. L., & Mench, J. A. (2003). Stereotypies in caged parrots, schizophrenia
and autism: evidence for a common mechanism. Behavioural Brain Research, 145, 125134.
Hauf, P., Elsner, B., & Aschersleben, G. (2004). The role of action effects in infants action
control. Psychological Research, 68, 115125.
Higgins, E. T. (2012). Beyond pleasure and pain: how motivation works. New York: Oxford
University Press.
Hollander, E., Novotny, S., Hanratty, M., Yaffe, R., DeCaria, C. M., Aronowitz, B. R., & Mosovich,
S. (2003). Oxytocin infusion reduces repetitive behaviors in adults with autistic and Aspergers
disorders. Neuropsychopharmacology, 28, 193198.
Hommel, B., Msseler, J., Aschersleben, G., & Prinz, W. (2001). The theory of event coding
(TEC): a framework for perception and action planning. Behavioral and Brain Sciences, 24, 849
937.
James, W. (1890). Principles of psychology. New York: Holt.
Karsh, N., & Eitam, B. (2015). I control therefore I do: judgments of agency influence action
selection. Cognition, 138, 122131.
Kay, A. C., Gaucher, D., Napier, J. L., Callan, M. J., & Laurin, K. (2008). God and the
government: testing a compensatory control mechanism for the support of external systems.
Journal of Personality and Social Psychology, 95, 1835.
Kelley, H. H. (1971). Attribution in social interaction. New York: General Learning Press.
Kumar, G., Pepe, D., & Steer, R. A. (2004). Adolescent psychiatric inpatients self-reported
reasons for cutting themselves. Journal of Nervous and Mental Disease, 192, 830836.
Leotti, L. A., & Delgado, M. R. (2011). The inherent reward of choice. Psychological Science, 22,
13101318.
Leotti, L. A., & Delgado, M. R. (2014). The value of exercising control over monetary gains and
losses. Psychological Science, 25, 596604.
Leotti, L. A., Iyengar, S. S., & Ochsner, K. N. (2010). Born to choose: the origins and value of the
need for control. Trends in Cognitive Sciences, 14, 457463.
Locke, E. A. (1982). Relation of goal level to performance with a short work period and multiple
goal levels. Journal of Applied Psychology, 67, 512514.
Maclean, K. (2003). The impact of institutionalization on child development. Development &

Psychopathology, 15, 853884.
Page 19 of 22
Moore, J., & Haggard, P. (2008). Awareness of action: inference and prediction. Consciousness
Moore, J. W., Middleton, D., Haggard, P., & Fletcher, P. C. (2012). Exploring implicit and explicit
aspects of sense of agency. Consciousness and Cognition, 21, 17481753.
Moore, J. W., & Obhi, S. S. (2012). Intentional binding and the sense of agency: a review.
Morris, G., Nevet, A., Arkadir, D., Vaadia, E., & Bergman, H. (2006). Midbrain dopamine
neurons encode decisions for future action. Nature Neuroscience, 9, 10571063.
Nader. J. N., Claudia, C., Rawas, R. E., Favot, L., Jaber, N., Thiriet, N., & Solinas, M. (2012).
Loss of environmental enrichment increases vulnerability to cocaine addiction.
Neuropsychopharmacology, 37, 15791587.
Niv, Y., Joel, D., & Dayan, P. (2006). A normative perspective on motivation. Trends in Cognitive
Sciences, 10, 375381.
ODoherty, J. P. (2004). Reward representations and reward-related learning in the human

brain: insights from neuroimaging. Current Opinion Neurobiology, 14, 769776.
Oztop, E., Bradley, N. S. & Arbib, M. A. (2004). Infant grasp learning: a computational model.
Palmiter, R. D. (2008). Dopamine signaling in the dorsal striatum is essential for motivated
behaviors: lessons from dopamine-deficient mice. Annals of the New York Academy of Sciences,
1129, 3546.
Patall, E. A., Cooper, H., & Robinson, J. C. (2008). The effects of choice on intrinsic motivation
and related outcomes: a meta-analysis of research findings. Psychological Bulletin, 134, 270
300.
Powell, S. B., Newman, H. A., Pendergast, J. F., & Lewis, M. H. (1999). A rodent model of
spontaneous stereotypy: initial characterization of developmental, environmental, and
neurobiological factors. Physiology & Behavior, 66, 355363.
Prinz, W. (1987). Ideo-motor action. In H. Heuer & A. F. Sanders. (Eds.), Perspectives on

perception and action. Hillsdale, NJ: Erlbaum Associates.
Redgrave, P., Gurney, K., & Reynolds, J. (2008). What is reinforced by phasic dopamine signals?
Brain Research Reviews, 58, 322339.
Redgrave, P., Prescott, T. J., & Gurney, K. (1999). The basal ganglia: a vertebrate solution to the
selection problem. Neuroscience, 89, 10091023.
Samejima, K., Ueda, Y., Doya, K., & Kimura, M. (2005). Representation of action-specic reward
values in the striatum. Science, 310, 13371340.
Page 20 of 22
Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journal of Neurophysiology,

80, 127.
Schultz, W. (2012). Updating dopamine reward signals. Current Opinion in Neurobiology, 23,
229238.
Seligman, M. E. (1972). Learned helplessness. Annual Review of Medicine, 23, 407412.
Shapiro, D. H., Schwartz, C. E., & Astin, J. (1996). Controlling ourselves, controlling our world:
psychology s role in understanding positive and negative consequences of seeking and gaining
control. American Psychologist, 51, 12131230.
Skinner, E. A. (2007). Secondary control critiqued: Is it secondary? Is it control? Comment on

Morling and Evered (2006). Psychological Bulletin, 133, 911916.
Stephens, J. M. (1934). The influence of punishment on learning. Journal of Experimental

Stoit, A. M. B., van Schie, H. T., Riem, M., Meulenbroek, R. G. J., Newman-Norlund, R. D.,
Slaats-Willemse, D. I. E., et al. (2011). Internal model decits impair joint action in children and
adolescents with autism spectrum disorders. Research in Autism Spectrum Disorders, 5, 1526
1537.
Sutton, R., & Barto, A. (1998). Reinforcement learning, an introduction. Cambridge: MIT Press/
Bradford Books.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: a social psychological perspective
on mental health. Psychological Bulletin, 103, 193210.
Thelen, E. (1980). Determinants of stereotyped behaviour in normal human infants. Ethology

and Sociobiology, 1, 141150.
Thelen, E. (1981). Kicking, rocking, and waving: contextual analysis of rhythmical stereotypies
in normal human infants. Animal Behavior, 29, 311.
Thirkettle, M., Walton, T., Shah, A., Gurney, K., Redgrave, P., and Stafford, T. (2013). The path
to learning: action acquisition is impaired when visual reinforcement signals must first access
cortex. Behavioral Brain Research, 243, 267272.
Thorndike, E. L. (1927). The law of effect. American Journal of Psychology, 39, 212222.
Tobler, P. N., Fiorillo, C. D., & Schultz, W. (2005). Adaptive coding of reward value by dopamine
neurons. Science, 307, 16421645.
Page 21 of 22
Tricomi, E. M., Delgado, M. R., & Fiez, J. A. (2004). Modulation of caudate activity by action
contingency. Neuron, 41, 281292.
Vallacher, R. R., & Wegner, D. M. (1987). What do people think theyre doing? Action
identification and human behavior. Psychological Review, 94, 315.
Walton, T., Thirkettle, M., Redgrave, P., Gurney, K. N., & Stafford, T. (2013). The discovery of
novel actions is affected by very brief reinforcement delays and reinforcement modality. Journal
of Motor Behavior, 45, 351360.
White, R. W. (1959). Motivation reconsidered: the concept of competence. Psychological Review,

66, 297333.
Wolpert, D. M., Doya, K., & Kawato, M. (2003). A unifying computational framework for motor
control and social interaction. Philosophical Transactions of the Royal Society of London, Series
B, 358, 593602.
Page 22 of 22
Comparators and Weightings

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Neurocognitive Accounts of Agency
Matthis Synofzik
DOI:10.1093/acprof:oso/9780190267278.003.0013

Research on the neurocognitive underpinnings and disruptions of the sense of agency has been
greatly inspired by a simple computational account of agency: the comparator model. However,
this account faces several shortcomings in explaining agency in healthy and neuropsychiatric
subjects and in explaining the distinction between feelings (or experiences) of agency and
judgments (or beliefs) of agency. This chapter suggests that integration of agency cues might be
governed by a specific weighting mechanism, namely a weighting mechanism that follows the
principles of optimal cue integration. According to this notion, internal action-related cues,
external action-related cues, and action-independent cues are each constantly weighted
according to their reliability in a given situation. Also prior information on ones action is
incorporated, probably along Bayesian principles. Such a process would provide the basis for a
much more flexible and at the same time more robust mechanism than proposed by previous
neurocognitive accounts of agency.
Keywords: agency, comparator model, action, optimal cue integration, neuropsychiatry
Introduction
The experience of agency, that is, the registration that one is the initiator of ones actions, is a
basic and constant underpinning of our interaction with the world: whenever we grasp, type, or
walk, we register the resulting sensory consequences as caused by ourselves. In the last two
decades, several different accounts have been proposed to explain the neurocognitive
underpinnings of this experience. The probably most popular accountthe comparator model
Page 1 of 16
puts a strong emphasis on a particular mechanism for installing an experience of agency,

namely motor processes preceding the execution of ones action. In contrast, more recent
accounts emphasize not only one type of mechanism or a particular comparison process, but
various agency cues that operate in different time frames (e.g., predictive and post-dictive cues)
and on different levels (e.g., on sensorimotor and conceptual levels).
Here we discuss the advantages and shortcomings of these accounts. By further elaborating
these accounts, we will outline a neurocognitive account of agency that provides the basis for a
much more flexible and at the same time more robust mechanism than that proposed by
previous accounts. This novel account will stimulate manifold questions that lead to
experimentally directly testable hypotheses.
(p.290) The Comparator Model of Agency

The comparator model of agency (CM) has become one of the most popular accounts of the
experience of agency. This model elaborates and extends computational models of sensorimotor
integration (Sperry, 1950; von Holst, 1954; von Holst & Mittelstaedt, 1950; Wolpert,
Ghahramani, & Jordan, 1995) to the domain of agency processing. According to the CM
(Blakemore, Wolpert, & Frith, 2002; Frith, Blakemore, & Wolpert, 2000), an internal prediction
about the sensory consequences of ones actions is generated on the basis of an efference copy
of the motor command. These predicted sensory consequences can be compared with the actual
sensory state after that action has been initiated. If the actual sensory state matches the
predicted one, it is registered as self-caused. In case of a mismatch, it is registered as externally
caused (Figure 13.1). This account is not a purely predictive account of agency, as agency
registration here requires the sensory feedback of ones action (and thus also a post-dictive
component) for the comparison process. The predictive component, however, is the distinctive
feature of the CM: it is the crucial signal for deciding which sensory feedback is the result of
ones own actions and which is not.
Page 2 of 16
Shortcomings of the Comparator

Model
If applied to agency processing, the CM
faces several challenges and biological or
explanatory disadvantages (Synofzik,
Vosgerau, & Newen, 2008a; Synofzik,
Vosgerau, & Voss, 2013; Vosgerau &
Synofzik, 2012).
Ignorance of Non-sensorimotor Agency

Signals
The CM builds on internal predictions and
immediate sensory feedback (e.g., Figure 13.1 The comparator model of the
proprioceptive or visual) as its essential experience of agency. On the basis of an
efference copy of a given motor command,
signals. However, it does not tell us whether the sensorimotor system predicts the
and how a variety of other action-related outcome of ones own behavior. These
information is integrated into agency predictions can be compared to the actual
outcome of the movement. If the predicted
experience and belief. In particular, it
and the actual sensory feedback match, the
cannot provide any account of whether and movement is registered as self-generated. If
how non-sensorimotor cues, such as there is a discrepancy between predicted
background beliefs, ad hoc reasoning, or and actual feedback, the movement is
registered as externally caused.
affective valences of the action outcome, are
integrated in agency processing.
The neglect of affective cues might be

particularly problematic. A recent study by Wilke and colleagues showed that the immediate
non-conceptual perception of ones actions isin addition to the comparison between internal
predictions and sensory feedbackalso determined by the affective valence of post hoc action
outcomes (Wilke, Synofzik, & Lindner, 2012). Thus, the output of the comparator model cannot
fully explain the direct non-conceptual perception of ones actions.
(p.291) In fact, there is evidence that emotion is operative at all levels of the action control
hierarchy (Reis & Gray, 2009). Thus, we should expect emotion to be relevant at different stages
of agency processing, and an approach that integrates emotions into a model of agency seems
not only possible, but obligatory.
Experiences of Agency Without Efference Copy-Based Internal Predictions

The CM has difficulties in explaining instances of agency experiences in contexts where no
precise efference copy and thus no subsequent precise efference copy-based internal predictions
are issued. For example, in a helping hands pantomime task, subjects experienced high
degrees of agency for movements that were performed by another agent, when the other agents
hands appeared in the place where subjects hands would normally appear and when subjects
could hear instructions previewing each movement (Wegner, Sparrow, & Winerman, 2004).
Since subjects own arms remained passive, there was most (p.292) plausibly no precise
efference copy tied to ones motor command that could be used for a specific and detailed
prediction about the upcoming event (but, if at all, only a general cognitive anticipatory or
Page 3 of 16
intentional state). This finding demonstrates that efference copy-based internal predictions
(which are only issued in case of active movements) are not necessary to induce an experience
of agency; external cues (here: externally provided prior instructions) can substitute for it. Thus,
a comparator processing might, at least in some instances, not be necessary for the experience
of agency.
One might try to argue that in the case of passive viewing of ones movements, the subjects
perform some kind of motor simulation, and that this simulation might explain why subjects
experience agency for the observed movements. This might or might not be the case, yet it
would require a broad extension of the CM account. One would have to postulate that even in
instances of motor simulation, precise efference copies are issued, on the basis of which one
could exactly predict the upcoming external movements. But why and how should one issue
efference copies if no motor commands are issued? And how could the efference copy-based
predictions about the actions of another agent be so precise that they exactly match the sensory
properties of these actions? Efference copies must work in a highly precise way to prevent
constant ambiguity about ones actions in everyday life, and it has been shown that indeed work
in a highly precise temporally and spatially tuned way (e.g., sensory attenuation of self-produced
tactile hand stimulation functions only in time windows of 300 ms; Bays, Wolpert, & Flanagan,
2005; Blakemore, Goodbody, & Wolpert, 1998).
Alternatively, one might try to develop an adjusted CM account that does not require efference
copies and efference copy-based predictions as an input signal. In this case, however, the CM
account would loose its distinctive feature: the comparison between efference copy-based
predictions and sensory feedback of ones movements.
The Need for Constant Adjustments

The CM account might explain some instances of the experience of agency, but needs various
adjustments for many other instances (Carruthers, 2012; Vosgerau & Synofzik, 2012). For
example, as outlined in the preceding paragraph, one might try to tackle the difficult task of
adjusting the CM to incorporate also motor simulation without losing the key elements of the
CM. Another example are priming studies. Here, the amount of modification to the
[comparator] model needed is becoming incredibly large and none of these modifications is
predicted by the initial [comparator] model (Carruthers, 2012, p. 43). It remains very
questionable whether it is indeed possible to integrate all (p.293) different adjustments into a
coherently adjusted comparator model. Moreover, the CM also does not specify a number of
problems, thus making various different adjustments possible and necessary, which can no
longer be extrapolated from the comparator model itself (Vosgerau & Synofzik, 2012). So one
might ask why one should try to adjust the CM in all different ways and hold on to this account,
rather than simply developing a novel account right away.
No Explanation for the Temporary Nature and Semantic Content of Misguided Agency Beliefs
One of the main initial aims of the CM was to explain delusions of influence in schizophrenia
(Frith, 1992). Indeed, different experimental methodologies have provided experimental support
for this model as a basis of delusions of influence (Synofzik, Thier, Leube, Schlotterbeck, &
Lindner, 2010; Voss et al., 2010). Nevertheless, this model fails to explain key aspects of these
delusions: it can explain neither the temporary instantiation of these delusions (why do they
Page 4 of 16
suddenly evolve? why do they only last for a certain period of time? why do they stop at certain
points in time?) nor their semantic content (why do delusions refer only to specific content, and
not to all actions alike? why do they refer to a very specific content and not to another content
that is very similar? [Synofzik et al., 2008a]).
The Difference Between Feelings and Judgments of Agency

Within the sense of agency, two levels must be distinguished: the feeling of agency, which
consists of a non-conceptual, automatic registration of whether I am the agent or not, and the
judgment of agency, which is the formation of a belief about who the initiator of the movement
was (Synofzik et al., 2008a; Synofzik, Vosgerau, & Newen, 2008b; for a partly different
distinction between two levels within the sense of agency see Bayne & Pacherie, 2007). The
automatic registration on the level of feeling can lead to the perception of a particular action or
sensory event as self-caused. Subsequently and based on this feeling, a judgment might be
established (depending on the demands of the context) that takes into account not only the
feeling itself but also context information, background beliefs, general social norms, and so on.
Although this step is one of judgment formation, it does not necessarily need to be conscious or
to involve complex inferences or active reasoning. In most everyday cases, this step will be
effortless, automatic, and quick. Moreover, in some instances the feeling of agency might be so
strong and irresistible that it directly triggers and constrains the corresponding agency
judgment (Vosgerau & Synofzik, 2012).
(p.294) The comparator model does not respect this distinction and cannot account for it.
Within certain limitations (see previous sections), it might explain aspects of the basic, non-
conceptual feeling of agency. For example, if the output of the comparator is effectively 0, then
the signal might usually be registered as self-generated on this level (if not outweighed by other
factors that also influence the non-conceptual feeling of agency, e.g. affective cues; Wilke et al.,
2012). In turn, if the output is non-zero, then the signal might usually be registered as externally
generated. Yet, in both instances, the output can be complemented, outweighed, and overruled
on the level of judgment formation. This is due to the fact that this attribution depends not only
on sensorimotor processes (including, inter alia, the comparator output), but also on context
cues, background beliefs, and post hoc inferences (Synofzik et al., 2008a).
In other words, the comparator output can certainly support a me-judgment (in case of a zero
output) or a not-me-judgment (in case of a non-zero output). But in both instances the final
weight of this signal depends on additional non-sensorimotor cues. This makes an external
action attribution possible even if the comparator output is zero, and a self-action attribution
possible even if the comparator output is non-zero (Synofzik et al., 2008a).
Weighting Models
Chris Frith pointed out that the comparator model in its original version needs to be replaced
by a model with a much greater degree of sophistication and specificity (Frith, 2012). And,
indeed, some of the shortcomings of the comparator model have been resolved by more recent
accounts that start off from a more complex notion of the sense of agency, emphasizing an
integration of various agency cues, which operate on different levels (Bayne & Pacherie, 2007;
Fletcher & Frith, 2009; Moore & Fletcher, 2012; Moore, Lagnado, Deal, & Haggard, 2009;
Moore, Wegner, & Haggard, 2009; Synofzik et al., 2008a; Wegner & Sparrow, 2004). Unlike the
Page 5 of 16
comparator model, these accounts do not restrict agency processing to only a particular subset
of cues (e.g., efference copy-based internal predictions and immediate sensory feedback), but
consider various action-related and action-independent cues. They also do not need to postulate
the existence of several comparators in order to solve the problem of intergrating multiple cues
for establishing a harmonious experience or judgment of agency (evoking the problem of how to
combine the different comparator outputs).
According to the prominent multifactorial weighting model (MWM; Synofzik et al., 2008a), the
feeling and the judgment of agency result from a flexible integration process of a large variety of
cues. These cues include internal action-related cues (e.g., internal predictions/the comparator
model (p.295) output), external action-related cues (e.g., visual action consequences, affective
valences of the action outcome) and action-independent cues (e.g., background thoughts, social
cues), which can all be integrated to influence both the feeling of agency and the judgment of
agency.
The following account emerges. The feeling of agencywhich operates on a non-conceptual

sensorimotor levelwould be largely determined by predictive cues in a sensorimotor format.
The brain probably employs manifold prediction mechanisms in parallel, which complement and
substitute one another. They might be recruited in different situations and contexts, depending
on the information signals that are available in a given situation and on the requirements that
need to be met in a certain situation (Hohwy, 2013). These predictive mechanisms include, for
example, efference copies, internal predictions based on an efferency copy (Frith et al., 2000),
or sensorimotor predictions based on automatic associations (e.g., through subliminal priming
priming; Aarts, Custers, & Wegner, 2005; Wegner, 2003; Wegner et al., 2004). Some
sensorimotor predictive cues can be modulated by cognitive cues like background beliefs or
knowledge about the world. For example, motor processing or sensorimotor predictions can by
influenced by autosuggestion or through supraliminal priming (Aarts et al., 2005; Wegner et al.,
2004) or through prior causal beliefs induced by contextual information (Desantis, Roussel, &
Waszak, 2011). In addition, the feeling of agency is determined by post-dictive cues, again
mainly in a sensorimotor format. These post-dictive cues include mainly the visual feedback of
ones action (Synofzik et al., 2010) and feedback in other sensory modalities (e.g.,
proprioception).
On the conceptual cognitive level, a judgment of agency is formed. This is largely based on the
feeling of agency, but also takes into account cognitive cues like background beliefs and
information about the environment (e.g., the post hoc observation that I am the only person in
the room (cf. de Vignemont & Fourneret, 2004).
At both levelsthe level of feeling and the level of judgment of agencythe multifactorial
weighting process can be modulated by affective components. Two studies using very different
paradigmsnamely a visual distortion paradigm and an intentional binding paradigmhave
already provided complementary evidence that signatures of the feeling of agency are
modulated by the affective valence of an action outcome (Wilke et al., 2012; Yoshie & Haggard,
2013). Future work has to investigate whether this is also true for other aspects of affect, for
example the motivational value of an action (salience) or the subjective phenomenology of a
feeling associated with a particular action. Moreover, future work has to unravel how affect
Page 6 of 16
influences agency processing, and in particular how these two levels of the MWM might be
differentially influenced by different affective components.
(p.296) Both levels of agency processing are also influenced by the context and the
environment (Synofzik et al., 2013). For example, the context can directly influence the
weighting of sensory cues; for example, lighting conditions influence the reliability of a visual
cue. And cognitive representations about the environment can influence the formation of agency
judgments; for example, one might tend to self-attribute to ones own agency those events that
occur in ones own private wellness room, rather than those events that occur in a foreign or
hostile novel environment.
Shortcomings of Recent Weighting Accounts

Although recent weighting accounts provide several advantages compared to the classic
comparator model, they still face various shortcomings.
Indetermination of Agency Cues

The weighting accounts do not provide criteria and details regarding which cues are integrated
and which are not. In other words, it remains undetermined which cues are agency cues and
which are not. In fact, without criteria, it almost appears arbitrary what we count as an agency
cue and what we do not. It is likely that not all signals that are perceived and computed by our
sensorimotor and cognitive systems are integrated in the agency process. Recent models fail to
provide an explanation for this distinction and the underlying processes that serve to draw it.
Indetermination of the Cue Integration Process

Most recent weighting models do not provide evidence for how the different agency cues are
combined to shape agency feelings, agency judgments, and their content. For example, by which
equation are efference copy information and feedback signals intergrated and weighted? It
remains even more speculative how emotions, background thoughts, and ad hoc reasoning
processes are integratedand how they interact with sensorimotor cues. Likewise, the exact
nature of the integration mechanism has not yet been studied. For example, the integration
mechanism might follow a winner takes all principle, highlighting just one dominant agency
cue (e.g., ones efference copies or, alternatively, ones background beliefs). Alternatively,
integration might be done in a weighted fashionbut according to which principle would this
weighting then occur?
(p.297) Optimal Cue Integration: Combining Agency Cues According to Their

Reliability
A new proposal has emerged, which fills some (albeit not all) of the explanatory gaps of the
weighting models proposed thus far. In particular, it proposes a mechanism of how the
weighting might occur. According to the framework of optimal cue integration, the brain
constantly integrates several different authorship cues and weights, each cue according to its
relative reliability in a given situation (Synofzik et al., 2010; Synofzik, Vosgerau, & Lindner,
2009; Synofzik & Voss, 2010). The reliability of a cue would be low if its variance is high or if it
is completely absent; in turn, its reliability would be high if it is present in a highly precise or
very salient way. This notion follows the framework of optimal cue integration established in the
field of object perception: according to this framework, no single information signal is powerful
Page 7 of 16
enough to convey an adequate representation of a certain perceptual entity under all everyday
conditions. Instead, depending on the availability and reliability of a certain information cue,
different combination and integration strategies should be used to frame the weighting of
sensory and motor signals. Usually, predictive efferent signals such as internal predictions serve
as the most reliable and robust agency cues, as they usually provide the fastest and least noisy
information about ones own actions (Wolpert & Flanagan, 2001). However, in some situations
and subjects, other cues might outweigh or even replace these efferent signals to install a basic
registration of agency. For example, if predictive cues like internal predictions are weak or
imprecise, post hoc cues like the action feedback or the affective action outcome should receive
a higher weight for determining ones experience of agency. In other words, the variance within
one agency cue should be directly related to the reliance on another. Thus, optimal cue
integration might not only allow robust perception of objects and the world (Ernst & Banks,
2002; Ernst & Bulthoff, 2004) and efficient sensorimotor learning (Kording & Wolpert, 2004); it
could also provide the basis for subjects robust, and at the same time flexible, agency
experience in variable contexts (Moore & Fletcher, 2012; Synofzik et al., 2009; Synofzik & Voss,
2010).
So far, experimental evidence supporting the notion of optimal cue integration as the
neurocognitive key principle underlying agency processing is still rare and only indirect.
Nevertheless, this account already provides a unified and parsimonious framework for many
heterogeneous and so far unconnected findings from recent studies of agency. For example, it
connects agency studies using priming methods and agency studies focusing primarily on
efference copy mechanisms. According to the optimal cue integration approach, both cues can
be seen as priors (though operating on different levels) that (p.298) would determine the
feeling of agency according to their reliability in a given situation. A study by Moore, Wegner,
and Haggard (Moore, Wegner, et al., 2009; Synofzik et al., 2009) supports this notion. This study
used a combination of an intentional binding and a priming paradigm and showed that even
effects of passive movements are adopted to ones own agencyif consistent primes are
available. Although the interpretation of the intentional binding paradigm for agency studies has
recently been challengedsince intentional binding might be linked specifically to neither motor
predictive processes (Desantis, Hughes, & Waszak, 2012; Hughes, Desantis, & Waszak, 2013)
nor agency (Buehner, 2012; Buehner & Humphreys, 2009; Dogge, Schaap, Custers, Wegner, &
Aarts, 2012), but rather to causality in generalthis finding preliminarily indicates that primes
and efference copy cues might be integrated according to their availability and reliability in a
given situation. Since internal predictions are not reliably available in the case of passive
movements (as they are only issued in case of active movements), primes become the more
reliable cue and, consequently, trigger an experience of agency.
Does Optimal Cue Integration Presuppose a Self or Self-Agency?

The hallmark of optimal integration models in perception is that the brain integrates and
combines different classes of informational content about a single source.1 But what would be
this single source in the case of agency? Moreover, might there be a danger of a petitio principii
in using optimal integration as a model of agency? It seems as if one first assumes a self, and
then analyzes how different classes of information are optimally integrated to describe the
action of this self.
Page 8 of 16
Optimal cue integration does indeed presuppose a certain core self as the common source of
different classes of information (like the CM also presupposes a certain self-related
representation for allowing a self-world distinction; Vosgerau & Newen, 2007). However,
making this presupposition is not problematic for the optimal cue integration approach. In its
essence, this very basic form of a self is not more than a simple contingency registration
mechanism that allows the detection of systematic contigencies between motor output and
sensory signals (Vosgerau & Newen, 2007). This registration of systematic contigencies
probably already starts in utero as a non-conceptual and momentary representation of action-
effect-couplings (Synofzik et al., 2008b). It is the very basis of the self-world distinction
(Vosgerau & Newen, 2007) and thus for any more elaborated form of the self. By detecting
systematic contingencies, sensory signals can be systematically learned as self-produced. These
self-produced sensory signals (reafferences) are then filtered out from the incoming sensory
(p.299) flow (von Holst, 1954; von Holst & Mittelstaedt, 1950), and they can be learned to be
systematically predicted given ones own motor output, thus serving, for example, as one of the
input signals of the comparator (Synofzik et al., 2008b). At the beginning, this self-world
distinction is still unstable and imprecise, but it already provides a first basis to learn more
systematic contingencies with other, novel information signals (cues). Throughout infancy and
even the entire life, the array of action-effect contingencies is continously extended and refined,
and different types and presentations of cues are learned to be systematically associated with
ones own actions. In other words, an increasing array of cues is internalized to ones agency
(Synofzik, Thier, & Lindner, 2006) and builds up ones agency cue system. Thus, in a nutshell,
the optimal cue integration approach does not run into a petitio principii as it presupposes only a
very basic self-world distinction, but not yet a cue integration system.
Optimal Cue Integration: The Neurocognitive Principle Underlying Disorders of

Agency?
The account of optimal cue integration might not only provide a unified framework to explain
thus far unconnected findings from different agency studies in healthy subjects. It might also
provide a parsimonious basis to explain various psychopathological disorders of agency (Moore
& Fletcher, 2012; Synofzik et al., 2010)2.
Schizophrenia patients with delusions of influence feel that their actions are no longer
controlled by themselves. Sometimes they not only experience their actions as not self-caused,
leading only to a vague and strange experience, but also attribute them to some specific other
agents (e.g., to a friend, neighbor, or the devil) (Frith, 1992). How can this experience be
explained by the optimal cue integration approach? Although several studies that argue for a
close link between delusions of influence and a deficit in internal motor predictions must be
interpreted with caution (Davies, Coltheart, Langdon, & Breen, 2001; Jeannerod, 1997; Synofzik
et al., 2008a), two recent studies using very different paradigmsnamely a visual distortion
paradigm and an intentional binding paradigmprovide complementary evidence that
schizophrenia patients might indeed show imprecise internal predictions about the sensory
consequences of their own actions (Synofzik et al., 2010; Voss et al., 2010). These studies also
show that this deficit correlates with the severity of the psychopathology: the higher the
imprecision in predicting the sensory consequences of ones own actions, the higher the score
for delusions of influence (Synofzik et al., 2010).
Page 9 of 16
Following the optimal cue integration approach, imprecise predictions should prompt the
perceptual system to rely more strongly on post hoc cues in (p.300) order to receive a more
reliable account of ones own actions. And indeed, the study by Synofzik and colleagues found
that schizophrenia patients relied more on post hoc information about their actions (in their
study: vision; Synofzik et al., 2010). Similarly, another study investigating schizophrenia
patients, as well a group of patients with a putative psychotic prodrome, showed that both
patient groups, compared to healthy individuals, relied more strongly on external additional
sensorimotor cues to agency in an ambiguous situation, where the reproduction of a drum-pad
sequence had to be judged with respect to self-agency (Hauser et al., 2011).
The approach of optimal cue integration might thus provide a common basis for the various
misattributions of agency in schizophrenia patients, including their episodic nature (Synofzik et
al., 2010; Synofzik & Voss, 2010). In schizophrenic patients with delusions of influence, internal
predictions about the sensory consequences of ones own actions could be frequently imprecise
and unreliable. Patients should therefore be prompted in certain situations to rely more on
(seemingly more reliable) alternative cues about self-action. These might be either post hoc
(e.g., vision, auditory input, affective valence of the action outcome, or post-dictive thoughts) or
predictive (e.g., prior sensorimotor expectations based on specific background beliefs or prior
emotional appraisal of the situation). The stronger weighting of these alternative cues could help
patients to avoid misattribution of agency for self-produced sensory events in the case of
imprecise internal action-related predictions. However, as a consequence of giving up the
usually most robust and reliable internal action information source, that is, internal predictions,
the sense of agency in psychotic patients is at constant risk of being misled by ad hoc events,
invading beliefs, and confusing emotions and evaluations. In other words, schizophrenia patients
would be at constant risk of becoming a slave to every environmental influence (Frith, 1994, p.
151) and to every affective and moral ad hoc evaluation. Different agency judgment errors may
result: patients might over-attribute external events to their own agency whenever these more
strongly weighted alternative agency cues are not veridical and misleading, as is the case in
delusions of reference (also referred to as megalomania). Conversely, if alternative cues are
temporarily not attended or unavailable, patients might fail to attribute self-produced sensory
events to their own agency and instead assume external causal forces (as is the case in delusions
of influence). A context-dependent weighted integration of imprecise internal predictions and
alternative agency cues may therefore reflect the basis of agency attribution errors in both
directions: over-attribution, as in delusions of reference/megalomania, and under-attribution, as
in delusions of influence (Synofzik et al., 2010; Synofzik & Voss, 2010).
(p.301) Agency attribution in patients with delusions of influence usually has a very specific
semantic content, differing from individual to individual (e.g., a delusional attribution of an
action to a particular neighbor, relative, or religious entity), and fails only episodically and only
in certain contexts. The cue integration approach might also provide an explaination for the
following features:
1. An imprecision in efferent action-related information leads generally to a fluctuating,

unreliable basis on which the sense of agency is built, prompting schizophrenia patients
to rely more on other alternative cues, which might be misleading in some situations.
Page 10 of 16
2. An altered weighting of affective cues and the well-established formal thought

disorders in schizophrenia will then lead to an unbalanced and disturbed integration of
different agency cues with a lack of coherency and consistency. Features of formal
thought disorder in schizophrenia patients that might be particularly detrimental in this
integration process include tangentiality, illogicality, loss of goal, deficits in probabilistic
reasoning, and a premature jumping to conclusions. Based on these deficits, patients
might give an inadequate probabilistic weight to each agency cue and reach conclusions
on the basis of significantly less evidence than healthy subjects and express more
confidence in their decisions (Fletcher & Frith, 2009). This might explain the clinical
observation that patients all too easily develop false beliefs, which they then hold with
great confidence and immunity to any counter evidence (Fletcher & Frith, 2009, p. 50).
3. This disturbed integration of different agency cues then leads to the formation of an
individual delusional belief, which results from an individuals weighting of cognitive and
affective cues in a particular situation and the individuals personal background beliefs
and history. As personal affects, background beliefs, and history differ from individual to
individual, the content of each persons delusional belief differs as well. Thus, the
account of optimal cue integration offers, for the first time, also an account for the
content of agency beliefsat least with respect to affectively laden agency beliefs. This
would also explain why delusions of control do mostly refer not to trivial, non-emotional
actions in daily life (e.g., brushing teeth or typing on a computer), but to very specific,
singular actions with high affective and/or moral value. Mostly, they refer to actions that
are morally and socially not acceptable or at least negatively connoted, for example,
causing an accident, hurting someone, or behaving inappropriately. Here the affective
and moral valence gains major influence on both the sensorimotor and the cognitive level
(which might lead to modulated predictions and perception, as well as to specific
negative beliefs), such that the action is consequently not attributed to ones own
agency.
It is important to note that the account of cue integration does not per se predict a higher
weighting of affective cues compared to other agency cues. (p.302) Prima facie, it considers
affect only as one of the different agency cues (in contrast to the CM, where affect has no role in
agency processing), without giving it a higher weight a priori. However, it does predict that if
affect has a high saliency or valence, it will have a strong weight in agency determination.
Some neuropsychiatric subject groupsfor example schizophrenia patientsmight be

particularly inclined to give a high weight to affective cues. This might be due to the fact that
they experience affects more intensely. In this case, the cue integration per se would function
correctly, as a heigher weighting of more intense cues reflects a correct weighting. Or it might
be due to the fact that they experience affects in a similar way to healthy persons, but
nevertheless give more weight to them. This would reflect a deficit in the cue-weighting process.
These two alternative hypotheses are directly testable in future experiments.
Conclusions and Outlook

At the current stage, experimental evidence for the optimal cue integration approach underlying
agency and its disorders is still rare and only indirect. Yet this framework stimulates a wide
Page 11 of 16
range of questions and hypotheses on agency processing in different subject groups that are
directly experimentally testable:
1. Agency processing resorts to a weighting of various agency cues operating on

different levels. But are they really integrated according to their relative reliability and
thus in a statistically optimal way?
2. How are action-independent cues like emotions, background thoughts, and ad hoc
reasoning processes integrated, and how do they precisely interact with sensorimotor
cues in framing agency experiences, agency beliefs, and its content?
3. Do disturbances of agency in neuropsychiatric diseases result from a deficit in one (or
several) of the agency cues (e.g., internal motor representations of the actions or
affective evaluation of the action outcome) or in the mechanism integrating these cues?
4. Schizophrenia patients with delusions of influence show imprecise internal motor
predictions, whichaccording to the principle of optimal cue integrationshould make
them rely more on action-independent factors, such as the affective outcome of an
action. This can be tested by experimental tasks that have already been tested in healthy
subjects and that were described in this chapter (Wilke et al., 2012; Yoshie & Haggard,
2013). Such studies should also test for a possible correlation between the degree of
imprecision in ones internal predictions and the weighting of the affective valence,
which would be expected according to the optimal cue integration approach. The
abnormally (p.303) weighted affective action outcome might then be falsely integrated
in patients belief-formation process given their formal thought disorders. If this was
true, a correlation between the weighting, the formal thought disorders, and the clinical
degree of delusions was to be expected. Experiments testing these notions would be very
elegant as they would provide, for the first time, a mechanistic explanation for both the
formation and the content of delusions in schizophrenia.
Acknowledgments
This work was supported by a research grant by the Volkswagen Stiftung (VW II/85158).
Notes
References
Aarts, H., Custers, R., & Wegner, D. (2005). On the inference of person authorship: Enhancing
experienced agency by priming effect information. Consciousness and Cognition, 14, 439458.
Bayne, T., & Pacherie, E. (2007). Narrators and comparators: the architecture of agentive self-
awareness. Synthese, 159, 475491.
Bays, P. M., Wolpert, D. M., & Flanagan, J. R. (2005). Perception of the consequences of self-
action is temporally tuned and event driven. Current Biology, 15(12), 11251128.
Page 12 of 16
Blakemore, S. J., Goodbody, S. J., & Wolpert, D. M. (1998). Predicting the consequences of our
own actions: the role of sensorimotor context estimation. Journal of Neuroscience, 18(18), 7511
7518.
Buehner, M. J. (2012). Understanding the past, predicting the future: causation, not intentional
action, is the root of temporal binding. Psychological Science, 23(12), 14901497. doi:
10.1177/0956797612444612
Buehner, M. J., & Humphreys, G. R. (2009). Causal binding of actions to their effects.
Psychological Science, 20(10), 12211228. doi: 10.1111/j.1467-9280.2009.02435.x
Carruthers, G. (2012). The case for the comparator model as an explanation of the sense of
agency and its breakdowns. Consciousness and Cognition, 21(1), 3045; discussion 5538.
Davies, M., Coltheart, M., Langdon, R., & Breen, N. (2001). Monothematic Delusions: Towards a
two-factor account. Philosophy, Psychiatry, and Psychology, 8(23), 133158.
de Vignemont, F., & Fourneret, P. (2004). The sense of agency: a philosophical and empirical
review of the Who system. Consciousness and Cognition, 13(1), 119.
Desantis, A., Hughes, G., & Waszak, F. (2012). Intentional binding is driven by the mere
presence of an action and not by motor prediction. PLoS One, 7(1), e29557. doi: 10.1371/
journal.pone.0029557
of agency. Consciousness and Cognition, 20(4), 12111220. doi: 10.1016/j.concog.2011.02.012
Dogge, M., Schaap, M., Custers, R., Wegner, D. M., & Aarts, H. (2012). When moving without
volition: implied self-causation enhances binding strength between involuntary actions and
effects. Consciousness and Cognition, 21(1), 501506. doi: 10.1016/j.concog.2011.10.014
Ernst, M. O., & Banks, M. S. (2002). Humans integrate visual and haptic information in a
statistically optimal fashion. Nature, 415(6870), 429433.
Ernst, M. O., & Bulthoff, H. H. (2004). Merging the senses into a robust percept. Trends in
Fletcher, P. C., & Frith, C. D. (2009). Perceiving is believing: a Bayesian approach to explaining
the positive symptoms of schizophrenia. Nature Review Neuroscience, 10(1), 4858.
Frith, C. (1992). The cognitive neuropsychology of schizophrenia. Hillsdale, NJ: Erlbaum

Associates.
Frith, C. (1994). Theory of mind in schizophrenia. In A. S. David & J. C. Cutting (Eds.), The
neuropsychology of schizophrenia (pp. 147161). Hillsdale, NJ: Erlbaum Associates.
Page 13 of 16
Frith, C. (2012). Explaining delusions of control: the comparator model 20 years on.
Consciousness and Cognition, 21(1), 5254. doi: 10.1016/j.concog.2011.06.010
control of action. Philosophical Transactions of the Royal Society of London B: Biological
Sciences, 355(1404), 17711788.
Hauser, M., Knoblich, G., Repp, B. H., Lautenschlager, M., Gallinat, J., Heinz, A., & Voss, M.
(2011). Altered sense of agency in schizophrenia and the putative psychotic prodrome.
Psychiatry Research, 186(23), 170176. doi: 10.1016/j.psychres.2010.08.003
Hohwy, J. (2013). The predictive mind. Oxford: Oxford University Press.
prediction. Psychological Bulletin, 139(1), 133151. doi: 10.1037/a0028566
Jeannerod, M. (1997). The cognitive neuroscience of action. Oxford: Blackwell.
Kording, K. P., & Wolpert, D. M. (2004). Bayesian integration in sensorimotor learning. Nature,
427(6971), 244247.
Moore, J. W., Lagnado, D., Deal, D. C., & Haggard, P. (2009). Feelings of control: contingency
determines experience of action. Cognition, 110(2), 279283. doi: 10.1016/j.cognition.
2008.11.006
external cues. Consciousness and Cognition, 18(4), 10561064.
Reis, D. L., & Gray, J. R. (2009). Affect and action control. In E. Morsella, J. A. Bargh, & P. M.
Gollwitzer (Eds.), Oxford handbook of human action (pp. 277297). Oxford: Oxford University
Press.
Sperry, R. (1950). Neural basis of the spontaneous optokinetic response produced by visual
inversion. Journal of Comparative and Physiological Psychology, 43, 482489.
ones actions. Brain, 133(Pt 1), 262271.
Synofzik, M., Thier, P., & Lindner, A. (2006). Internalizing agency of self-action: perception of
ones own hand movements depends on an adaptable prediction about the sensory action
outcome. Journal of Neurophysiology, 96(3), 15921601.
Page 14 of 16
Synofzik, M., Vosgerau, G., & Lindner, A. (2009). Me or not mean optimal integration of
agency cues? Consciousness and Cognition, 18(4), 10651068.
Synofzik, M., Vosgerau, G., & Newen, A. (2008a). Beyond the comparator model: a multifactorial
Synofzik, M., Vosgerau, G., & Newen, A. (2008b). I move, therefore I am: a new theoretical
framework to investigate agency and ownership. Consciousness and Cognition, 17(2), 411424.
Synofzik, M., Vosgerau, G., & Voss, M. (2013). The experience of agency: an interplay between
prediction and postdiction. Frontiers in Psychology, 4, 127. doi: 10.3389/fpsyg.2013.00127
Synofzik, M., & Voss, M. (2010). Disturbances of the sense of agency in schizophrenia. In M.
Balconi (Ed.), Neuropsychology of the sense of agency (pp. 145156). New York; Heidelberg:
Springer.
von Holst, E. (1954). Relations between the central nervous system and the peripheral organs.
British Journal of Animal Behavior, 2, 8994.
von Holst, E., & Mittelstaedt, H. (1950). Das Reafferenzprinzip. Naturwissenschaften, 37, 464
476.
Vosgerau, G., & Newen, A. (2007). Thoughts, motor actions, and the self. Mind and Language,
22, 2243.
Vosgerau, G., & Synofzik, M. (2012). Weighting models and weighting factors. Consciousness
Voss, M., Moore, J., Hauser, M., Gallinat, J., Heinz, A., & Haggard, P. (2010). Altered awareness
of action in schizophrenia: a specific deficit in predicting action consequences. Brain, 133(10),
31043112.
Wegner, D. (2003). The mind`s best trick: How we experience conscious will. Trends in
Wegner, D., & Sparrow, B. (2004). Authorship processing. In M. S. Gazzaniga (Ed.), The new
cognitive neurosciences (3rd ed., pp. 12011209). Cambridge, MA: MIT Press.
Wegner, D. M., Sparrow, B., & Winerman, L. (2004). Vicarious agency: experiencing control
over the movements of others. Journal of Personality and Social Psychology, 86(6), 838848.
Wilke, C., Synofzik, M., & Lindner, A. (2012). The valence of action outcomes modulates the
perception of ones actions. Consciousness and Cognition, 21(1), 1829.
Wolpert, D. M., & Flanagan, J. R. (2001). Motor prediction. Current Biology, 11(18), R729R732.
Page 15 of 16
Yoshie, M., & Haggard, P. (2013). Negative emotional outcomes attenuate sense of agency over
voluntary actions. Current Biology, 23(20), 20282032. doi: 10.1016/j.cub.2013.08.034
Notes:
(1) I am indebted to Patrick Haggard (London, UK) and Gottfried Vosgerau (Dsseldorf,
Germany) for raising and discussing this issue.
(2) This section closely follows an argument outlined by us elsewhere (Synofzik et al., 2013).
Page 16 of 16
Action Control and the Sense of Agency

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001

Bernhard Hommel
DOI:10.1093/acprof:oso/9780190267278.003.0014

Peoples behavior is often impossible to predict and to understand based on the stimulus
condition and the environment in which the behavior occurs. People carry out very different
actions in response to the same stimuli, and respond to very different stimuli by means of the
same action, if they only have reasons to do so or a goal that has motivated and is satisfied by
that action. Planning and executing goal-directed behavior requires knowledge about means-
ends relationships (i.e. which goals can be reached by means of which actions/movements),
which implies knowledge about which goals can be reached by the agent him- or herself and
which cannot. In the literature, the latter kind of knowledge has been assumed to emerge from a
so-called sense of agency. Unfortunately, the term is rather ill-defined and used differently by
different authors, creating considerable confusion with respect to at least three different,
relatively orthogonal dimensions.
Keywords: environment, stimulus condition, goals, actions, sense of agency
The Sense of Agency

Peoples behavior is often impossible to predict or to understand based on the stimulus condition
and the environment in which the behavior occurs. We can carry out very different actions in
response to the same stimuli, and respond to very different stimuli by means of the same action,
if we only have reasons to do so (to refer to the personal level of description) or a goal that has
motivated and is satisfied by that action (to refer to the functional level of description).
Therefore, people are commonly assumed to be in control of their actions, so that their actions
often reflect their intentionswhich again allows some degree of mind reading in social
encounters (e.g., Heider & Simmel, 1944).
Page 1 of 17
Planning and executing goal-directed behavior require knowledge about means-ends

relationships (i.e., about which goals can be reached by means of which actions/movements).
That is, the goal-directed agent needs to have knowledge about which goals can be reached by
him- or herself and which cannot. In the literature, this kind of knowledge has been assumed to
emerge from a so-called sense of agency. Unfortunately, the term is rather ill-defined and is
used differently by different authors (for recent overviews, see Gallagher, 2012; Haggard &
Tsakiris, 2009). This has created considerable confusion with respect to at least three different,
relatively orthogonal dimensions.
(p.308) First, authors tend to confuse objective agency, the question of whether a given
individual was actually producing a particular action, with subjective (or perceived) agency,
that is, with the question of whether the agent or non-agent is actually sensing, experiencing, or
reporting to have some sort of authorship. For instance, Haggard and Tsakiris (2009) discuss
three empirical observations that provide strong evidence that objective agency matters (e.g.,
objectively self-produced events are perceived to be closer to ones action; see Haggard, Clark,
& Kalogeras, 2002) and take that to speak to the issue of subjective agency (which actually is not
assessed). In the following, I will mainly restrict myself to the discussion of how objective
agency operates (i.e., how self-performed actions are cognitively controlled) and which aspects
of these operations are likely to inform subjective/perceived agencywithout attempting to
provide a full-fledged account for the latter.
Second, in discussions of the sense of agency it often remains unclear what the concept of
sense is actually referring to. On the one hand, the term may be used the same way it applies
to vision, audition, and other sensory systems. These senses can be defined as physiological
capacities of organisms that provide data for perception (Wikipedia). According to this
definition, having a sense need not imply its proper use. For instance, one can easily imagine
that ones visual sensory system provides complete information about particular states of affairs
(e.g., that one is facing fresh powder snow rather than packed powder or crud), while the
perceiver makes very little use of that information (and simply perceives snow). From this
perspective, investigating the sense of agency should focus on the origin and availability of
information about whether it was the agent or someone else who carried out a particular action
irrespective of whether that information is actually picked up and used appropriately by that
agent (Synofzik, Vosgerau, & Neven, 2008). On the other hand, however, everyday use of
language often takes the term sense to imply some degree of sensing, so that having a
sense of agency would imply that an agent engages in some sort of perception related to his or
her agenthood. Indeed, some authors relate the term to the experience of controlling ones own
actions (Chambon & Haggard, 2013), which goes way beyond the mere availability of
information but implies its active and appropriate use for creating particular mind states. In the
following, I restrict myself mainly to the first use of the term and focus on the origin and
availability of agency information. One reason for that choice is that I will be discussing findings
from infant research suggesting that the availability of information about agency precedes the
use of this information for action control. This implies that requiring appropriate use of agency
information, as in studies asking for agency judgments, tends to underestimate the actual
availability of agency-relevant information. Another reason is that multiple sources for (p.309)
agency judgments are likely to exist (Synofzik et al., 2008), and it makes sense to assume that
people differ both intra- and inter-individually with respect to which sources of information are
Page 2 of 17
considered and how the different sources are weighted in making an agency judgment (Synofzik,
Chapter 13 of this volume).
Third, even though most authors seem to share some implicit agreement that it is individual
agents that are the attributional targets of the experience or the judgment of agency, this is
likely to reflect but a widely shared cultural bias. Indeed, most articles on agency restrict their
analysis to an I-perspective, on whether and how a single individual agent is able to perceive
him- or herself as being in control of his or her self-performed action. However, even though this
seems to be the obvious perspective for most readers with a Western background, members of
Eastern cultures tend to have a more extended perspective that includes family, peers, and
colleagues in the perception of agency (Markus & Kitayama, 2003). This is likely to be a
consequence of the wider definition of the self in Eastern as compared to Western cultures:
while in Westerners the borders of the perceived self coincide more or less with ones skin,
Easterners often have a more socially extended self-concept (Markus & Kitayama, 1991).
Accordingly, while the former commonly perceive some sort of individual agency, the latter will
often experience what Markus and Kitayama (2003, 2010) have called conjoint agency. If one
assumes that culture operates on cognition mainly by providing selective reward for a particular
cognitive style (Hommel & Colzato, 2010), one would expect that other kinds of social systems
that operate similarly can exert comparable effects. Indeed, there is evidence that Buddhists
(i.e., members of a religion that emphasizes social concern and de-emphasizes self-other
distinction) spontaneously relate their own action to the action of a co-actor more strongly than
culture-matched atheists do (Colzato et al., 2012). Interestingly, neither cultural background nor
religion seems to create fixed, hard-wired agency models, but rather implements default
biases toward one or the other alternative modelwhich leaves room for short-term adaptation.
For instance, participants are more likely to relate their own action to someone elses action
after having circled relational pronouns in a text (such as we, our, or us) than after having
circled pronouns emphasizing social independence (such as I, my, or me; Colzato, de
Bruijn & Hommel, 2012). These and other demonstrations of considerable inter- and intra-
individual variability in distinguishing between oneself and other agents (e.g., Hommel, Colzato,
& van den Wildenberg, 2009; Kuhbandner, Pekrun, & Maier, 2010); Khnen & Oyserman, 2002)
provide a substantial theoretical challenge for agency modelsin which the identity of the agent
is taken as a given. While I will not attempt to provide a comprehensive account for variability in
self-other discrimination (see Hommel & Colzato, 2010), I will briefly get back to its implications
below.
(p.310) Approaches to Agency

Theoretical treatments of the sense of agency commonly focus on the predictability of the
perceivable outcomes of actions. Carrying out a goal-directed action implies knowledge about
the expected outcome of the action, and this knowledge must play some role in judging whether
it is oneself or another person who actually created this outcome. Experimental analyses of
agency have been mainly motivated by four theoretical frameworks, which differ in emphasis
and explanatory goals (i.e., most were actually not developed to account for agency), but which
can nevertheless be considered functionally equivalent with respect to their agency-relevant
assumptions. In the following, I will briefly discuss these four frameworks in turn and then
explain where and in which sense they are functionally equivalent. Next, I will discuss how they
Page 3 of 17
can be integrated into one single model that might provide a solid basis for understanding the
relationship between action control and the sense of agency.
Associative Learning Models

Associative learning approaches to action control have focused on three concepts: the stimulus,
which is commonly assumed to trigger associated responses (after sufficient stimulus-response
learning); the response, which might be triggered by a stimulus or by the expectation of reward;
and the response outcome (e.g., Mackintosh, 1974). From this perspective, perceiving oneself to
be the agent of a given action should be a function of acquired response-outcome associations: if
performing a particular response has often produced a particular outcome, one would be likely
to perform that action again in order to create that outcome (if wanted) and to perceive oneself
to be the agent responsible for that outcome. This hypothesis has been confirmed in various
studies, which for instance show that the causality that agents perceive between their actions
and related outcomes increases with degree of contiguity (i.e., temporal and spatial proximity)
and contingency (i.e., constant covariation) between action and outcome (for overviews, see
Shanks & Dickinson, 1987; Wasserman, 1990; Young, 1995). Interestingly, these are the same
factors that determine the degree of action-outcome learning in higher and lower animals as
well (e.g., Meck, 1985; Rescorla, 1992; Urcuioli & DeMarse, 1996), which suggests that the
mechanisms providing the information underlying the sensing of action-effect causality are
cognitively rather undemanding.
Ideomotor Theory of Action Control

Whereas associative learning approaches tend to downplay cognitive factors, ideomotor theories
have the explicit goal to account for a role of intentions (p.311) (i.e., internal states that drive
the body toward reaching particular goals) in action control. Ideomotor theorizing has a long
and varied history (for extended reviews, see Prinz, 1987; Stock & Stock, 2004), but it is fair to
say that the most systematic theoretical treatments were developed by Lotze (1852), Harless
(1861), and James (1890). And yet, it took many decades of oblivion before the theory was
revived (Greenwald, 1970) and updated, modern versions were suggested (Hommel, Msseler,
Aschersleben, & Prinz, 2001a, 2001b). It is interesting to consider why this theoretical approach
has suffered from such a bad press. As I will explain, it assumes that mental states become
conditioned to patterns of the motor system, and that the resulting association is sufficient to
explain voluntary action. On the one hand, the consideration that mental states play a role in
action control must have provoked considerable resistance in the behaviorist movement that
dominated the research on action from the 1910s on. It is thus not surprising that influential
theorists like Thorndike (1913, p. 113) considered ideomotor theorizing a descendant of, the
crassest forms of imitative magic. On the other hand, the suggestion that something as central
to human self-understanding as voluntary action should emerge from mere associations must
have provoked fierce resistance in cognitive scientists. Indeed, the assumption that simple
associations may suffice to create voluntary action seemed so absurd to Miller, Galanter, and
Pribram (1960) that they caricatured the intellectual contribution of ideomotor thinking as
merely inventing the (sometimes used) hyphen between the words ideo and motor. This
difficult start notwithstanding, ideomotor theorizing has received ample empirical support, and
it has certainly benefited from the great interest in mirror neurons and other neural mechanisms
linking perception and action. In fact, computational models that were developed to capture the
Page 4 of 17
way mirror systems operate in humans (as that by Keysers & Perrett, 2004) can be considered
highly instructive reinventions of the ideomotor principle.
The basic problem the first ideomotor theorists aimed to tackle might be coined the riddle of
executive ignorance (to borrow the term from Turvey, 1977). We can carry out all sorts of
voluntary action whenever we want, and yet we know nothing at all about how we actually
achieve this. Indeed, asking an agent to describe how she carried out a particular action
commonly triggers one of two strategies: either she carries out the action on the spot and
describes her perceptions while doing it (suggesting the use of reafferent information), or she
tries to recall an earlier occasion on which she carried out that action and tries to remember the
reafferent information available back then (a strategy that is very close to Lotzes particular
consideration). In other words, agents do not seem to have any sort of privileged access to their
motoric means to execute actions, but rather refer to perceptual knowledge that any (p.312)
other attentive observer might have collected as well (apart from interoceptive simulation, that
only the agent herself could perceive). How can it be that this executive ignorance nevertheless
allows us to orchestrate all the motor processes necessary to carry out the action?
Figure 14.1 sketches the basic mechanism

that according to ideomotor theories
underlies voluntary action control. They
take as a given that activating particular
motor patterns (Motor codes in Harlesss
model and the M neuron in Jamess) leads
to the execution of overt movements, which
again creates reafferent stimulation
(activation of codes of the Sensorium in Figure 14.1 Harle (1861) and
James (1890) models of the ideomotor
Harlesss model and the kinesthetic K
mechanism (left and right, respectively).
neuron in Jamess). Frequent overlap of Reproduced from Harle (1861) and James
activation of motor neurons/codes and (1890), respectively (in the public domain).
reafferent neurons/codes induces an
association between the twoan early
example of what is now known as Hebbian
learning. Importantly, this association can include any kind of motor and sensory code, and it is
bidirectional, so that the sensory code becomes a mental cue (James, 1890) of the action. All
the agent needs to do to activate an action pattern would thus consist in reactivating the
representation of the sensory action outcome; this would prime the corresponding action pattern
and (ceteris paribus) drive it to execution. In other words, voluntary actions are carried out by
anticipating their sensory consequences.
As reviewed elsewhere (Hommel, 2009; Shin, Proctor & Capaldi, 2010), the ideomotor approach
has received ample empirical support, but it was not developed to address perceived agency. In
fact, James (1890) explicitly denies conscious access to outflowing (efferent) information, thus
leaving no direct information from action production about action production. The only
information that could be used to arrive at agency-related judgments arises from a comparison
between expected and actual outcomeas in the basic cybernetic control loop (Wiener, 1948)
or, as we will see below, in comparator models of action control. In terms of Harlesss model: if
Page 5 of 17
action A is selected by (p.313) anticipating (i.e., activating the internal code of) the sensory
action effect a but for some reason produces effect b, anticipation and outcome would differ,
which could be expected to create internal conflict. Such a conflict could provide important
information for the perception of agency. Indeed, after being exposed to regularities between
actions and sensory action effects, irregular (i.e., experience-incongruent) action effects induce
surprise (Verschoor, Spap, Biro, & Hommel, 2013) and a decreased sense of agency (Spengler,
von Cramon, & Brass, 2009), accompanied by electrophysiological indicators of internal conflict
(a so-called feedback-related negativity, NFB, which is commonly observed if agents are
informed to have committed an error; Band, van Steenbergen, Ridderinkhof, Falkenstein, &
Hommel, 2009).
Comparator Models of Action Control

Comparator models seek to use cybernetic control principles to provide a computationally
transparent account of action control. Particularly influential in the discussion of agency is the
model presented by Frith, Blakemore, and Wolpert (2000; Blakemore, Wolpert, & Frith, 2002)
shown in Figure 14.2. It translates the representation of a wanted outcome (a desired state)
into motor commands (which may be fine-tuned on-line through environmental information:
affordances). The reafferent information produced by the action is compared to outcome
expectations, which can be improved in the case of a mismatch, and to the desired state. This
latter comparison is essential for the perception of agency: the greater the mismatch, the less
agency is perceived.
As with ideomotor theory, the original

motivation for the comparator framework
was not to address agency but, rather, to
provide a general processing model for
action control. In contrast to ideomotor
theory, however, the comparator framework
targets processes rather than the cognitive
structures these processes operate on, and
it emphasizes the evaluation of the action
rather than its selectionthe main focus of
idemotor theory. Considering these
differences in design, ideomotor and
comparator approaches can be considered
complementary: the ideomotor approach
provides a cognitive architecture that
implies the processes that the comparator
Figure 14.2 Frith, Blakemore and Wolperts
approach explicitly suggests, and the (2000) comparator model of voluntary
comparator approach suggests operations action. Redrawn (modified) from Frith et al.
that imply a cognitive architecture similar (2000).
to that suggested by the ideomotor
approach. Considering this relationship, the
two approaches can be considered to be
more or less identical (cf., Chambon & Haggard, 2013): the desired state corresponds to the
activation of the intended action effect (e.g., of K in Jamess model); the translation into motor
Page 6 of 17
commands to the spreading of activation from action-effect representations to the associated

motor patterns (from K to (p.314) M in Jamess model); the estimated actual state to the
registration of the actual sensory action effects (the muscle-produced activation of K in Jamess
model); and the comparison between the estimated actual state to the desired state the
matching of, or competition between, the representations of the anticipated action effect and the
actually registered action effect (an aspect that the original version of ideomotor theory did not
devote much attention to). Accordingly, the comparator approach can be considered a
translation of the ideomotor control model into processing termsa translation that also
highlights the role of action evaluation.
Apparent Mental Causation Model

The main question that is driving the apparent mental causation model suggested by Wegner
(2003) relates to the connection between the experience of conscious will and action. While,
according to Wegner, human agents have the experience that their thought of a particular
action is causally responsible for (p.315) its execution (a process that ideomotor theories try to
explain), this connection should be considered an illusion. Wegner claims that voluntary actions
are actually driven by unconscious processes that are not further specified (the actual causal
path in his model), which also trigger conscious thoughts of the action in a direct or indirect
fashion. Given that performing an action takes time, the conscious thought will often precede
the action in time, which creates the illusion that it was the thought that created the action (see
Figure 14.3).
The apparent mental causation model

suggests that the degree of agency that
agents perceive depends on the match
between pre-actional thought and the
actually produced action effects. Indeed,
Wegner and Wheatley (1999) observed that
people experienced more agency for the
appearance of an object on a screen if they
were presented with a word describing that
object about half a second before the
appearance. Hence, correctly anticipating
action effects (i.e., thinking of them before
they appear) makes one believe that one has
Figure 14.3 Wegners (2003) model of
produced these effects oneself. apparent mental causation. Redrawn
(modified) from Wegner (2003).
Theoretical Integration
The four approaches to agency I have
discussed so far overlap to a substantial
degree, especially with respect to the crucial assumption that perceived agency (p.316) is, or
at least can be, informed by relating representations of expected action effects to
representations of actual action effects. Even the many assumptions that go beyond the issue of
agency seem to be compatible, suggesting that these four approaches merely highlight different
aspects and implications of the underlying cognitive machinery. And yet, there are a few
assumptions and implications of these approaches that are somewhat more difficult to reconcile,
Page 7 of 17
and it is these details, so I will argue, that are informative and helpful for building a more
comprehensive model of action control in general and the sense of agency in particular. I will
discuss the most important of these discrepancies in turn.
How Specific Are Action-Effect Representations?

While associative-learning approaches are notoriously silent with respect to internal
representations, and the apparent mental causation model is not overly specific with respect to
the original cause of both thoughts and actions, ideomotor theory and the comparator approach
both assume that voluntary action begins with a representation of the wanted effect. According
to ideomotor theory, an agent creates some sort of active representation of the intended
outcome, such as a visual image or a verbal description of the wanted effect. Effect
representations become associated with the motor patterns that have created the corresponding
effects in the past, which renders these effect representations retrieval cues of the motor part of
the action. If we assume that wanted effects are represented in the same format as the effects of
actions in the agents repertoire (Hommel et al., 2001a; Prinz, 1990), the identification of
intention-compatible actions boils down to a feature-matching process. As a result of this
process, action representations are primed (pre-selected or pre-activated) to the degree that
their expected outcome corresponds to the wanted outcome (i.e., to the degree that their
features overlap). Hence, provided sufficient practice and experience of the agent, the
ideomotor principle provides a computationally transparent and undemanding, but highly
effective means to prepare a relevance-ranked repertoire of suitable actions for every given
intention.
Ideomotor approaches are not particularly specific with respect to the amount of detail that
action-effect representations are likely to have. As they assume that action-effect knowledge is
created through association, driven by repeated experience, the amount of detail is unlikely to
be high. No two actions or action-effect experiences are identical, as they depend on, and are
modulated by, the context, the current body posture, and so on, suggesting that the information
integrated into action-effect representations is confined to the most relevant, invariant features
of an action. Interestingly, this is also (p.317) implied by the comparator model of Blakemore et
al. (2002). Note that the action outcome is determined by two factors: the desired state
(corresponding to the wanted action effect) and so-called affordances, that is, context-specific
environmental information necessary to fine-tune an action. Indeed, there is ample evidence that
cognitive action planning considers intention-relevant features only, such as the bottle to be
grasped in order to drink, but not the situational specifics, such as the precise landing position
or the kinematics of the approaching movementthese specifics are likely to be added on-line
through fast-acting but consciously inaccessible sensorimotor loops (Hommel, 2010; Milner &
Goodale, 1995). However, the comparator model assumes that this dually determined action
outcome is compared against the wanted outcome, which again is not informed by affordances.
Accordingly, the comparison would always result in some degree of mismatch, which should
tend to reduce perceived agency. To make the model realistic, one would thus need to assume
that the comparison is not precise enough to consider the modification of the action through
affordances. In other words, the comparison must relate relatively abstract representations of
wanted and actual effect, just as implied by the ideomotor approach. Indeed, pointing
movements have been shown to immediately adjust to small and unnoticed changes of the goal
Page 8 of 17
location (Prablanc & Plisson, 1990), suggesting that goal states do not contain high-resolution
spatial information.
If the main contribution of action control to the sense of agency refers to relatively abstract
information, this must create quite some degree of uncertainty with respect to agencyat least
in the absence of other, control-unrelated information (Synofzik et al., 2008). This explains why
agents can be fooled so easily when it comes to agency judgments, so that they compensate for
movement errors of limbs that are actually not their own (Nielsen, 1963) and claim authorship
for anticipated but objectively random events on a screen (Wegner & Wheatley, 1999).
Is Outcome Prediction Derived from Action Selection?

Comparator models assume that the representation of wanted action effects is translated into
two states, resulting in the activation of the motor program that is needed to produce the
wanted effect and in the representation of the expected outcome. In particular, the latter is
assumed to derive from the former: agents are claimed to first specify the appropriate motor
program and translate that back, as it were, into a sensory expectation. As ideomotor theory is
more interested in the role of action effects in selecting an action than in the monitoring of its
outcomes, it does not say much about how action effects are used to evaluate action outcomes.
One possible scenario (p.318) could be just like the one suggested by the comparator
approach: matching the image of the wanted effect against previously acquired action-effect
codes would lead to the selection of an appropriate action-effect/motor pattern compound, and
the action-effect part of this compound could be used as a reference to evaluate actually
produced effects. Even the apparent mental causation model could be interpreted along these
lines: the actual cause of the action (the motor program or action pattern) could inform the
actual cause of the thought (the action-effect expectation), which could then become conscious.
Unfortunately, however, there are reasons to assume that the actual mechanism is more
complex. For instance, Elsner and Hommel (2004) presented participants with key-pressing-
contingent auditory action effects. After an extended practice phase, participants were
presented with a new task that required key-pressing responses to auditory stimuli that were
the same as the previous action effects. As reported earlier (Elsner & Hommel, 2001),
participants were faster if the new key-tone mapping heeded the previous key-tone mapping;
that is, people were faster pressing a key to a tone that they had previously produced by
pressing that key. The size of this mapping effect was modulated by temporal contiguity (largest
effect with zero delay between key press and tone in the practice phase) and contingency
(largest effect with high correlations between key press and tone and/or high probability of tone
in the practice phase). In the contingency experiment, Elsner and Hommel also assessed the
participants perceived agency, the degree to which they felt that they produced the tones by
means of their key pressing. The outcome mirrored the mapping-effect profile, with strongest
perceived agency when the key-press-tone contingency and/or told probability was high.
However, correlation analyses did not reveal any relationship between these measures,
suggesting that they reflected different processes.
A similar conclusion is suggested by a recent infant study of Verschoor et al. (2013). The study
investigated oculomotor action-effect learning by presenting 7- and 12-month-olds (in addition to
adults) with tones that were contingent on the direction of horizontal saccades. In a test phase,
Page 9 of 17
participants were again to make saccades to left or right while task-irrelevant tones (the
previous action effects) were presented. Replicating the findings from manual studies, reaction
times were slower if the eye was moved into a direction that did not match the direction of the
saccades that had produced the current tone in the practice phase. This suggests that
participants had acquired bidirectional associations between saccade directions and the
particular tones, as predicted by ideomotor theory. Being presented with a tone would then tend
to reactivate the previously associated saccade, which would compete with the tendency to
perform a saccade in the opposite direction. Only 12-month-olds and adults (p.319) showed this
effect; 7-month-olds did not, again replicating findings from manual tasks (Verschoor, Weidema,
Biro, & Hommel, 2010).
Importantly, Verschoor et al. (2013) also measured task-evoked pupillary responses as an index
of surpriseprediction failure, that is. All three age groups showed evidence of surprise when
moving their eyes to a tone-incompatible location. That is, even though the 7-month-olds had not
yet acquired reliable associations between actions and action-effect representations, they did
make accurate predictions of action outcomes. This dissociation between action-effect
representation for action selection on the one hand and action-effect anticipation on the other
suggests that these two processes are independent and develop at different rates. And if we take
the ability to correctly predict action outcomes to be at the basis of, or at least provide strong
input to, the conscious perception of agency, it suggests that perceived agency is rather
independent of the causal connection between actions and the sensory outcomes they produce.
In other words, the degree of perceived agency seems to depend more on the accuracy of our
prediction than on our actual authorship for a given action. If so, this provides strong support
for Wegners (2003) claim that agency judgments are rather unrelated to actual action
production and do not provide privileged access to action-control operations.
What Is the Functional Role of Conscious Will?

Common sense has it that voluntary actions are initiated and accompanied by a conscious
representation of its goal (Hommel, 2007). Scientific investigations did not provide much
evidence for this apparently overly optimistic scenario, however, which led Wegner (2003) to
conceptualize conscious will as a functionally irrelevant byproduct of intentional action control.
Ideomotor theory originally presumed the conscious representation of the wanted outcome to be
the trigger of voluntary action; after all, it was developed to understand how this representation
can get muscles moving. However, more modern versions have dropped this implicit assumption
and do not leave any particular functional role for conscious experience (Hommel et al., 2001a).
Similarly, Frith et al. (2000; Blakemore et al., 2002) consider which action-related information
might be consciously accessible, but their comparator model does not require any particular
contribution of conscious representations.
Irrespective of this widespread disenchantment with consciousness, it is interesting to consider

which aspects of voluntary action control might be available to consciousness. The most obvious
candidate would be the representation of the wanted action effect, an assumption that James
(1890) would share. Indeed, the consciousness theory of Baars (1988) suggests that action goals
would be a prime candidate for conscious representation. The standard (p.320)
counterargument against this possibility is based on findings in the tradition of Libet, Wright,
and Gleason (1982), who reported physiological indicators of action preparation to produce
Page 10 of 17
measurable effects some hundreds of milliseconds before the agent even feels the urge to act. It
is this finding that motivated Wegner (2003) to distinguish between the true cause of voluntary
actions (which would produce the physiological markers observed by Libet et al.) and its
conscious representation. And yet, this argument is by no means watertight. Tasks in the
tradition of Libet and colleagues require participants to perform tens or hundreds of equivalent
actions in a row. Actions of that sort have been suspected to be automatized ahead of time by
authors as early as Exner (1879), and more recent studies have indeed revealed that task
instructions are automatically translated into mental sets that, in turn, enable more or less
automatic performance (Bargh, 1989; Hommel, 2000). If conscious action representations would
play a role in Libet-type experiments, they would be expected to occur while participants
prepare for the task but not while carrying it out. There is thus little reason to search for
functionally relevant conscious representations a few hundred milliseconds before the actual
action is carried out, and it is difficult to understand why Libets findings have played such a
dominant role in discussions of the role of conscious action representations and free will (e.g.,
Klemm, 2010).
Given the lack of evidence that humans can consciously access internal processes (rather than
the states they produce) directly, it is not very likely that the translation of action-effect
representations into motor activation is a reasonable candidate for conscious representation,
and the absence of suitable receptors for activity within the motor cortex makes that activity an
unlikely candidate as well. This leaves the perception of self-produced action effects. There is
some evidence that the representations of expected action-effects are attenuated in the process
of planning and executing the action (Blakemore, Wolpert & Frith, 1998). This suggests that the
expected action effects are more difficult to consciously perceive than non-expected stimuli, and
there is indeed evidence supporting this prediction, such as the you-cant-tickle-yourself
effect (Blakemore et al., 1998). Similar effects have been obtained for other modalities as well
(Weiss, Herwig, & Schtz-Bosbach, 2011), even though more research on this issue is needed.
Taken together, the available, very preliminary observations suggest that, if conscious
experience is related to action control in some systematic fashion at all, it most likely refers to
the anticipatory (i.e., pre-actional) representation of the wanted action effect. Even if that would
be correct, it would not necessarily mean that conscious experience plays a causal role in the
sense that preventing conscious experience would make voluntary actions impossible. And yet, it
may well be that under normal circumstances the wanted action effect is (p.321) always
available for conscious consideration, a kind of standard companion. As pointed out by Hommel
(2013), consciousness is commonly operationalized as communicability of the represented
information. While communicability does not seem to have much use for the online control of
action, it does allow informing other people about ones intended actions, instructing others to
carry out particular actions, or discussing the pros and cons of alternative actions. Hence,
communicability allows one to explain ones action to others and to relate it to theirs, thus
providing the opportunity for self-reflection and social impression management. As important as
these functions are, they do not seem to reflect immediate causation in action control. Thus,
there seem to be good reasons to consider the possibility that consciousness is more important
for the social communication of ones action rather than for their actual performance
(Masicampo & Baumeister, 2013).
Page 11 of 17
Conclusion
As I have argued, action control provides information about agency, not by providing copies of
any outcome of the action-selection process (as comparator models claim), but mainly by
specifying the expected perceptual consequences of a given actiona process that has been
neglected by ideomotor theories but focused on by comparator approaches and predictive-
coding accounts (Friston, 2012). This comparison between anticipated and actual action
outcomes is likely to provide input for judgments of agency, as indicated in Figure 14.4. (p.322)
Wanted action effects are likely to be familiar because something similar hasbeen previously
experienced (otherwise they could not be expected). This would imply that the feeling of
familiarity provides a cue for agency or that the feeling of unfamiliarity provides a cue for non-
agency. Indeed, the observation that the perception of an expected action effect is attenuated
(Blakemore et al., 1998) suggests that the expected effects are effectively nullified by
subtracting the expected from actual effects. If so, unexpected effects would attract quite some
attention and would surprise the agenta cue for non-agency.
These considerations do not exclude that

there are other cues as well, as Figure 14.4
indicates. For instance, Chambon and
Haggard (2013) have argued that the
fluency of action selection might provide
information about agency, which fits with
Wenke, Fleming, and Haggards (2010)
observation that facilitating action selection
through subliminal response-compatible Figure 14.4 An integrative model combining
cues produced a heightened sense of ideomotor action selection, predictive coding
of anticipated action outcomes, and action
control over the action effects that their evaluation driven by comparing predicted
responses generated. Under other and actual action outcomes. Note that
circumstances, perceived agency might be agency is informed by only one of the two
suspected to increase with the effort the main control operations (action evaluation
but not action selection) and that it does not
agent is exerting: agents are likely to feed back directly to action control. Also
perceive more agency when climbing Mount note that the outcome of action evaluation
Everest instead of passing a bump on the provides just part of the information agency
perception relies on.
street, and when making a difficult rather
than a simple and obvious decision. That is,
it will often be the lack of fluency that
informs us about agency. Indeed, participants perceive the temporal distance between their
action and an auditory action-contingent effect to be shorter if they judge this relationship while
performing an effort-demanding task (Demanet, Muhle-Karbe, Lynn, Blotenberg, & Brass, 2013).
And there are likely to be more sources for agency judgments, such as contextual plausibility,
past experience, and the agent-specific typicality of the action. Hence, action control provides
some input to agency judgments but certainly not all, and sometimes perhaps not even the most
important one (Synofzik et al., 2008).
It is also interesting to consider that none of the action-control models I have discussed leaves
any functional role for the perception or representation of agencywhich is why in Figure 14.4
Page 12 of 17
agency judgments do not feed back directly to action control. And indeed, if everything goes as
expected and if the agent gets what he or she wants, it does not seem to be particularly
important whether it is he or she who was actually responsible, or someone else instead.
Moreover, there are interesting cultural differences with respect to what counts as agent and as
oneself. While Western authors commonly refer to individual agents, often even without
justifying this theoretical choice, Easterners seem to carve the world in different ways. Indeed,
what counts as an agent and as a causal factor is open to interpretation. If, for instance, a speed
runner beats the world record, one can argue whether it is she who was the responsible agent
or an entire team including support staff, family members, and friendsthe people Eastern
sportsmen indeed tend to refer to when attributing responsibility for individual performance
(Markus & Kitayama, 2003). Considering this (p.323) kind of conjoint agency (Markus &
Kitayama, 2003) has considerable impact on our theoretical conception of perceived agency and
the kind of information it is likely to reflect. And yet, this is unlikely to affect the internal
organization of action control, as in the end it is the individual runners brain and muscles that
must realize the record-beating run. Accordingly, it makes a lot of sense that action control
proper is independent of perceived agency and the cultural context into which it is woven.
Acknowledgment
The preparation of this work was supported by the European Commission (EU Cognitive
Systems project ROBOHOW.COG; FP7-ICT-2011).
References
Baars, B. J. (1988). A cognitive theory of consciousness. New York: Cambridge University Press.
Band, G. P. H., van Steenbergen, H., Ridderinkhof, K. R., Falkenstein, M., & Hommel, B. (2009).
Action-effect negativity: irrelevant action effects are monitored like relevant feedback. Biological
Bargh, J. A. (1989). Conditional automaticity: varieties of automatic influence in social

perception and cognition. In J. S. Uleman, & J. A. Bargh (Eds.), Unintended thought (pp. 351).
London: Guilford Press.
Blakemore, S. J., Wolpert, D. M., & Frith, C. D. (1998). Central cancellation of self-produced
tickle sensation. Nature Neuroscience, 1, 635640.
Chambon V., & Haggard P. (2013). Premotor or ideomotor: how does the experience of action
come about? In W. Prinz, M. Beisert, & A. Herwig (eds.), Action science: foundations of an
emerging discipline (pp. 359380). Cambridge, MA: MIT Press.
Colzato, L.S., de Bruijn, E., & Hommel, B. (2012). Up to me or up to us? The impact of self-
construal priming on cognitive self-other integration. Frontiers in Psychology, 3, 341.
Page 13 of 17
Colzato, L. S., Zech, H., Hommel, B., Verdonschot, R., van den Wildenberg, W., & Hsieh, S.
(2012). Loving-kindness brings loving-kindness: the impact of Buddhism on cognitive self-other
integration. Psychonomic Bulletin & Review, 19, 541545.
Demanet, J., Muhle-Karbe, P. S., Lynn, M. T., Blotenberg, I., & Brass, M. (2013). Power to the
will: how exerting physical effort boosts the sense of agency. Cognition, 129, 574578.
Psychology: Human Perception and Performance, 27, 229240.
Elsner, B., & Hommel, B. (2004). Contiguity and contingency in the acquisition of action effects.
Psychological Research, 68, 138154.
Exner, S. (1879). Physiologie der Grosshirnrinde. In L. Hermann (Ed.), Handbuch der

Physiologie, 2. Band, 2. Theil (pp. 189350). Leipzig: Vogel.
Friston, K. (2012). Prediction, perception and agency. International Journal of Psychophysiology,

83, 248252.
control of action. Philosophical Transactions of the Royal Society of London B: Biological
Sciences, 355, 17711788.
Gallagher, S. (2012). Multiple aspects in the sense of agency. New Ideas in Psychology, 30, 15
31.
Greenwald, A. G. (1970). Sensory feedback mechanisms in performance control: with special

reference to the ideomotor mechanism. Psychological Review, 77, 7399.
Haggard, P., & Tsakiris, M. (2009). The experience of agency: Feelings, judgments, and
responsibility. Current Directions in Psychological Science, 18(4), 242246.
Harless, E. (1861). Der Apparat des Willens [The apparatus of will]. Zeitschrift fuer Philosophie
und philosophische Kritik, 38, 5073.
Heider, F., & Simmel, M. (1944). An experimental study of apparent behaviour. American
Journal of Psychology, 57, 243259.
Hommel, B. (2000). The prepared reflex: automaticity and control in stimulus-response

translation. In S. Monsell & J. Driver (Eds.), Control of cognitive processes: attention and
performance (Vol. XVIII, pp. 247273). Cambridge, MA: MIT Press.
Hommel, B. (2007). Consciousness and control: not identical twins. Journal of Consciousness
Studies, 14, 155176.
Page 14 of 17
Hommel, B. (2009). Action control according to TEC (theory of event coding). Psychological
Research, 73, 512526.
Hommel, B. (2010). Grounding attention in action control: the intentional control of selection. In
B. J. Bruya (Ed.), Effortless attention: a new perspective in the cognitive science of attention and
action (pp. 121140). Cambridge, MA: MIT Press.
Hommel, B. (2013). Dancing in the dark: no role for consciousness in action control. Frontiers in
Psychology, 4, 380.
Hommel, B., & Colzato, L. S. (2010). Religion as a control guide: on the impact of religion on
cognition. Zygon: Journal of Religion & Science, 45, 596604.
Hommel, B., Colzato, L. S., & van den Wildenberg, W. P. M. (2009). How social are task
representations? Psychological Science, 20, 794798.
Hommel, B., Msseler, J., Aschersleben, G., & Prinz, W. (2001a). The theory of event coding
(TEC): a framework for perception and action planning. Behavioral and Brain Sciences, 24, 849
878.
Hommel, B., Msseler, J., Aschersleben, G., & Prinz, W. (2001b). Codes and their vicissitudes.
Behavioral and Brain Sciences, 24, 910937.
James, W. (1890). The principles of psychology (Vol. 2). New York: Dover Publications.
Keysers, C., & Perrett, D. I. (2004). Demystifying social cognition: a Hebbian perspective. Trends
in Cognitive Sciences, 8, 501507.
Klemm, W. R. (2010). Free will debates: simple experiments are not so simple. Advances in
Cognitive Psychology, 6, 4765.
Kuhbandner, C., Pekrun, R., and Maier, M. A. (2010). The role of positive and negative affect in
the mirroring of other persons actions. Cognition & Emotion, 24, 11821190.
Libet, B., Wright, E. W., & Gleason, C. A. (1982). Readiness potentials preceding unrestricted
spontaneous and preplanned voluntary acts. Electroencephalography and Clinical
Lotze, R. H. (1852). Medicinische Psychologie oder die Physiologie der Seele. Leipzig:
Mackintosh, N. J. (1974). The psychology of animal learning. New York: Academic Press.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: implications for cognition, emotion,
and motivation. Psychological Review, 98, 224253.
Page 15 of 17
Markus, H. R., & Kitayama, S. (2003). Models of agency: sociocultural diversity in the
construction of action. In V. M. Berman & J. J. Berman (Eds.), Nebraska symposium on
motivation: crosscultural differences in perspectives on the self (Vol. 49, pp. 158). Lincoln:
University of Nebraska Press.
Markus, H. R., & Kitayama, S. (2010). Cultures and selves: a cycle of mutual constitution.
Perspectives on Psychological Science, 5, 420430.
Masicampo, E. J., & Baumeister, R. F. (2013). Conscious thought does not guide moment-to-
moment actionsit serves social and cultural functions. Frontiers in Psychology, 4, 478.
Meck, W. H. (1985). Postreinforcement signal processing. Journal of Experimental Psychology:

Animal Behavior Processes, 11, 5270.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the structure of behavior. New
York: Holt, Rinehart & Winston.
Milner, A. D., & Goodale, M. A. (1995). The visual brain in action. Oxford: Oxford University
Press.
Nielsen, T. I. (1963). Volition: a new experimental approach. Scandinavian Journal of

Prablanc, C., & Plisson, D. (1990). Gaze saccade orienting and hand pointing are locked to their
goal by quick internal loops. In M. Jeannerod (Ed.), Attention and performance (Vol. XIII, pp.
653676). Hillsdale: Erlbaum.
Prinz, W. (1987). Ideo-motor action. In H. Heuer & A. F. Sanders (Eds.), Perspectives on

perception and action. Hillsdale, NJ: Lawrence Erlbaum Associates.
Prinz, W. (1990). A common coding approach to perception and action. In O. Neumann & W.
Prinz (Eds.), Relationships between perception and action (pp. 167201). Berlin: Springer.
Rescorla, R. A. (1992). Response-outcome vs. outcome-response associations in instrumental

learning. Animal Learning and Behavior, 20, 223232.
Shanks, D. R., & Dickinson, A. (1987). Associative accounts of causality judgment. Psychology of
Learning and Motivation, 21, 229261.
Shin, Y. K., Proctor, R. W., & Capaldi, E. J. (2010). A review of contemporary ideomotor theory.
Spengler, S., von Cramon, D. Y. & Brass, M. (2009). Was it me or was it you? How the sense of
agency originates from ideomotor learning revealed by fMRI. Neuroimage, 46, 290298.
Stock, A., & Stock, C. (2004). A short history of ideo-motor action. Psychological Research, 68,
176188.
Page 16 of 17
Thorndike, E. L. (1913). Ideo-motor action. Psychological Review, 20, 91106.
Turvey, M. T. (1977). Preliminaries to a theory of action with reference to vision. In R. Shaw & J.
Bransford (Eds.), Perceiving, acting and knowing: Toward an ecological psychology. Hillsdale,
NJ: Lawrence Erlbaum Associates.
Urcuioli, P. J., & DeMarse, T. (1996). Associative processes in differential outcome

discriminations. Journal of Experimental Psychology: Animal Behavior Processes, 22, 192204.
Verschoor, S. A., Spap, M., Biro, S., & Hommel, B. (2013). From outcome prediction to action
selection: Developmental change in the role of action-effect bindings. Developmental Science,
16, 801814.
Verschoor, S. A., Weidema, M., Biro, S., & Hommel, B. (2010). Where do action goals come
from? Evidence for spontaneous action-effect binding in infants. Frontiers in Cognition, 1, 201.
Wasserman, E. A. (1990). Detecting response-outcome relations: toward an understanding of the

causal texture of the environment. Psychology of Learning and Motivation, 26, 2782.
Weiss, C., Herwig, A., & Schtz-Bosbach, S. (2011). The self in action effects: selective
attenuation of self-generated sounds. Cognition, 121, 207218.
Wegner, D. M., & Wheatley, T. P. (1999). Apparent mental causation: sources of the experience
of will. American Psychologist, 54, 480492.
Wenke, D., Fleming, S.M., & Haggard, P. (2010). Subliminal action priming influences sense of
agency. Cognition, 115, 2638.
Wiener, N. (1948). Cybernetics, or communication and control in the animal and the machine.
Young, M. E. (1995). On the origin of personal causal theories. Psychonomic Bulletin & Review,
2, 83104.
Page 17 of 17
Control and Truth Working Together

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
The Agentic Experience of Going in the Right Direction
E. Tory Higgins
DOI:10.1093/acprof:oso/9780190267278.003.0015

The human life experience is mostly an experience of the process of goal pursuit rather than the
outcomes from goal attainment. As Carl Rogers said, the good life is a direction, not a
destination. In this way, peoples sense of agency during the process of goal pursuit is
fundamental to the good life. People experience a sense of agency when they manage to make
things happen, that is, when they are effective at control. But people can be effective at
managing to change from state to state regardless of the direction of that change. To manage
change in the direction of their goal, people also need to know what is the best or correct
direction. Control effectiveness, then, needs to work together with truth effectiveness. When
they do work together effectively, humans experience a special kind of sense of agencythe
experience of going in the right direction.
Keywords: agency, control, truth, effectiveness, goal attainment
Introduction
Let me begin with a classic storyperhaps the classic storyabout control effectiveness. It is
the epic story of Odysseus in Homers Odyssey. In his voyage to return home with his men,
Odysseus faced constant control challenges in managing what would happen to him and his
crew. The story is clearly not about being effective in fulfilling the goal of safely returning home
(value effectiveness) because by the end of the voyage not one member of his crew had survived.
Odysseus himself did return home, but he was in bad shape. Instead, the story of Odysseus is
about managing what is needed to make something happen and, especially in this story, to make
Page 1 of 15
something not happen. Perhaps the best-known episode in the story is the challenge of the
Sirens.
The Sirens were enchantresses who lived on a rocky island and used their music and voices to
lure mariners toward them, where they would be shipwrecked on the rocks. To manage his
sailors temptation, Odysseus had their ears stuffed with wax so they could not hear the music.
Odysseus could have used the same control tactic, but he was curious to hear the Sirens song.
It turned out that the Sirens words were even more enticing than their beautiful voices because
they promised to give great wisdom to whomever came to themthe great temptation of
learning the truth about the world (p.328) (truth effectiveness). Odysseus managed his strong
temptation for knowledge by having his followers lash him securely to the mast of the ship.
Odysseus, then, succeeded at controlling what happened. He experienced control effectiveness

a sense of agency. When people are effective at control, they manage what is required, such as
managing procedures, competencies, and resources, to make something happen or not happen.
Having control relates to exercising direction or restraint upon action, to having power or
authority to guide or manage. As illustrated in this story, experiencing control effectiveness, or a
sense of agency, can be distinct from being effective in goal fulfillment (value effectiveness).
As I discuss in more detail later in this chapter, value effectiveness relates to having desired (vs.
undesired) results or outcomes at the end of the goal pursuit, and truth effectiveness relates to
establishing what is real or right (vs. illusion or what is wrong). In contrast, control effectiveness
relates to having a strong versus weak influence over somethingmanaging to have an effect.
While control can increase the likelihood of beneficial results or outcomes, it is independent of
outcomes. Control is also independent of truth. Someone, for example, can have the truth
effectiveness of knowing that some future negative event is certain to happen (e.g., the falling
glass will land on the kitchen floor), without having any control over whether it happens. Thus,
experiencing control effectiveness, or a sense of agency, is not the same as experiencing value
effectiveness or truth effectiveness.
The major objective of this chapter is to review how control effectiveness, although distinct from
truth effectiveness, partners with truth functions in order to succeed (for a fuller review, see
Higgins, 2012). Although control and truth are independent, they work together as partners.
When they work together, people have the experience of going in the right direction. It is this
partnership between control functions and truth functions that merits more explicit emphasis in
the motivational literature. People can have a sense of agency by managing just to go
controlling changes of state in any direction. But by working together with truth effectiveness,
people with control effectiveness can experience going in the right direction. This special kind
of sense of agency is the focus of this chapter. As background for this discussion, I begin by
saying something more about how value, truth, and control are three different ways of being
effective in life pursuits.
Three Ways of Being Effective in Life Pursuits

Value Effectiveness
By value effectiveness I mean that actors are successful in ending ultimately with the
outcomes they desire. Value effectiveness is about success with (p.329) respect to outcomes,
Page 2 of 15
about the consequences of goal pursuitsuccess in ending with benefits versus costs, pleasure
versus pain, biological needs satisfied versus unsatisfied. Simply put, value effectiveness is being
successful in having what is desired. It should be emphasized that what matters for value
effectiveness is ending with the desired outcomesnot how this ending came about, whether
through a proxy, through collaboration with others, or through our own actions.
Value effectiveness was emphasized by drive theories (e.g., Hull, 1943, 1952) and the hedonic
principle. For drive theories, it was the value derived from the benefits of satisfying primary
biological needs, such as reducing hunger (e.g., finding food) or reducing fear (e.g., escaping
danger). For the hedonic principle, it was the value derived from maximizing pleasant outcomes
and minimizing painful outcomes. Goal theories have also emphasized value effectiveness, with
motivation constituting forces within us that are goal-directed or purposive (see Elliott & Dweck,
1988; Elliot & Fryer, 2008; Kruglanski et al., 2002; McDougall, 1914; Pervin, 1989). In social
psychology, at least, a major influence on this conceptualization of motivation was Kurt Lewins
work on goal-directed action and goal striving within a field of forces where positive value
relates to a force of attraction and negative valence relates to a force of repulsion (Lewin, 1935,
1951). Woodworth said it clearly: What persists, in purposive behavior, is the tendency towards
some end or goal. The purposeful person wants something he has not yet got, and is striving
towards some future result (Woodworth, 1921, p. 70).
Truth Effectiveness
By truth effectiveness I mean that actors are successful in knowing what is real. The root
meaning of truth (as well as trust) relates to true; truth is the quality of being true.
Something being true means being in accordance with an actual state of affairs, being
consistent with the facts; conforming to or agreeing with an essential reality; being that which is
the case, representing things as they arein brief, knowing whats real, whats reality (Oxford
English Dictionary, 1971). True also relates to accuracy; to being correct, right, and
legitimate; to being genuine, honest, and faithful. It is contrasted with being imaginary,
spurious, or counterfeit. Thus, truth effectiveness is being successful in establishing what is real.
Value effectivenesshaving desired resultsis critical for humans and other animals. But so is
truth effectivenessknowing what is real in the world, representing things as they are. Without
truth effectiveness we would bump into walls, we would live in a world that William James
([1890] 1948, p. 462) referred to as one great blooming, buzzing confusion.
(p.330) Young children, and sometimes adults as well, find it difficult to distinguish reality from
fantasy. Children may fear what is hiding in their closet, and some adults have paranoid
delusions. What is reality to one religious group is mere illusion or delusion to another. But what
is clear is that each individual and each group is strongly motivated to know what is realto
attain truth effectiveness. This plays out in various ways, including wanting to know what is
accurate, or what is correct or incorrect, right or wrong, legitimate or illegitimate, honest or
deceitful, genuine or fraudulent.
Given the dominant position of the hedonic principle within motivation, the difference between
truth effectiveness and the hedonic principle needs to be emphasized. It is common knowledge
that learning the truth about ourselves or those we care about can be painful, and yet we often
Page 3 of 15
seek the truth, even when we know it will be painful. When something pleasant but unexpected
happens to people, they often want to know why it happened. Others will tell them to just enjoy
it and not be concerned about why it happened. Yet, they still want to know the truth.
The movie The Truman Show illustrates this clearly. Trumans entire life is a reality TV show
that began airing at his birth. Everyone he knows are actors playing their roles. Trumans life on
the reality TV show is designed to be all pleasure and no pain. But there is one exception. He is
anxious about being on the water because he believes his father drowned in a boating accident.
However, like his entire life, this event was manufactured as part of the TV show. The actor who
played his father was simply a victim of his character being written out of the show. When,
accidentally, Truman sees him againalong with some other accidents on the showhe finally
realizes that his life has been manufactured. Everyone tries to reassure him, including the
executive producer of the show, who argues that there is no more truth in the real world than
there is in his own artificial one. Despite everyones reassurances and despite having a life of
pleasure and no pain, Truman risks actual death by sailing across the water he fears because he
needs to follow up his discovery and seek the truth.
The movie The Matrix provides another compelling example. In the movie, there is a future
where the reality perceived by humans is actually a simulated realitythe Matrixthat provides
people with a hedonically positive life to pacify them. Morpheus, the leader of the rebels, gives
Neo, the hero of the tale, a choice between a blue pill that will keep him in this comfortable
simulated reality or a red pill that offers only the truth. Morpheus tells Neo, All Im offering is
the truth, nothing more. Neo chooses the red pill. Neos motivation, like Trumans, is truth
effectiveness, which trumps the hedonic principle.
(p.331) And, of course, it is not just movie characters who illustrate the power of truth
effectiveness. Many people make life choices on the basis of their religious or political beliefs
about what is right or proper. All too frequently, individuals will give up their lives for truth, as
evidenced by suicide bombers and protesters who set themselves on fire.
Control Effectiveness
By control effectiveness, I mean actors experiencing success at managing what is required
(procedures, competencies, resources) to make something happen (or not happen). Having
control relates to exercising direction or restraint upon action; having power or authority to
guide or manage; having influence over something (Oxford English Dictionary, 1971). Control
effectiveness is being successful in managing what happens. Whereas value effectiveness relates
to outcomes (benefits versus costs) and truth effectiveness relates to reality (real versus
illusion), control effectiveness relates to strength (strong versus weak influence over something).
It is very general. People can have strong versus weak muscles, intellect, character, arguments,
willpower, teamwork, and so on. Managers, leaders, and administrators can be strong or weak.
While high control effectiveness increases the likelihood of beneficial outcomes, it is separate
from outcomes, as reflected in maxims such as its not whether you win or lose, it is how you
play the game and in victory or defeat, you play with skill and couragewith strength.
Indeed, control effectiveness can trump value effectiveness. Consider the phenomenon of
contra-freeloading (for a review, see Osborne, 1977). In one study showing this phenomenon,
Page 4 of 15
rats learned that by pressing a lever they could make a food pellet fall into a food tray where
they could eat it. In one experimental condition, a food dish was placed in the cage, which meant
that the rats could obtain the same food pellets for free (i.e., without having to work for them).
On occasion a rat would accidentally push the free food dish in front of the food tray. Despite the
fact that they could effortlessly attain the food from the free food dish in front of them, the rats
actually pushed the food dish out of the way (not eating from it), and then pressed the lever to
make a food pellet fall into the food tray, where they ate it (see Carder & Berkowitz, 1970). Such
behavior is about control effectiveness and not just about value effectiveness. If it was just about
value effectiveness, the rats would eat from the free food dish, thereby maximizing the benefits/
costs ratio given that it would be the same beneficial food for less cost in effort. (For a human
analog, see Eitam, Kennedy, & Higgins, 2013; Karsh & Eitam, Chapter 12 of this volume).
(p.332) As I mentioned earlier, control effectiveness is also separate from truth effectiveness.
High predictability provides truth effectiveness, but it need not provide control effectiveness.
And for most people high predictability, even when it is combined with high desired outcomes, is
not enough if personal control is lacking. Consider again Neo in The Matrix. Morpheus asks Neo,
Do you believe in fate, Neo? Neo answers, No! Why not? asks Morpheus. Neo replies,
Because I dont like the idea that Im not in control of my life. For Neo, fate as high
predictability or truth is not enough. He also wants personal control.
As another illustration of the difference between wanting control effectiveness even when you
have both value and truth effectiveness, consider the following thought experiment offered by
Robert Nozick in his book Anarchy, State and Utopia (Nozick, 1974). He asks us to imagine an
experience machine that could give us whatever desirable or pleasurable experiences we could
possibly want. If you get into this machine, your pleasant experiences will be fully convincing;
you would not be able to tell that they were not veridicalthat is, you will experience both value
effectiveness and truth effectiveness. The inventor promises that for the rest of your life in the
machine you will have a fully convincing experience of a life that is better than whatever would
have happened to you outside the machine. For example, whatever your salary would be outside
the machine, it will be experienced as higher within the machine; you will marry a prettier wife;
your children will be better behaved; your career promotions will occur sooner; and so on,
regarding all your life pursuit outcomes. And it will all be experienced as real. Nonetheless, most
people choose not to get into the machine because the inventor rather than they will be in
control of what happens.
Truth in the Service of Control Effectiveness

Now that I have distinguished control effectiveness from value and truth effectiveness, I will
review some mechanisms of effective control that have been emphasized in the motivational
literature. As will be seen, these mechanisms often involve control and truth working together.
Being effective in control often involves making use of truth functions. This fact is
underappreciated in the literature.
Mechanisms of Self-Control
Sigmund Freud provides a salient starting point for a discussion of control motivation. It was
Freud who first suggested that controlling conflicts between (p.333) inner motivational forces
was the major psychological problem that people faced (see Freud, 1961b). The most
Page 5 of 15
fundamental conflict was between the motivational forces of the id and the motivational forces
of the superego. The motivational forces of the id are primitive, impulsive forces that follow the
program of the pleasure principle. They relate to instinctual wishes and desires, such as those
for sex and aggression and egoistic self-satisfaction. In contrast, the motivational forces of the
superego are learned demands and prohibitions, such as societal norms, prescriptions, and rules
(see Freud 1961a, 1961b). In one of his last books, Civilization and Its Discontents, Freud
describes how civilization demands that individuals curb their personal pleasures. Freud
suggests that this is best accomplished through the superego controlling the id.
How is the conflict between the id and the superego managed? It is not enough for the superego
to control the id because the superego can be overly harsh in its demands. There is a need to
manage the conflict between the id and the superego because neither motivational force is
realistic about the balance between the legitimate wants of both the individual and society.
According to Freud, it is the role of the ego, with its reality principle, to find the balance, to
resolve the conflict between the id and the superego according to what is real (see Freud, 1961a;
see also Block, 2002; Block & Block, 1980). What has been underappreciated is the fact that
when Freud proposes that the id-superego conflict should be controlled by the ego, he is
proposing the use of a truth functionthe reality principlein the service of control
effectiveness.
Since Freud, the theory of control motivation that perhaps has received the most attention in
motivation science is Walter Mischels theory of delayed gratification. While Freuds problem
concerned individuals suppressing wished-for acts that were forbidden by others, Mischels
problem concerned individuals giving up something they want now in order to attain something
better in the future. Mischel considers such delay of gratification to be a basic self-control task
that is at the core of willpower. In his so-called marshmallow test to measure self-control (see
Mischel, 1974; Mischel & Ebbesen, 1970), preschoolers were brought into a room one at a time
and were seated at a table and shown two objects, such as a marshmallow or a pretzel. It is
known from pretesting that, although the test child likes both of the objects, one of them is
clearly preferred, such as the test child preferring the marshmallow. In order to attain the
preferred object, the child must wait alone with the two objects on the table until the
experimenter returns to the room. At any time while the experimenter is away, the child can
ring a bell that is on the table to signal the experimenter to return. But the children know that if
they hit the bell, then it is the less preferred object, the pretzel, that they will eat, rather than
the more preferred object, the marshmallow. Preschoolers performance on Mischels
marshmallow test has been found to predict school-related competencies (p.334) many years
later (see Mischel, Shoda, & Rodriguez, 1989); Shoda, Mischel, & Peake, 1990).
What strategies do children use for control effectiveness in this task? What did not work was
having the children think about the marshmallows yummy properties and how delicious it would
be to eat it if you were able to wait. What did work for children was to mentally transform the
marshmallow into a non-consummatory object, such as a white fluffy cloud. The explanation
given for this was that the concrete hot marshmallow object had been mentally transformed
into an abstract cool object (see Metcalfe & Mischel, 1999). However, it is also possible that,
once again, effective control was assisted by a truth function. Mental transformation may have
succeeded by reducing the objects truth. The marshmallow, after all, was not really a white
Page 6 of 15
fluffy cloud. Thus, this mental transformation made the tempting object less real. By being less
real, its motivational force, its pull, was reduced.
From this perspective, we can reconsider other studies by Mischel and his colleagues that also
demonstrated the power of the mental transformation strategy (see Mischel & Moore, 1980;
Mischel, 1974). While the children waited, they were shown color pictures of the objects rather
than the actual objects, or they faced the actual objects but were asked to think of them as color
pictures. How long the children were willing to wait also increased in these treat as a picture
conditions. Note that when the children in these studies were shown color pictures of the
objects or were told to think of the actual objects as color pictures, they were explicitly
instructed to think of them as not real (The ____ arent real; theyre just a picture.). The
properties of the actual objects were depicted, but the objects were being experienced as not
real; they were either just pictures, or were thought of as being just pictures. Being experienced
as not real made the tempting objects easier to resist.
The difference between the Freud case and the Mischel case in how truth serves control needs
to be highlighted. In Freuds interpretation of controlling the conflict between id temptations
and superego demands, it is asserting the egos reality principle that supports success. In
contrast, in Mischels studies of controlling immediate temptations for future goals, it is
reducing the reality of the immediate temptations that supports effective control in the situation.
Thus, the use of truth functions for effective control does not always means establishing more
reality, or more truth. Rather, truth or reality can be manipulated in whatever way best serves
successful control. This aspect of controls use of truth has received insufficient attention in the
literature.
Another example of successful control that involves reducing the current reality or truth of
something is the self-control mechanism of repressing unwanted inner states. Let me begin
again with Freud. The prototypic case of unwanted inner states was people feeling anxious or
guilty about their (p.335) thoughts, feelings, or desires about another person, and the classic
example of this was very young boys (between 35 years of age) feeling anxious and guilty about
wanting to kill their father in order to possess their motherthe Oedipus complex. By repressing
the unwanted inner states, the superego is formed (Freud, 1961a, 1965). From a psychoanalytic
perspective, controlling the unwanted Oedipal thoughts, feelings, and desires has special
significance to human development. Indeed, controlling Oedipal thoughts and feelings through
repressionkeeping them unconsciousis considered to be a continuing self-control task for all
well-functioning adults. It should be emphasized that the purpose of this repression is to keep
the thoughts and feelings from consciousness because, if conscious, they would make people
feel extremely anxious and guilty. In this case, the self-control problem is resolved by reducing
the experienced reality of the unwanted thoughts and feelings through repression.
The role of truth functions in controlling unwanted inner states is not restricted to repression.
And it need not involve only reducing a current reality or truth. It can also involve changing it.
For example, rather than repressing emotionally stressful events, people can talk or write about
them, and such verbalization has been found to be effective in reducing emotional stress. The
benefits to adjustment from verbalizing about (rather than repressing) them have been
explained in terms of what happens when people write or talk about traumatic experiences (see
Page 7 of 15
Pennebaker, Mayne, & Francis, 1997). First, an organized and coherent explanation or story is
constructed about the trauma. Second, labeling the emotions associated with the trauma helps
individuals to integrate them into their general understanding of the traumatic event. Note that
both of these are truth functions.
It is important, however, to recognize that trying to get to the bottom of things, asking why?
questions, is not always beneficial. Indeed, when individuals go over and over in their mind
about why something happened, trying incessantly to make sense of what happened, that is,
rumination, it can actually intensify the negative affect because asking questions and not getting
answers is a problem (Nolen-Hoeksema, 2000). It is precisely those individuals who do not
arrive at an explanation or find insight who continue to ruminate (Pennebaker, Mayne, &
Francis, 1997). Thus, trying to find the truth is not in itself beneficial. What makes verbalization
or looking for an explanation effective as a control mechanism is feeling that you have been
successful at finding the truth.
Another alternative to repression is distancing as a self-control strategy. A tactic for distancing

ourselves when thinking about a past traumatic event is to take a third-person perspective, to
step back and imagine watching the event as if it were happening again and you were watching
it from a distance, (p.336) as if you were watching a video of what happened. Studies have
shown that self-distancing, along with why? questioning, is effective (Kross, Ayduk, & Mischel,
2005). The distancing reduces the intensity of the experienced reality of the traumatic event,
while the questioning makes sense of what happenedtwo truth functions in the service of
successful control.
As a final example of using truth functions for effective control of unwanted thoughts or feelings,
consider constructive alternativism as a self-control strategy. George Kelly (1955, 1969)
described how constructive alternativism could be a strategy for improving control
effectiveness. His core idea was that events could always be construed differently, and thus
people could change their views about past events and themselves in ways that worked better
for them (cf. Ochsner & Gross, 2004). Note that it is not about accuracy per se but constructing
a reality that worksself-control as the ability to change self-views to produce a better fit
between the self and the world (see Rothbaum, Weisz, & Snyder, 1982). By supporting
alternative construals of past events, this truth function can also contribute to the control
effectiveness of verbalization and of self-distancing plus asking why? questions.
Other Control Mechanisms Using Truth Functions: Selection, Commitment, and Feedback
Effective control requires not only self-control mechanisms but other mechanisms as well. Let us
briefly consider how truth functions support some control effectiveness mechanisms involved in
the self-regulatory processes involved in selection, commitment, and feedback.
Selection
Effective management of goal pursuit requires making different selections during different
phases of the goal-pursuit process. In their Rubicon model of self-regulation (see Gollwitzer,
1990; Heckhausen & Gollwitzer, 1987), Heinz Heckhausen and Peter Gollwitzer describe
different selections at different goal-pursuit phases. Within the pre-actional or goal-setting
phase (see Lewin, Dembo, Festinger, & Sears, 1944), they distinguish between a pre-actional
Page 8 of 15
deliberative phase and a pre-actional implementation phase. I describe these two pre-actional
phases next.
The deliberation phase is when goals are selected. People must deliberate over which of their
many different wishes and needs they prefer to pursue because not all of them can be pursued
at the same time, even if a few can be pursued simultaneously. This deliberation phase involves
assessing and comparing both the desirability and the feasibility of different wishes or needs in
order to construct goal preferences. Notably, assessing and comparing are (p.337) truth
functions. Moreover, feasibility concerns a determination of the likelihood that certain outcomes
will occur and thus also entails a truth function. During the implementation phase, strategies,
tactics, and specific behavioral intentions need to be planned. Implementation planning involves
selecting those strategies, tactics, and behaviors to be initiated, executed, and terminated
during the goal-pursuit process. Such planning requires addressing questions of when to start
acting, where to act, how to act, and how long to act. Addressing such questions involves yet
another kind of truth function.
Commitment
Managing commitment is necessary for effective goal pursuit. Where does the feeling of
determination to pursue a goal come from? The standard answer, including in the Rubicon
model, is that it comes from the utility of a goal pursuit. There are two factors that contribute to
the utility of a goal pursuitthe subjective value of successful goal pursuit and the subjective or
perceived likelihood of successful goal pursuit. Commitment to pursuing a goal will be stronger
when the subjective value of success is high (vs. low). Commitment to pursuing a goal will also
be stronger when people perceive a high (vs. low) likelihood of success. But, again, perceived
likelihood concerns a belief about whether something is true or will be true, and thus the
contribution of perceived likelihood to the commitment needed for effective control constitutes a
contribution to control from a truth function.
Feedback
The feedback function of goal pursuit involves both evaluating the goal-pursuit activity while it is
still ongoing (how am I doing?) and evaluating the success or failure of the goal pursuit after it
is completed (how did I do?). Regarding how am I doing? one of the best-known feedback
models is Carver and Scheiers control model of self-regulation. This model proposes that there
are two layers of managing what happens through feedback that keep a person on track during
goal pursuit (see Carver, 2004; Carver & Scheier, 1998, 2008). The first layer of feedback
concerns goal attainment or maintenance and consists of an input, a reference value, a
comparison, and an output (for other control models, see also Miller, Galanter, & Pribram, 1960;
Powers, 1973; Wiener, 1948). The input is information about the present condition, or the
current state. The goal, or desired end-state, provides the reference value. The input (current
state) is compared to the reference value (the desired end-state). If a discrepancy between the
input (current state) and the reference value (desired end-state) is detected in this comparison,
then there is an error signal and an output of taking action to reduce (or eliminate) the
discrepancy.
Page 9 of 15
(p.338) The second layer of feedback involves affect and provides the degree of urgency behind
the action to reduce a detected discrepancy. This second layer functions simultaneously with the
first layer and is monitoring or checking on how well the first process is doing in its goal
attainment or maintenance. Specifically, the input for the second layer of feedback is the rate of
discrepancy reduction over time. There is a rate criterion that determines what affect occurs.
When the rate of progress in reducing a discrepancy is below this criterion, negative affect
occurs; when the rate of progress is above this criterion, positive affect occurs.
These feedback processes are a critical part of effective control regulation. Once again, note
how they involve different kinds of truth functions, including a comparison process, an error
signal when a discrepancy is detected, and an affective signal concerning whether the rate of
discrepancy reduction over time is above or below a rate criterion. According to control models,
effective control requires all of these feedback truth functions. It should be noted that the
affective signal concerning discrepancy reduction could be related to an actors sense of agency.
When the affective signal is positive, people would experience control effectiveness and a sense
of agency, whereas they would not when the affective signal is negative.
Regarding how did I do? the post-actional assessment phase of the Rubicon model includes
feedback about whether the actual value of goal success (or failure) matches the expected value.
This question is trying to establish what was real in the past. By knowing the truth about how we
did in our completed goal pursuit, we can better manage future goal pursuits by retaining
control strategies or tactics that were effective and efficient, or searching for new strategies or
tactics that could be more effective or efficient. This feedback, then, is another case in which
truth is in the service of control.
Going in the Right Direction from Control and Truth Working Together
The great thing in the world is not so much where we stand, as in what direction we are
moving.
OLIVER WENDELL HOLMES
The good life is a process, not a state of being. It is a direction, not a destination.
CARL ROGERS
As the renowned Supreme Court Justice Oliver Wendell Holmes pointed out, life is not so much
about the state of having desired results, or where we stand, as it is about the direction in which
we are moving. Carl Rogers, a founder of the (p.339) humanistic approach in psychology and
psychotherapy, agreed; having a good life is not about some state of being, or some destination,
as a desired result, but, rather, it is the process of moving in some direction. The message is that
we must manage to move in some direction, which is control. But to have a good life we cannot
move just anywhere. We need to move in the right direction. And to know which direction is
right, we need to have truth as well. This is why it is essential that control and truth work
together as partners. In this last section I will illustrate this by briefly describing how the
regulatory mode functions of locomotion-control and assessment-truth work together effectively.
Page 10 of 15
Locomotion-Control and Assesment-Truth

When people self-regulate, they decide what they want that they do not currently have. They
then figure out what they need to do to get what they want, and then they do it. Two key
functions of self-regulation are captured in this conception. First, people assess both the
different goals to pursue and the different means to pursue them. Second, people locomote or
move from their current state in pursuit of some alternative goal pursuit state (see Higgins,
Kruglanski, & Pierro, 2003; Kruglanski, Thompson, Higgins, Atash, Pierro, Shah, & Spiegel,
2000).
Assessment is the aspect of self-regulation that is concerned with the truth, with determining
what is best and what is right. It establishes this truth by critically evaluating entities or states
in relation to alternatives, such as critically evaluating alternative goals or alternative means, in
order to judge relative quality. Individuals with strong assessment concerns want to compare all
options and search for new possibilities before making a decision, even if that process takes
time and delays the decision. They relate past and future actions to critical standards. They want
to choose the option that has the best attributes overall compared to the alternative options.
They want to make the correct choice. They want to get it right.
By contrast, the locomotion mode is the aspect of self-regulation that is concerned with control,
with making things change and happen. It exerts control by moving or changing from one state
to another state. Individuals with strong locomotion concerns want to take action, to get started,
even if that means not considering all the options fully. Once the task is initiated, they want to
maintain it and complete it without undue disruptions or delays. They want to just do it.
From the perspective of traditional control-system models, it is not reasonable to ask whether it
is the monitoring function or it is the operating function that is more important for successful
achievement because both functions are equally critical in these models. In contrast, by
separating the independent (p.340) functions of locomotion and assessment, it is possible for
regulatory mode theory to address the question of whether strong locomotion concerns or
strong assessment concerns are more important for successful achievement. In research on
students cumulative GPA achievement and on soldiers completion of an extremely demanding
military training program, what was found was that both locomotion and assessment mattered,
but higher locomotion was critical (Kruglanski et al., 2000). Specifically, higher achievement
was found for students and for soldiers who had stronger locomotion concerns, but this higher
achievement was more pronounced for those who also had stronger assessment concerns.
Stronger assessment, by itself, did not predict higher achievement. However, a certain level of
assessment concerns had to be reached before stronger locomotion concerns translated into
higher achievement. As we have seen consistently throughout this chapter, a truth function,
assessment, contributed to effective control from strong locomotion.
Strong assessment alone has the potential downside of leaving people lost in thought and
taking no action. Strong locomotion, on the other hand, wants to take action, wants to get going,
wants to effect change. Taking action is necessary for achievement, but there is a downside of
strong locomotion as wellit could motivate going in any direction, including a bad direction.
This is where strong assessment helps out. It works together with locomotion to ensure that the
Page 11 of 15
direction taken is the right direction, while locomotion ensures that some going takes place
going in the right direction.
Conclusion
The fact that stronger locomotion-control concerns do not translate into more effective
achievement unless a certain level of assessment-truth concerns is reached makes a lot of sense.
After all, effective achievement requires having some idea of which goals are good and which
means are good. Without this truth, individuals with strong locomotion-control concerns could
be effective in moving from state to state without the movement being directed toward a better
state than the current state. Something has to establish a basic reality about where the
movement should be headed. That is, it is important to go in the right direction. Once
established, however, then the strength of motivation to initiate and maintain smooth and steady
movement or change becomes the essential factor for achievement. This provides a new insight
into how the partnership of control and truth provides the motivational underpinnings of
effective achievement. And when control and truth do work together effectively, people
experience a special kind of sense of agencythe experience of going in the right direction.
And, importantly, when people go in the right direction, it does not necessarily mean that they
will end with experiencing desired results (i.e., value (p.341) effectiveness). Neo and Truman
also had direction for the first time when they chose truth and control over hedonic pleasure,
but again this led initially to a dangerous situation for them. And all of us are engaged daily in
managing our actions to move in the direction of a chosen goal, even though we recognize that
we might not attain that goal and we cannot be sure what will happen if we do attain it.
Nonetheless, we feel good about going in the right direction. As Carl Rogers said, the good life is
a direction, not a destination.
References
Block, J. H. (2002). Personality as an affect-processing system: toward an integrative theory.

Mahwah, NJ: Erlbaum Associates.
Block, J. H., & Block, J. (1980). The role of ego-control and ego-resiliency in the organization of
behavior. In W. A. Collins (Ed.), Minnesota symposium on child psychology (Vol. 13, pp. 39101).
Hillsdale, NJ: Erlbaum Associates.
Carder, B., & Berkowitz, K. (1970). Rats preference for earned in comparison with free food.
Science, 167, 12731274.
Carver, C. S. (2004). Self-regulation of action and affect. In R. F. Baumeister & K. D. Vohs

(Eds.), Handbook of self-regulation: research, theory, and applications (pp. 1339). New York:
Guilford Press.
Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. New York: Cambridge
University Press.
Page 12 of 15
Carver, C. S., & Scheier, M. F. (2008). Feedback processes in the simultaneous regulation of
action and affect. In J. Y. Shah and W. L. Gardner (Eds.), Handbook of motivation science (pp.
Research, 229, 475484.
Elliot, A. J., & Fryer, J. W. (2008). The goal construct in psychology. In J. Y. Shah and W. L.
Gardner (Eds.), Handbook of motivation science (pp. 235250). New York: Guilford Press.
Freud, S. (1961a). The ego and the id. In J. Strachey (Ed. and Trans.), Standard edition of the
complete psychological works of Sigmund Freud (Vol. 19, pp. 366). London: Hogarth Press.
(Original work published 1923.)
Freud, S. ([1930] 1961b). Civilization and its discontents. Edited and translated by J. Stachey.
New York: W. W. Norton. (Original work published 1930).
Freud, S. ([1933] 1965). New introductory lectures on psychoanalysis. Edited and translated by
J. Stachey. New York: W. W. Norton. (Original work published 1933).
Elliott, E. S., & Dweck, C. S. (1988). Goals: an approach to motivation and achievement. Journal
Gollwitzer, P. M. (1990). Action phases and mind-sets. In E. T. Higgins and R. M. Sorrentino

(Eds.), Handbook of motivation and cognition: foundations of social behavior (Vol. 2, pp. 5392).
New York: Guilford.
Heckhausen, H., & Gollwitzer, P. M. (1987). Thought contents and cognitive functioning in
motivational versus volitional states of mind. Motivation and Emotion, 11, 101120.
Higgins, E. T. (2012). Beyond pleasure and pain: how motivation works. New York: Oxford
University Press.
Higgins, E. T., Kruglanski, A. W., & Pierro, A. (2003). Regulatory mode: Locomotion and
assessment as distinct orientations. In M. P. Zanna (Ed.), Advances in experimental social
Hull, C. L. (1943). Principles of behavior. New York: Appleton-Century-Crofts.
Hull, C. L. (1952). A behavior system: an introduction to behavior theory concerning the

individual organism. New Haven, CT: Yale University Press.
James, W. ([1890] 1948). Psychology. New York: The World Publishing Company. (Original
publication, 1890).
Kelly, G. A. (1955). The psychology of personal constructs. New York: W. W. Norton.
Page 13 of 15
Kelly, G. A. (1969). Clinical psychology and personality: the selected papers of George Kelly. Ed.
B. Maher. New York: Wiley.
Kross, E., Ayduk, O., & Mischel, W. (2005). When asking why does not hurt: Distinguishing
rumination from reflective processing of negative emotions. Psychological Science, 16, 709715.
Kruglanski, A. W., Shah, J. Y., Fishbach, A., Friedman, R., Chun, W. Y., & Sleeth-Keppler, D.
(2002). A theory of goal systems. In M. P. Zanna (Ed.), Advances in experimental social
psychology (Vol. 34, pp. 331378). San Diego, CA: Academic Press.
Kruglanski, A. W., Thompson, E. P., Higgins, E. T., Atash, M. N., Pierro, A., Shah, J. Y., &
Spiegel, S. (2000). To do the right thing or to just do it: locomotion and assessment as
distinct self-regulatory imperatives. Journal of Personality & Social Psychology, 79, 793815.
Lewin, K. (1935). A dynamic theory of personality. New York: McGraw-Hill.
Lewin, K. (1951). Field theory in social science. New York: Harper.
Lewin, K., Dembo, T., Festinger, L., & Sears, P. S. (1944). Level of aspiration. In J. McHunt (Ed.),
Personality and the behavior disorders (Vol. 1, pp. 333378). New York: Ronald Press.
McDougall, W. (1914). An introduction to social psychology (8th ed.). Boston: Luce.
Metcalfe, J., & Mischel, W. (1999). A hot/cool-system analysis of delay of gratification: dynamics
of willpower. Psychological Review, 106, 319.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the structure of behavior. New
York: Holt, Rinehart, & Winston.
Mischel, W. (1974). Processes in delay of gratification. In L. Berkowitz (Ed.), Advances in

experimental social psychology (Vol. 7, pp. 249292). New York: Academic Press.
Mischel, W., & Ebbesen, E. B. (1970). Attention in delay of gratification. Journal of Personality
and Social Psychology, 16, 329337.
Mischel, W., & Moore, B. (1980). The role of ideation in voluntary delay for symbolically
presented rewards. Cognitive Therapy and Research, 4, 211221.
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science,
244, 933938.
Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed anxiety/
depressive symptoms. Journal of Abnormal Psychology, 109, 504511.
Nozick, R. (1974). Anarchy, state, and utopia. New York: Basic Books.
Page 14 of 15
Ochsner, K. N., & Gross, J. J. (2004). Thinking makes it so: a social cognitive neuroscience
approach to emotion regulation. In R. F. Baumeister & K. D. Vohs (Eds.), Handbook of self-
regulation: research, theory, and applications (pp. 229255). New York: Guilford Press.
Osborne, S. R. (1977). The free food (contrafreeloading) phenomenon: a review and analysis.
Animal Learning & Behavior, 5, 221235.
Oxford English Dictionary (1971). Compact Edition, Vol. I, II. Oxford: Oxford University Press.
Pennebaker, J. W., Mayne, T. J., & Francis, M. E. (1997). Linguistic predictors of adaptive
bereavement. Journal of Personality and Social Psychology, 72, 863871.
Pervin, L. A. (Ed.) (1989). Goal concepts in personality and social psychology. Hillsdale, NJ:
Erlbaum Associates.
Powers, W. T. (1973). Behavior: the control of perception. Chicago: Aldine.
Rothbaum, F., Weisz, J. R., & Snyder, S. S. (1982). Changing the world and changing the self: a
two-process model of perceived control. Journal of Personality and Social Psychology, 42, 537.
Shoda, Y., Mischel, W., & Peake (1990). Predicting adolescent cognitive and self-regulatory
competencies from preschool delay of gratification: identifying diagnostic conditions.
Developmental Psychology, 26, 978986.
Wiener, N. (1948). Cybernetics: Control and communication in the animal and the machine.
Woodworth, R. S. (1921). Psychology: a study of mental life. New York: Holt.
Page 15 of 15
Sense of Agency and Its Disruption

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Clinical and Computational Perspectives
Paul Fletcher
Aikaterini Fotopoulou
DOI:10.1093/acprof:oso/9780190267278.003.0016

Sense of agencythe feeling of being the author of ones actionsmay be a critical component
of ones sense of self and of ones interaction with the world. Insights from clinical and
experimental neuropsychology, as well as cognitive and computational neuroscience, have
provided complementary evidence that the sense of agency arises from the integration of an
array of internal and external cues. These frameworks can help to explain how disruptions in
one or more of these cues may result in altered experiences of agency. This chapter reviews
these explanatory frameworks and shows how important and useful they have become in making
sense of an array of clinical observations, from the disorders of control and agency that result
from circumscribed brain damage to the widespread attenuation of agency that may
characterize psychosis in which no clear brain lesion has been identified.
Keywords: experimental neuropsychology, cognitive, computational neuroscience, agency, brain damage,

psychosis
Introduction
In this chapter, we examine the nature of sense of agency from a clinical and a computational
perspective. Beginning with a consideration of the nature of the subjective experience of agency
and the complexity of factors that may underlie, we discuss briefly the emergent models seeking
to understand the processes that govern and shape it. In doing so, we point to certain areas of
contention before taking these ideas forward into a discussion of agency in neuropsychiatry,
Page 1 of 23
focusing on agency in psychosis (notably delusions of control), Anosognosia for Hemiplegia and
Anarchic Hand syndrome. In doing so, we will attempt to apply the insights afforded by the
theoretical models of SoA and motor awareness that we consider. We believe such consideration
may enhance the understanding of the pathogenesis of these disorders as well as the
mechanisms underlying the SoA and motor awareness more generally.
We begin by considering the sense of agency and how it might be quantified.
What Is Sense of Agency and How Do We Measure It?

A fundamental aspect of our self-consciousness is the compelling feeling of authorship and
control that accompanies and characterizes many of our (p.348) actions (Gallagher, 2000). We
see ourselves as the wilful originators of our acts and, moreover, when an action is performed,
we recognize ourselves as its prime controller. In the last two decades there has been a
burgeoning body of neuroscientific and neurophilosophical literature on such notions of
agentive self-awareness (Pacherie, 2007). Despite such progress, however, we are faced with
inconsistency, complexity, and contention. In fact, most of the processes underlying our
movements are unconscious; we seem to become aware only of a small proportion of them. Key
facets of this limited and elusive agentive self-awareness include (1) the sense of agency (SoA),
or the subjective feeling that we, and not somebody else, have caused and controlled an action
and its effects (a narrow, heuristic definition of the SoA); (2) the subjective feeling that we are
moving, or have just executed a movement (motor awareness); and (3) the awareness of our own
intention or urge to act (ranging from prior intentions, such as intending to go for a swim, to
intentions-in-action, such as intending to take a dive; Searle, 1983).
It should be noted that, despite such seemingly clear definitions and demarcations, these
awareness concepts are themselves multifaceted. The relation between them is, at times,
unclear, and their phenomenal contents are rather slippery. For example, in many, if not most,
cases, we would not necessarily experience ongoing consciousness of initiating, executing, and
controlling actions, yet we would argue with someone who denied that it was we who caused,
executed, and controlled the actions in question. Under certain circumstances, healthy people
may incorrectly attribute agency to actions that they do not really govern, or may fail to
attribute agency to those that they do. One example of the former is water-divining, in which
movements of a rod, caused by the carrier, are actually attributed to the presence of some force
exerted by a nearby body of water. Conversely, many people are unaware of the degree to which
the movements of a cursor on a screen are actually assisted by the computer software,
attributing it entirely to themselves (Fourneret & Jeannerod, 1998; Sarrazin et al., 2008). Added
to this is a degree of contention over what sorts of cues (internal or external) primarily engender
and shape our SoA and how these may be integrated. The precise mechanisms for such
integration and the nature of signalsboth those originating internally and externallythat are
critical to SoA are also unclear. Similarly, in the domain of motor awareness, we have the
everyday belief that we are aware of the kinds of actions we execute as we perform them. In
reality, so long as an expected action goal is achieved, we are largely unaware of precisely how
we executed the movements involved, including the errors and adjustments we had to make on
the way (Fourneret & Jeannerod, 1998; Sarrazin et al., 2008). This implies that our normal
subjective feeling of action execution is, to a degree at least, non-veridical and distinct from
motor control. Thus, the precise mechanisms (p.349) by which mostly efficient and non-
Page 2 of 23
conscious processes of motor control give rise to our non-veridical and conscious feelings of
movement execution remain unknown.
It is worth noting at the outset that successful scientific study of SoA and related awareness
notions is obviously reliant on the degree to which such senses may be identified reliably and
quantified accurately. Generally, subjective reports have been usedexperimental participants
are required to indicate their feeling that an action has been carried out or not and, if yes,
whether it was they who caused it or some external agent. Such judgments may be binary (did I
or didnt I initiate this movement?) or they may be rated in terms of the strength of the sense
under consideration. There are other, more indirect, measures of both agency and motor
awareness. While their indirectness may be disadvantageous, they do offer complementary
quantitative measures of SoA and motor awareness, allowing subtle experimental assessments
that might not otherwise be possible. One phenomenon that has been suggested to provide an
indirect measure of agency is intentional binding, which is a temporal measure of the degree to
which an action and an ensuing outcome are bound in time. The key finding with such
measures is that, when an action is intentional, the actor perceives it to be closer in time to the
outcome than it actually is. If the action is not internally generated but is produced by, for
example, transcranial magnetic stimulation, the binding effect is lost, suggesting that it is
peculiar to instances in which there is agency. This therefore offers the possibility that binding
may be a reliable surrogate measure for agency. Interestingly, this binding effect appears to be
produced not just by the presence of the actions outcome but also by the agents prediction of
the outcome (the perception of the time of action shifts, when there is a strong prediction of an
outcome, even when that outcome does not actually ensue). Using combinations of subjective
judgment of agency and intentional binding, it has been possible to explore the impact of
external factors and expectations (Moore et al., 2013) on both explicit and implicit measures of
SoA. The intentional binding approach has, moreover, proven useful and sensitive in studies of
conditions such as schizophrenia in which agency is thought to be altered.
Similarly, within the domain of motor awareness, one could trace at least two major distinctions,
each with related empirical implications. First, as with other domains of awareness, there is the
distinction between implicit and explicit indices. Explicit measures of awareness relate to
conscious, subjective feelings and thoughts that are available for verbal report, while implicit
ones relate to knowledge that is expressed in task performance unintentionally and with little
or no phenomenal awareness (Schacter, 1990, p. 157). Thus, while motor awareness can be
measured by verbal report, implicit measures include reaction times (Nardrone et al., 2007;
Fotopoulou et al., 2010) or (p.350) choice of behavioral strategy (e.g., Cocchini et al., 2010;
Moro et al., 2011). Second, there is a distinction between on-line (or emergent) and off-
line (or anticipatory) motor awareness (Carruthers, 2008; Crosson et al., 1989; see also
Tsakiris & Fotopoulou, 2008). The former terms refer to the subjective feeling of moving in the
moment or having just moved, while the latter refer to a more general expectancy and inference
about ones ability to move and execute future actions, as for example when one is feeling able
to reach a target. This distinction also necessitates different types of measurements, including,
for example, confrontation tasks where participants are asked to execute movements and report
on their experience versus estimation tasks where subjects are asked to estimate their future
performance in given motor tasks. The exact relation between these facts of motor awareness,
as well as between their respective measurements, remains currently unclear. However, studies
Page 3 of 23
in neuropsychiatric populations have demonstrated that such aspects of awareness may

dissociate between and even within patients (Cocchini et al., 2010; Fotopoulou et al., 2010; Moro
et al., 2011), and hence they warrant separate and careful consideration.
We will return to examples of studies using both implicit and explicit measures of agency and
action awareness in the context of neuropsychiatric disease. Before this, given some of the
above contentions over the nature of motor agency and awareness, we will consider existing
models of how SoA arises and what factors may shape it, including internal and external cues, as
well as the levels of uncertainty and noise that modulate the impact of these cues. In considering
these factors we will introduce computational models of agency appealing to optimal motor
control, predictive coding, and active inference and discuss their respective contributions to our
understanding of these fundamental aspects of the bodily self.
Internal and External Cues to Agency

One simple distinction in models of SoA concerns the extent to which this sense emerges as a
consequence of internal cues (for example, those concerned with sensorimotor experiences) or
of external cues (for example, priming with cues suggestive that one is indeed the agent of
action). We have reviewed this elsewhere (see Moore & Fletcher, 2012) and it has been more
extensively treated by Synofzik, Vosgerau, and Newen (2008). Indubitably, both forms of cue can
engender SoA, even to the extent of producing a false SoA for an action for which one had not
been responsible (Wegner & Wheatley, 1999; Moore, Wegner & Haggard, 2009). To be a little
more specific, the internal cues providing a signal that a movement was agentic involve the
conscious experience of the intention or the desire to move, as well as the proprioceptive and
other sensory changes (p.351) signaling that a movement has indeed occurred. One
suggestion, for example, is that a true sense of agency emerges when the predicted sensory/
proprioceptive consequences (based on a forward model shaped by a computation of the
actions required to achieve the goal or intention) closely match those consequences that actually
ensue. A mismatch, it is suggested, may well designate a movement that was not intentional or
under ones own control. This comparator view of the emergence of SoA will be considered
briefly below, as well as models that reject this view of SoA arising from an absence of
mismatch.
Perhaps less obviously, external cues may influence the experience of agency. Wegner and
Wheatley (1999) pointed out that the experience of willed action relies on a tripartite experience
connected with an internal thought about the act in question: specifically, that this thought is
prior to the act, that it is consistent with this act, and that it alone can account for that act. In
this respect, the SoA is a causal attribution, entailing the same criteria as if one were making a
causal judgement about any two events. The experience of willing the action is a candidate
cause, but not the only cause, of bodily movement. Such a perspective offers insights to
erroneous SoA and also allows for the possibility that external factors may modulate the
experience of action in a number of ways. To highlight this, Wegner and Wheatley elegantly
demonstrated the impact of external cues by showing that, in a situation in which agency for an
action (cessation of a continuous movement controlling a cursor on a screen) was ambiguous,
the presence of a relevant, external, auditorily presented, prime word enhanced a participants
sense of agency for this action.
Page 4 of 23
In short, therefore, SoA is shaped by both internal and external cues. This raises a question: If
the agent receives an array of cuesboth internal and externalpotentially relevant to agency,
how might these be combined optimally? Moreover, what might be the outcome when cues are
contradictory? We have previously speculated that, just as has been considered extensively for
sensory cues (e.g., Ernst & Bulthoff, 2004), cues to agency may be combined and integrated
according to their estimated reliability or precision (Moore & Fletcher, 2012). Put simply, if two
cues offer information about agency, then they may be optimally combined by taking into
account the precision (the inverse variance) of each and weighting them accordingly, cues
estimated to be more precise being accorded a greater weighting. This optimal integration can
be represented formally in terms of maximum likelihood estimation. A consequence of such
integration is an overall reduction in variance or an enhancement of precision. As an example, in
sensory processing, integration of visual and haptic information is likely to weight the former
more heavily due to its (generally) superior precision.
Of course, as encapsulated by Bayess Theorem, one critical factor that can be added to the
maximum likelihood estimation to further optimize ones (p.352) estimation or conclusion is
prior belief. Priors, according to Bayesian models, form a critical part of the equation and, in the
face of noisy/imprecise internal or external cues, provide a way of more completely combining
cues. This combination or integration is discussed more fully elsewhere (see, for example, Moore
& Fletcher, 2012). For the purposes of the current chapter, we wish to focus on overarching
models of action based upon these principles before we go on to consider neuropsychiatric
perturbations in light of these models.
Frameworks for Agency: The Comparator Model, Predictive Coding, and Active
Inference
Influential current theories of motor control are based on the pioneering work of Todorov and
Jordan, who argued that sensory signals are passed down the motor hierarchy as motor
commands in order to achieve desired action goals (Todorov & Jordan, 2002; Todorov, 2004).
According to computational models adhering to this now well-established optimal control theory,
this transformation and selective use of sensory signals to specify motor commands entails two
types of models (Wolpert & Kawato, 1998). First, an inverse model selects appropriate motor
commands that would achieve a desired goal. As these motor commands are sent to the muscles,
an efference copy of the same commands is sent to an internal predictive, or forward, model.
This models task is to estimate the likely sensory consequences of the motor command and thus
the intended action. The sensory predictions of this forward model are used to optimize the
estimated state of the motor plan required by the inverse model. Sensory feedback provides
additional information about the executed movement, but such sensory transmission is relatively
slow. The advantage of the joined action of the inverse and forward models is that it can bypass
these sensory delays, allowing rapid adjustments and fluent movements toward a desired goal.
More generally, the optimal control of action is thought to depend to a large extent on the
coordination of inverse and forward models through a series of comparators, the results of the
comparisons being used to correct errors, deviations, and other regulatory purposes.
These theories were designed to understand motor control and performance, rather than the
subjective experience of action and its control. In fact, as mentioned, several components of
such motor control schemes are considered to be unconscious. Nevertheless, such models have
Page 5 of 23
proven useful in guiding investigations into which aspects of such motor generation schemes are
linked to the subjective feelings of action awareness and agency (for reviews, see Desmurget &
Sirigu, 2009; Frith et al., 2000). Thus, according to such perspectives, our SoA depends on the
degree of congruency between the predicted and actual consequences of our movements, with
this mismatch being (p.353) continuously monitored as an action unfolds. Moreover, and
crucially, predictions made by forward models are used to filter sensory information and to
attenuate the sensory effects of self- versus other-generated movements, thus generating a
marker of agency (which, as mentioned, is signified by an absent or low mismatch). Similarly,
action awareness is thought to rely mainly on forward signals and related comparisons, while
actual sensory feedback may not be necessary to construct motor awareness, particularly when
the desired goal is achieved and there are no unexpected delays (Fourneret & Jeannerod, 1998;
Sarrazin et al., 2008). If, however, the goal is not achieved, or there are large errors or delays in
the process of execution, then a comparator detects the mismatch between the expected and
actual sensory feedback and awareness is updated.
Predictive coding models form a general and, in neuroscience, increasingly influential class of
models accounting for perception and action (see also Box 16.1). In essence, the predictive
coding perspective represents a means of (p.354) (p.355) estimating the likely cause of a
given set of sense data based upon the data themselves as well as prior information. In the case
of SoA and motor awareness, this amounts to the estimation of the probability that one was the
agent of an action given the sensory and perceptual attributes of that action (for example,
proprioception, sensorimotor feedback, external cues) and the prior probability that the action
was executed and one was indeed the agent. Such a framework may, of course, exist at multiple
levels arranged hierarchically. In the context of a model of brain function that posits the brain as
seeking to minimize prediction error, active inference refers to the attempt to predict future
movements through the representation of intentions (Adams et al., 2013; Friston et al., 2011).
Box 16.1 A Computational Neuroscience Framework for Examining Agency
The starting point of the free energy framework (Friston, 2005) is that the world is an
uncertain place for self-organizing biological agents to survive. The signals that an organism
may receive from the world may be caused by several, unknown causes. This inherent
ambiguity of the world specifically threatens our need to occupy a limited repertoire of
sensory states (e.g., humans need certain ranges in environmental temperature in order to
survive). If, however, we cannot predict the causes of possible changes in the world (e.g.,
the weather) with any certainty, we may find ourselves in surprising states for longer
periods than those we could biologically sustain (e.g., in cold climates). We thus come up
with a cheeky and yet no less ingenious solution. We base our predictions about our sensory
states on unconscious inferences about their causes in the world (von Helmholtz, 1866). On
the basis of limited or noisy information, our brain engages in some form of probabilistic
representation of the causes of our future states in an uncertain world so that it maintains
hypotheses (generative models) of the hidden causes of sensory input. Theoretical
neuroscientists use Bayesian theory to formalize this kind of inference and a number of
Page 6 of 23
other computational terms about probability distributions, such as free energy,

uncertainty, and surprise that have the advantage of being formally (mathematically)
defined. We have not addressed the issues of interest in mathematical ways here. Instead we
have attempted to find faithful psychological translations for the mathematical definition
of these concepts in order to examine clinical and experimental evidence on motor agency
and awareness. Later in the chapter we outline the general ideas behind such a
psychologized version of the free energy framework.
The central tenet of this framework is that the brain attempts to reduce the probability of
being surprised by the world. It does this by derivingfrom genes and experience
inferential, predictive models of possible causes of its sensory input. Errors in the accuracy
of such representations have been conceptualized as free energy, on the basis of the formal
definition of the lattera quantity from informational theory that bounds (is greater than)
the evidence for a model of data (Feynmann 1972; Hinton & van Camp, 1993). The brains
data are sensory, and free energy bounds the negative log-evidence (surprise) inherent in
sensory data, given a model of how the data were caused. Furthermore, in agreement with
the so-called predictive coding scheme (Rao & Ballard, 1999), our brain is assumed to
achieve the minimization of free energy by recurrent message passing among hierarchical
level of cortical systems, so that various neural subsystems at different hierarchical levels
minimize uncertainty about incoming information by structurally or functionally embodying
a prediction (or a prior) and responding to errors (mismatches) in the accuracy of the
prediction: so-called prediction errors. Such prediction errors are passed forward to drive
the units in the level above that encode conditional expectations that optimize top-down
predictions to explain away (reduce, inhibit) prediction error in the level below until
conditional expectations are optimized. Such message passing is considered
neurobiologically plausible on the basis of functional asymmetries in cortical hierarchies
(see Mesulam, 2012), where forward connections (which convey prediction errors) are
driving and backward connections have both driving and modulatory characteristics (thus
modeling the nonlinear generation of sensory input). Given some mathematical assumptions,
free energy can be thought of as the amount of prediction error in any given level of the
system, including both exteroceptive and interoceptive (Seth et al., 2012) prediction errors.
Minimizing free energy then corresponds to explaining away prediction errors following the
principles of Bayes (Friston, 2010).
However, representing the world in constructive ways (perceptual inference) cannot take us
far in terms of our ultimate goal, which is surviving in an uncertain world. Psychologically
speaking, we may become better in predicting (mentalizing) the changes in the
environment that act to produce sensory impressions on us, but we cannot on this basis
change the sensations themselves and hence ultimately their surprise. A highly innovative
conceptual move in the free energy principle framework allows us to understand how we do
just that. By acting upon the world, we can change its states and therefore re-sample the
world to ensure that we satisfy our predictions about the sensory input we expect to receive.
By selectively sampling the sensory inputs that we expect, we add accuracy to our
Page 7 of 23
predictions about sensory states. This view suggests that action is best understood as being
elicited to fulfill prior expectations about proprioceptive sensations, not desired sensory
states (as classic optimal motor control theory would suggest). Thus, action has an intimate
relationship with perception, both being governed by the same master principle, namely
reduction of prediction error. Action therefore becomes a means of adhering to the central
principle of avoiding surprising states or minimizing prediction error. Thus action can
reduce free energy by changing sensory input, while perception reduces free energy by
changing predictions. According to the framework, our predictions (or priors) thus become a
constantly updated, iterative, self-fulfilling prophecy that allows us to evade the inherent
surprise of the world.
Finally, according to the framework, the organism needs to probabilistically infer two
properties of the world: its states (content; mathematically this can be thought of as the
center of a probability distribution), and the uncertainty (context; the dispersion of such
distribution) about such states. Thus, optimal inference in both perception and action
requires optimizing the precision (mathematically inverse amplitude or variance, and hence
the inverse of uncertainty) of sensory signals (Feldman & Friston, 2010; Friston et al.,
2012a). Uncertainty is thought of as encoded mainly by synaptic gain that encodes the
precision of random fluctuations about predicted states. It follows that neuromodulations of
synaptic gain (such as dopamine and acetylcholine) do not signal (reward or pleasure)
prediction errors about sensory data but the context in which such data were encountered.
In other words, such neuromodulators report the salience of sensorimotor representations
encoded by the activity of the synapses they modulate. This is important, especially in
hierarchical schemes, where precision controls the relative influence of bottom-up
prediction errors and top-down predictions. As regards exteroception, this processing of
salience can be seen as attention in perceptual inference (Feldman & Friston, 2010), and as
affordance (latent action possibilities of cues in the environment) in active inference (Friston
et al., 2012a). It has recently been proposed that optimizing the precision of internal body
signals can be seen as increased interoceptive sensitivity and related feelings of arousal in
perceptual inference and as increased seeking behaviours in active inference (Fotopoulou,
2014).
While it is beyond the scope of this chapter to deal with these models in detail, it is worth
mentioning that the comparator and active inference models of SoA are different in certain
fundamentals. While both speculate on the (p.356) importance of minimizing a mismatch or
prediction error signal in controlling movement and generating SoA, the comparator model
relates this to motor commands, while active inference focuses on proprioceptive predictions
wherein a movement is enacted as a means of resolving a proprioceptive prediction error
(emerging from the mismatch between the goal position [i.e., the intention] and the current
position). Given that the movement would violate a prediction that the person is not moving (and
therefore should militate against movement given the framework that actions are taken to
minimize prediction error), it has been speculated (Brown et al., 2013) that the key occurrence
that releases the desired movement is a reduction in the precision of sensory prediction error.
Page 8 of 23
This reduced precision of sensory prediction error is the cause of sensory attenuation (rather
than an absence of mismatch as posited by the comparator model) and thus, like the comparator
model, the active inference model can account for the characteristic sensory illusion seen in a
force-matching task (Brown et al., 2013), though they invoke quite different processes to explain
this effect. Although we cannot cover this evidence in full here, it is worth mentioning that this
model addresses a broader concept of sensory attenuation than the comparator model, including
changes in the sensitivity of responses to external stimuli and not just the criteria of responses
(see Brown et al., 2013, p. 3 for further discussion). The emergence of abnormal SoA in
schizophrenia is likewise accounted for in quite different ways by the two models, but these
precise differences are necessarily beyond the scope of this chapter. We will return to some of
the fundamental difference in the sections on abnormalities of agency.
Aberrant Agency and Motor Awareness

As mentioned, sense of agency and motor awareness are not, even in health, infallibly accurate.
Individuals may falsely claim to be the agents of actions or outcomes that they did not cause.
Conversely, they may fail to attribute agency to actions that they did indeed cause. A celebrated
example of the latter is the table-turning phenomenon of the nineteenth century. Here, the
experience among groups of people was of a table moving under their hands, propelled, they
believed, by members of the spirit world. This strong but erroneous SoA was, presumably, a
consequence of prior beliefs (that spirits could and would move the table) acting on weak
sensorimotor experiences. The latter would be especially vulnerable given that this occurred
within a group such that the force exerted by any individual would be insufficient to move the
table, while the experience of the table turning (as a consequence of the group effort) would be
striking. In addition to a failure to experience agency even when one was (p.357) indeed the
author of an action, the converse is possible. Wegner and Wheatley have, as described above,
shown how appropriately presented cues or primes can produce a feeling that one was the
primary agent of an action that was, in fact, initiated by another (Wegner & Wheatley, 1999).
Given the objective imperfections of SoA as a marker for self- versus externally generated
actions, it is unsurprising that a number of neuropsychiatric illnesses are associated with
profound alterations in the experience of agency. Moreover, while recent studies in cognitive
neuroscience have used several experimental tricks to systematically manipulate sensorimotor
signals, to promote their integration, or to generate conflicts and illusions with a view to
studying their role in SoA and motor awareness, these studies in healthy volunteers are by
necessity constrained by the duration and setup of the experiment. To date, neuropsychiatric
disorders that entail relatively long-lasting and biologically induced abnormalities in the
subjective experience of agency and motor awareness represent an additional, indispensable
window of insight into the neurocognitive mechanisms underlying such subjective feelings. Here
we will attempt to apply the insights afforded by the aforementioned theoretical models of SoA
and motor awareness in the consideration of three prototypical disorders of agency and motor
awareness, namely (1) delusions of control in psychosis; (2) anosognosia for hemipleagia
following stroke; (3) anarchic hand syndrome following brain damage. We believe such
consideration may enhance the understanding of the pathogenesis of these disorders, as well as
the mechanisms underlying the SoA and motor awareness more generally.
Page 9 of 23
Agency and Psychosis: Delusions of Control

Psychosis refers to a loss of contact with reality. It is characterized by abnormal perceptions
(specifically, perceptions in the absence of objectively quantifiable causal stimuli) and abnormal
(bizarre, apparently irrational, persistent) beliefs, referred to as hallucinations and delusions,
respectively. Psychosis can occur across an array of psychiatric and neurological disorders and
is perhaps most traditionally associated with schizophrenia. An intriguing and compelling point
that has been frequently made with respect to psychosis is that a core abnormality appears to lie
in misattributionspecifically, misattribution of self-generated thoughts or acts to the external
world. Thus, for example, an auditory hallucination in the form of a voice has been conceived as
a misattribution of internal speech to an external agent and, hence, its perception as some
external agent speaking. A particularly interesting feature of schizophrenia in this regard is the
passivity phenomenon wherein the sufferer perceives him- or herself to be the passive recipient
of actions, thoughts, or emotions. Such experiences are highly complex but do seem to entail, at
least (p.358) in part, a disturbed SoA. One striking experience is the delusion of motor control,
wherein someones motor actions feel as though they have been externally caused. Indeed, they
may be attributed to some unseen persecutor who, in some mysterious way, has attained control
over the sufferers body. Notably, this may occur in the context of movements that were
preceded by an intention to move and that fulfilled the object of this intention.
Aside from the fact that they pose something of a challenge to any simple distinction between
hallucinations and delusions (seeming simultaneously to entail both an abnormal perception and
an abnormal belief), delusions of motor control represent a profound explanatory challenge. In
the context of intended movements, successfully carried out, what might account for such an
unusual experience? One compelling response to this challenge emerges from the comparator
model of motor control alluded to earlier. To recap, this model suggests that an intention toward
action engenders, via an inverse model, a series of planned movements that leads, via a forward
model, to a predicted state (the new set of sensory and proprioceptive input that should obtain
following the successful completion of those movements). Critically, there is a comparison
between these two states (the predicted and the actual state), which of course is necessary to
modulating movement in order to achieve ultimately the intended goal. A small or absent
discrepancy between the two states signifies a successful movement, but it is also a hallmark of
agency: after all, a large discrepancy, signaling a surprising state following a movement, would
be a sign that the movement was neither intended nor under control. This simple model makes a
prediction elegantly tested and confirmed by Shergill and colleagues: specifically, actions for
which one is the agent will have predictable sensory consequences, which can then be canceled
or ignored; that is, self-generated actions will be associated with dampened sensory
consequences. They tested this using a force-matching task in which participants used self-
generated force to match a force that had just been applied to them externally. Results suggest
that self-generated force is indeed experienced as less than externally generated force (Shergill
et al., 2003).
The comparator model may be simply extended to explain delusions of motor control. The failure
to construct a forward model leads to erroneous predictions of the consequences of ones
actions, which are therefore not dampened and, hence, the action has the hallmarks not of self-
but of external generation. The evidence for this comes from observations that people with
Page 10 of 23
schizophrenia show results on the force-matching task suggestive of failed sensory dampening
and, indeed, find their own actions more intense than do control subjects (Shergill et al., 2005).
As discussed, an alternative perspective is offered by an Active Inference model in which it is a

prediction error generated by a mismatch between the (p.359) current state and the predicted
future state that forms the drive to action (mediated by reflex arcs). Importantly, in such models,
the instantiation of an action necessarily entails an alteration (attenuation) in the precision, and
hence weighting, of sensory evidence during the movement. This attenuation prevents a sensory
prediction error from stifling the movement and, furthermore, can be invoked to explain (see
Brown et al., 2013) the sensory attenuation illusion as well as the absence of such attenuation in
noted in schizophrenia.
Anosognosia for Hemiplegia

Anosognosia for hemiplegia (AHP) is the apparent inability to acknowledge ones contralesional
paralysis following stroke-induced perisylvian lesions, typically to the right hemisphere (though
there are exceptions; Cocchini et al., 2009). This counterintuitive but prototypical neurological
disorder of motor unawareness typically lasts from days to weeks, but even at the acute stages
of the illness patients may show strong delusional beliefs and corresponding emotional attitudes
toward the paralyzed body parts. In some patients such beliefs seem to be linked to their claims
that their limbs have moved even upon demonstration of the opposite (illusory movements;
Feinberg et al., 2000; Fotopoulou et al., 2008), while others may admit their on-line failures, but
fail to update their off-line body awareness, or acknowledge the functional consequences of
their paralysis (Carruthers, 2008; Marcel et al., 2004). Interestingly, despite such illusions and
delusions, some of these patients may show implicit awareness of their deficits in verbal
(Fotopoulou et al., 2010) or behavioral tasks (Cocchini et al., 2010; Moro et al., 2011; Nardrone
et al., 2007), and they may be more aware of their motor failures when taking a third- as
opposed to a first-person perspective (Fotopoulou et al., 2009, 2011; Marcel et al., 2004).
Initially, AHP was explained as the secondary consequence of one or more of the concomitant
sensorimotor and cognitive impairments that frequently accompanied it, such as primary
sensorimotor deficits, generalized cognitive impairment, or neglect (for reviews, see Jehkonen et
al., 2006; Orfei et al., 2007). Several studies, however, have revealed double dissociations
between AHP and such impairments (e.g. Bisiach et al., 1986; Marcel et al., 2004), suggesting
that they are not necessary for AHP to occur, although they could act as contributing factors.
Influenced by the aforementioned comparator models, some investigators have argued that
AHP is not the secondary consequence of deficits in other domains, but rather the primary
outcome of abnormalities in encapsulated and modular mechanisms of anticipatory motor
awareness (Berti et al., 2005; Frith et al., 2000; Heilman et al., 1998). Not all of these
perspectives, however, (p.360) propose the same component of the comparator model as
responsible for AHP. Heilman and colleagues (Heilman et al., 1998) have proposed that AHP
arises from a failure to form motor intentions, resulting in the forward model not priming the
comparator to expect movement, and hence patients never discover that they cannot move.
Frith and colleagues (2000) alternatively proposed that although patients with AHP are able to
predict the expected sensory consequences of intended movements, they fail to register the
Page 11 of 23
discrepancy between predicted and actual sensory feedback because of visuospatial neglect or
other sensory deficits. Berti and colleagues (Berti et al., 2005) follow Frith and colleagues in
proposing that patients are unaware of the discrepancy between intended and actual movement;
however, they suggest that this failure to detect discrepancies is the result of damage directly to
the comparator. The latter two hypotheses are capable of accounting for both the negative
(unawareness of motor failures) and the positive (the illusory awareness of having moved) signs
of AHP, while the theory of Heilman and colleagues addresses only the negative signs.
Empirical findings thus far have supported mainly the explanation put forward by Berti and
colleagues (see Jenkinson & Fotopoulou, 2010, for review). A lesion-mapping study (Berti et al.,
2005) revealed that the brain areas involved in monitoring the correspondence between motor
commands and sensory feedback (i.e., Brodmann premotor areas 6 and 44 and the insular
cortex) are selectively damaged in patients with AHP, while areas typically responsible for motor
planning (e.g., supplementary motor cortex) are intact in these patients. Moreover, there is
physiological (Berti et al., 2007; Hildebrandt & Zieger, 1995; but see Gold et al., 1994) and
behavioral (Garbarini et al., 2012; Jenkinson, Edelstyn, & Ellis, 2009) evidence for the presence
of intact motor intentions in AHP. A further study showed for the first time the direct relation
between motor intention and awareness (Fotopoulou et al., 2008). Specifically, patients illusory
awareness of movement, and related feelings and judgements of agency (was it you or someone
else who performed the action?) reflected an abnormal, selective dominance of motor intentions
over visual feedback about the actual effects of movement (elicited by a realistic rubber hand
that patients assumed was their own). Further, this effect could not be explained by neglect.
Despite the clear value of the feed-forward hypotheses, it has become apparent that a strictly
modular, motor explanation is not sufficient to account for all the manifestations of AHP (e.g.,
Fotopoulou et al., 2010; Orfei et al., 2009; Vocat et al., 2010). Feed-forward theories are
valuable in explaining the illusion of moving, but AHP patients do not simply claim that they
have the phenomenal experience of moving. Instead, they ignore the wealth of contrary evidence
and medical signs indicating that they are paralyzed (e.g., their medical results, disabilities,
occasional accidents, and others feedback) and they (p.361) adhere to the delusional belief
that they have functional limbs, showing corresponding emotional attitudes. The explanation of
such beliefs and attitudes requires the postulation of additional dysfunctions that prevent
sensorimotor and other failures from being re-represented at a higher level of cognitive and
emotional self-representation, one that is beyond the sensorimotor domain.
This is the point at which predictive coding schemes become useful. Unlike optimal motor
control theory, such models can envision a mismatch between prediction and experience in
various levels of the neurocognitive hierarchy and in relation to several cognitive and emotional
domains. For example, they can explain the motor illusions of patients who claim they have
moved their arms as planned, even upon demonstration of the contrary (Fotopoulou et al.,
2008), but they can also explain the more general, obstinate adherence of other patients to their
premorbid everyday habits (of course, I can walk) despite implicit knowledge of their paralysis
(Fotopoulou et al., 2010).
Specifically, aberrant perceptual inference (suboptimal synaptic activity; Friston, 2010) can be
caused by deficits that lead to weak, absent, or unreliable prediction errors, and hence lead
Page 12 of 23
patients to base their inference on premorbid, non-updated predictions about their motor
abilities (see also Fotopoulou, 2012, 2013). Such deficits may well occur in relation to
exteroceptive signals about the left side of the body as represented in the connections of right
hemisphere subcortical areas (e.g., the thalamus), or re-represented and organized in cortical
functional networks of the right hemisphere (Berti et al., 2005; Fotopoulou et al., 2010; Moro et
al., 2011; Vocat et al., 2010). However, as aforementioned, these deficits are not necessary to
cause AHP. They can readily explain illusions of moving (like the comparator model does), but
not the more general delusion of being able to move. Interestingly, recent lesion-mapping
studies have highlighted that areas such as the insula, limbic structures, and subcortical white
matter connections may be selectively associated with AHP (Fotopoulou et al., 2010; Karnath et
al., 2005; Moro et al., 2011; Vocat et al., 2010). Such areas are linked with interoception and
motivation and are specifically implicated in bodily salience and interoceptive awareness (Craig,
2003; Critchley et al., 2004). Thus, we propose that weak or imprecise (see also below)
interoceptive signals about the current state of the body may lead to persistent adherence to
past expectations of how the affected body parts should feel leading to the ensuing aberrant
beliefs. Moreover, given the position of such priors in the neurocognitive hierarchy, such faulty
inference may also explain away contrary exteroceptive signals in some patients. To use the
words of one anosognosic patient who also denied the ownership of his paralyzed limbs, But my
eyes and my feelings dont agree, and I must believe my feelings. I know they [left arm and leg]
look like mine, but I can feel they are not, and I cant believe my eyes (C. W. Olsen, 1937, cited
in Feinberg, 1997).
(p.362) It is worth noting that the principle of Free Energy minimization may enhance the
understanding of AHP and motor awareness in at least one additional way. First, the recently
identified lesions in frontostriatal circuits (Fotopoulou et al., 2010; Moro et al., 2011; Venneri &
Shanks, 2004; Vocat et al., 2010) may have a modulatory role in AHP, leading to dopamine
depletion and a more general difficulty in optimizing the precision (uncertainty) of prediction
errors (Friston et al., 2012), affecting their salience and ultimately both short- and long-term
learning (suboptimal synaptic gain and plasticity; Friston, 2010). Indeed, the functional role of
the basal ganglia and particularly the striatum has been linked with prediction error-driven
learning (ODoherty et al., 2003) as well as the aberrant salience theories of psychosis (Gray et
al., 1991; Kapur, 2003), which as we saw have direct implications for the formation of delusions
about ones agency in psychosis. In AHP such deficits can be linked with both specific instances
of aberrant motor monitoring in functionally specialized systems (Berti et al., 2005), or more
generally in global error monitoring (Davies et al., 2005; Venneri & Shanks, 2004; Vocat et al.,
2012), mental flexibility (Levine et al., 1991), and surprise detection (Ramachandran, 1995)
deficits. Indeed, a recent study showed that AHP patients had the tendency to jump to
conclusions on the basis of limited and rather vague information and then to subsequently get
stuck to their former false beliefs instead of modifying them based on novel, arguably more
salient information (Vocat et al., 2012).
Anarchic Hand
Another counterintuitive neuropsychiatric syndrome challenges the common-sense notion that
our actions are caused by a central, unitary will (Libet, 1983). Contrary to anosognosic patients
who may erroneously feel that they have executed intended movements with their paralyzed
Page 13 of 23
limbs, patients with anarchic hand sign (AHS) have usually distressing experience of one of
their arms acting without being guided by the patients conscious will (for a review, see Fisher,
2000). The hand is indeed executing the movements that patients report, but it appears to act
entirely on its own accord, grabbing objects that the patient had no known intention of wanting
to grab, or not releasing objects the patient wishes to release. Interestingly, the actions that the
affected arm performs, although not intended, are nevertheless purposeful in themselves (e.g.,
open a door, take off clothes) and are typically completed successfully. In some patients the
affected arm may even hinder purposeful actions of the other arm (intermanual conflict, or
diagonistic dyspraxia). The patients do not deny the ownership of the hand (see also below), nor
the fact that their own (p.363) body is actually executing the actions. However, they experience
the hand as having a will of its own, or of being controlled by external agents.
There appears to be some long-standing and persevering taxonomical and nosological confusion
in the neurological literature of this syndrome. For starters, the term alien hand syndrome or
sign is frequently used to refer not only to the presence of involuntary, uncontrolled movements
in ones limbs but also to feelings of non-belonging (lack of ownership) for the affected limb in
the absence or presence of such involuntary movements. Thus, Della Sala and colleagues (e.g.
Marchetti & Della Sala, 1998) proposed restricting the term alien hand to conditions involving
the feeling of non-belonging of a hand (lack of the sense of ownership for the arm) and
introduced anarchic hand to refer to conditions where subjects perform involuntary
movements with their hand but acknowledge the ownership of the same hand. We follow this
distinction in this chapter and below we attempt to explain the neuropsychological mechanisms
that are responsible for the AHS specifically.
The AHS can be caused by stroke, midline tumors, corticobasal degeneration, or callosotomy for
epilepsy, and it can be as variable as it is rare. In its most typical and dramatic form, AHS arises
following lesions to the medial frontal lobes and the corpus callosum. However, the syndrome
can also be observed in association with subcortical (mainly to the thalamus and basal ganglia)
and posterior (parietal cortex and posterior corpus callosum) lesions. Various related
distinctions have been proposed, such as anterior versus posterior AHS, or frontal versus
callosal AHS, but their validity and relation to laterality remains unclear (see Kikkert et al.,
2006, for review). It seems more appropriate to label the syndrome according to the affected
hemisphere in each case, as well as to specify which of the common behavioral symptoms are
present in each case; for example, Aboitiz et al. (2003) suggest that apart from the spontaneous,
involuntary movements of the affected hand, the general syndrome could comprise (1)
diagnostic dyspraxia (intermanual conflict), (2) alien hand (disownership of the hand, see also
above), (3) supernumery hands (the experience of additional arms belonging to the patient), and
(4) agonistic dyspraxia (involuntary movements of the affected arm with temporary inhibition of
the other intact arm following bimanual instructions).
Experimental and functional neuroimaging studies in AHS are rare, and comparisons between
them are hampered by the above taxonomical issues. However, most studies seem to have
focused on two functional abnormalities of the anarchic behavior. First, in patients with AHS,
simple observation of certain objects with strong motor affordances (Gibson, 1979) might be
sufficient to elicit the associated motor plan for interacting with that object and lead to the
corresponding action, even when such action conflicts with other motor goals (e.g., Humphreys
Page 14 of 23
& Riddoch, 2000; Riddoch et al., 1998). Such (p.364) affordances effects have been long
shown in healthy individuals, for example in paradigms where visual objects and shapes can
prime specific actions. The existence of such effects makes sense if one thinks that, in order to
act successfully upon the world, we need to have readily available information about the position
and shape of objects in relation to both our potentially moving body and the environment. This
means that we need to be able to represent the relation between the two in various frames of
reference and translate between them, as for example when we need to use visual cues, coded in
retinotopic coordinates, to make a limb movement (coded in body-centered coordinates) toward
an object (Jeannerod et al., 1995). There is well-established evidence that the parietal cortex
supports such transformation abilities and maintains representations of the appropriate
movement trajectories needed to reach and grasp objects of previously learned affordances.
However, in patients with AHP the effects of visual affordances on action appear to be stronger
in the anarchic hand relative to the unaffected hand (McBride et al., 2013), and such effects
seem to actually lead to involuntary actions.
The latter observation brings us to the second reported abnormality in AHS. It has been
proposed that these patients lack the ability to voluntarily suppress or inhibit the actions primed
by the perceptual processing of objects (e.g., Biran et al., 2006; Giovannetti et al., 2005;
Schaefer et al., 2010). However, more recently there is growing evidence in healthy volunteers
that unconsciously primed responses (for reviews, see Eimer and Schlaghecken, 2003; McBride
et al., 2012), or responses afforded by the properties of objects themselves (e.g., Vainio, 2009;
Vainio & Mustonen, 2011), can also be automatically and unconsciously inhibited. Moreover,
such inhibition seems to be causally linked to the functional role of medial frontal cortex
regions. For example, while the use of a recently developed masked-prime task (Eimer &
Schlaghecken, 2003) leads to an automatic inhibition of unwanted motor plans (activated
unconsciously by condition-action associations) in healthy individuals, a recent study
administered this paradigm to two patients with highly selective lesions to the supplementary
eye field and motor area and found that such inhibition does not take place for hand and eye
movements, respectively (Sumner et al., 2007). Consistently, the same priming paradigm has
been used in patients with AHS (where damage to such medial frontal cortex areas is common
but may not be as specific) to show that such automatic inhibition no longer takes place, and
instead perceptually processed affordances are automatically translated into the execution of
the corresponding actions (McBride et al., 2013).
From the perspective of the comparator model, the AHS has been explained as an inappropriate
activation of the parietal cortex areas responsible for the perceptual and visuomotor processing
of objects due to damage to the supplementary motor cortex (SMA), which is normally
responsible for action selection and (p.365) corresponding inhibitory functions (Frith et al.,
2000; see also above). In terms of the relation of such damage to the motor control model, it is
assumed that the current motor intentions of the patient are no longer able to inhibit the
visuomotor effects afforded by the environment and thus patients execute whatever action is
afforded by the object in front of them. Furthermore, according to the model, our motor
awareness derives from the various comparisons between the desired, predicted, and actual
state of the body and, as aforementioned, we normally have a limited awareness of our motor
commands themselves and the way immediate sensory information (affordances) are used to
fine-tune such commands (the ways in which the controller selects and corrects the precise
Page 15 of 23
commands required for an action). The areas that underlie the desired, predicted, and actual
states of the body are considered intact in patients with AHS and hence these patients are
aware of the fact that their arm executes actions that differ from their own, conscious
intentions.
As described above, from the perspective of predictive coding and active inference, descending
signals from higher motor areas to the periphery are conceptualized as proprioceptive
predictions (not motor commands) and the corresponding ascending signals (proprioceptive
prediction errors) are understood to fulfill such priors by leading to action and correcting
predictions at lower levels of the hierarchy, rather than passing up to higher areas and changing
the intended actions. Thus, in the context of this model, the fact that patients with AHS do
execute successful movements with their affected arm and are aware of having executed them
suggests that at some level in the sensorimotor hierarchy proprioceptive predictions were
formed and fulfilled. Moreover, unlike in the case of delusions of alien control (see above), one
may expect that the precision of sensory prediction errors is attenuated to some degree as
patients do not attribute their actions to a different agent. It seems, however, that the formed
and fulfilled proprioceptive predictions have been generated in parietal or subcortical
sensorimotor areas as a result of external affordances and without the top-down involvement of
certain damaged, or disconnected (by callosal damage) higher-order motor areas such as the
SMA. Thus, we speculate that large prediction errors arise at high levels of the hierarchy
because of the discrepancy between the sensory predictions of spared higher order
sensorimotor motor areas and the prediction errors (posterior beliefs about the executed
movements) conveyed by lower motor areas.
Conclusion
Though ubiquitous and compelling, our sense of being the agent of our actions is complex,
mutable, and unreliable. Moreover, the scientific frameworks and means of measurement used
to research this subjective experience are necessarily approximate and, at times, questionable.
Against this background we (p.366) have reviewed two types of model used as frameworks for
understanding SoA and have highlighted key areas in which they differ. Importantly, we argue
that neuropsychiatric disorders may provide very useful windows onto agency. Interestingly,
with some exceptions, such disorders have not, yet, been the subject of comprehensive and
systematic studies capitalizing on technological advances in neuroscience. It seems likely that
the increasing sophistication of computational theories of motor function may provide a useful
platform in this regard. It is also worth noting in closing that such models, as they become more
general and depart from a strictly motor emphasis (for example, moving from comparator
models to predictive coding and active inference models), may offer ways of providing a more
comprehensive understanding of the whole range of features accompanying such syndromes.
References
Aboitiz, F., Carasco, X., Schrter, C., Zaidel, D., Zaidel, E., & Lavados, M. (2003). The alien hand
syndrome: classification of form reported and discussion of a new condition. Neurological
Sciences, 24, 252257.
Page 16 of 23
Adams, R. A., Shipp, S., & Friston, K. J. (2013). Predictions not commands: active inference in
the motor system. Brain Structure & Function, 218(3), 611643.
Berti, A., Bottini, G., Gandola, M., Pia, L., Smania, N., Stracciari, A., Castiglioni, I., Vallar, G.,
Paulesu, E. (2005). Shared cortical anatomy for motor awareness and motor control. Science,
309, 488491.
Biran, I., Giovannetti, T., Buxbaum, L., and Chatterjee, A. (2006) The alien hand syndrome: what
makes the alien hand alien? Cognitive Neuropsychology, 23(4), 563582.
Bisiach, E., Vallar, G., Perani, D., Papagno, C., Berti, A. (1986). Unawareness of disease
following lesions of the right hemisphere: anosognosia for hemiplegia and anosognosia for
hemianopia. Neuropsychologia, 24, 471482.
Brown, H., Adams, R. A., Parees, I., Edwards, M., Friston, K. (2013). Active inference, sensory
attenuation and illusions. Cognitive Processing, 14, 411427.
Carruthers, G. (2008). Types of body representation and the sense of embodiment.

Cocchini, G., Beschin, N., Cameron, A., Fotopoulou, A., & Della Sala, S. (2009). Anosognosia for
motor impairment following left brain damage. Neuropsychology, 23(2), 223230.
Cocchini, G., Beschin, N., Fotopoulou, A., & Della Sala, S. (2010). Explicit and implicit
anosognosia for upper limb motor impairment. Neuropsychologia, 48(5), 14891494.
Craig, A. D. (2003). Interoception: the sense of the physiological condition of the body. Current
Opinion in Neurobiology, 13, 500505. doi: 10.1016/S0959-4388(03)00090-4
Critchley, H. D., Wiens, S., Rotshtein, P., hman, A., & Dolan, R. D. (2004). Neural systems
supporting interoceptive awareness. Nature Neuroscience, 7, 189195.
Crosson, C., Barco, P., Velozo, C., Bolesta, M., Cooper, P., Werts, D., et al. (1989). Awareness
and compensation in postacute head injury rehabilitation. Journal of Head Trauma
Rehabilitation, 4, 4654.
Davies, M., Davies, A. A., & Coltheart, M. (2005). Anosognosia and the two-factor theory of
delusions. Mind and Language, 20, 209236.
Desmurget, M., & Sirigu, A. (2009) A parietal-premotor network for movement intention and
motor awareness. Trends in Cognitive Sciences, 13, 411419.
Eimer, M., & Schlaghecken, F. Response facilitation and inhibition in subliminal priming.
Biological Psychology, 64(1e2): 7e26, 2003.
Ernst, M. O., & Bulthoff, H. H. (2004). Merging the senses into a robust percept. Trends in
cognitive sciences, 8(4), 162169
Page 17 of 23
Feinberg T. E. (1997). Anosognosia and confabulation, In: Feinberg T.E., Farah M. editors.
Behavioral neurology and neuropsychology. New York: McGraw-Hill, 1997: 369390.
Feinberg, T. E., Roane, D. M., & Ali, J. (2000). Illusory limb movements for anosognosia for
hemiplegia. Journal of Neurology, Neurosurgery and Psychiatry, 68, 511513.
Feynman, R. P. (1972). Statistical Mechanics. Reading: Benjamin.
Fisher, C. M. (2000). Alien hand phenomena: a review with the addition of six personal cases.
The Canadian Journal of Neurological Sciences, 27(3): 192203.
Fotopoulou, A. (2014). Time to get rid of the modular in neuropsychology: a unified theory of
anosognosia as aberrant predictive coding. Journal of Neuropsychology, Mar;8(1), 119.
Fotopoulou, A. (2014) Beyond the reward principle: consciousness as precision seeking.

Neuropsychoanalysis, 15(1), 3338.
Fotopoulou, A. (2012). Illusions and delusions in anosognosia for hemiplegia: from motor
predictions to prior beliefs. Commentary. Brain, 135, 13441347.
Fotopoulou, A., Tsakiris, M., Haggard, P., Rudd, A., & Kopelman M. (2008). The role of motor
intention in motor awareness: an experimental study on anosognosia for hemiplegia. Brain, 131,
34323442.
Fotopoulou, A., Jenkinson, P. M., Tsakiris, M., Haggard, P., Rudd, A., & Kopelman, M. D. (2011).
Mirror-view reverses somatoparaphrenia: dissociation between first- and third-person
perspectives on the body. Neuropsychologia, 49(14), 39463955.
Fotopoulou, A., Pernigo, S., Maeda, R., Rudd, A., & Kopelman, M. A. (2010). Implicit awareness
in anosognosia for hemiplegia: unconscious interference without conscious re-representation.
Brain, 133(Pt 12), 35643577.
Fotopoulou, A., Rudd, A., Holmes, P., & Kopelman, M. (2009). Self-observation reinstates motor
awareness in anosognosia for hemiplegia. Neuropsychologia, 47(5), 12561260.
Friston, K. (2005). A theory of cortical responses. Philosophical transactions of the Royal Society
B: Biological sciences, 360(1456): 815836.
Friston, K. (2010). The free-energy principle: a unified brain theory. Nature Reviews:
Friston, K. J., Shiner, T., FitzGerald, T., Galea, J. M., Adams, R., Brown, H., Dolan, R. J., Moran,
R., Stephan, K. E., & Bestmannm, S. (2012). Dopamine, affordance and active inference. PLoS
Computational Biology, 8(1), e1002327.
Page 18 of 23
Frith, C. D., Blakemore, S. J., & Wolpert, D. M. (2000) Abnormalities in the awareness and
control of action. Philosophical Transactions of the Royal Society of London, Series B. Biological
Sciences, 355, 17711788.
Gallagher, S. (2000). Philosophical concepts of the self: implications for cognitive sciences.
Garbarini, F., Rabuffetti, M., Piedimonte, A., Pia, L., Ferrarin, M., Frassinetti, F., Gindri, P.,
Cantagallo, A., Driver, J., & Berti, A. (2012). Moving a paralysed hand: bimanual coupling
effect in patients with anosognosia for hemiplegia. Brain, 135(Pt 5), 14861497.
Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mifflin.
Giovannetti, T., Buxbaum, L. J., Biran, I., & Chatterjee, A. (2005) Reduced endogenous control in
alien hand syndrome: evidence from naturalistic action. Neuropsychologia, 43(1), 75e88.
Gold, M., Adair, J. C., Jacobs, D. H., & Heilman, K. M (1994). Anosognosia for hemiplegia: an
electrophysiologic investigation of the feed-forward hypothesis. Neurology, 44, 18041808.
Gray, J. A., Feldon, J., Rawlins, J. N. P., Hemsley, D., & Smith, A. D. (1991). The neuropsychology
of schizophrenia. Behavioral and Brain Sciences, 14, 184.
Heilman, K. M., Barret, A. M., & Adair, J. C. (1998). Possible mechanisms of anosognosia: a
defect in self-awareness. Philosophical Transactions of the Royal Society of London, Series B.
Biological Sciences, 353, 19031909.
Hildebrandt, H., & Zieger, A. (1995). Unconscious activation of motor responses in a hemiplegic
patient with anosognosia and neglect. European Archives of Psychiatry and Clinical
Hinton, G. E., & van Camp, D. (1993). Keeping neural networks simple by minimising the
description length of weights. In: COLT 93 Proceedings of the Sixth Annual Conference on
Computational Learning Theory. New York: Association of Computing Machinery, 513.
Humphreys, G. W., & Riddoch, M. J. (2000). One more cup of coffee for the road: object-action
assemblies, response blocking and response capture after frontal lobe damage. Experimental
Brain Research, 133(1): 8193.
Jehkonen, M., Laihosalo, M., & Kettunen, J. (2006). Anosognosia after stroke: assessment,
occurrence, subtypes and impact on functional outcome reviewed. Acta Neurologica
Scandinavica, 114, 293306.
Jenkinson, P., & Fotopoulou, A. (2010). Motor awareness in anosognosia for hemiplegia:
experiments at last. Experimental Brain Research, 204(3), 295304.
Jenkinson, P. M., Edelstyn, N. M., Drakeford, J. L., & Ellis, S. J. (2009). Reality monitoring in
anosognosia for hemiplegia. Consciousness and Cognition, 18, 458470.
Page 19 of 23
Jeannerod, M., Arbib, M. A., Rizzolatti, G., & Sakata, H. (1995). Grasping objects: the cortical
mechanisms of visuomotor transformation. Trends in Neurosciences, 18, 314320.
Kapur, S. (2003). Psychosis as a state of aberrant salience: a framework linking biology,

phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry, 160, 13
23.
Karnath, H. O., Baier, B., & Ngele, T. (2005). Awareness of the functioning of ones own limbs
mediated by the insular cortex? Journal of Neuroscience, 3, 25(31), 71347138.
Kikkert, M. A., Ribbers, G. M., & Koudstaal, P. J. (2006). Alien hand syndrome in stroke: a report
of 2 cases and review of the literature. Archives of Physical Medicine and Rehabilitation, 87,
728732.
Levine, D. N., Calvanio, R., & Rinn, W. E. (1991). The pathogenesis of anosognosia for
hemiplegia. Neurology, 41, 17701781.
in relation to onset of cerebral activity (readiness potential): the unconscious initiation of a
freely voluntary act. Brain, 106, 623642.
Marcel, A. J., Tegner, R., & Nimmo-Smith, I. (2004). Anosognosia for plegia: specificity,
extension, partiality, and disunity of bodily unawareness. Cortex, 40, 1940.
Marchetti, C., & Della Salla, S. (1998) Disentangling the alien and anarchic hand. Cognitive
Neuropsychiatry, 3, 191208.
McBride, J., Sumner, P., Jackson, S. R., Bajaj, N., & Husain, M. (2013). Exaggerated object
affordance and absent automatic inhibition in alien hand syndrome. Cortex, 49(8), 20402054.
McBride, J., Boy, F., Husain, M., & Sumner, P. (2012). Automatic motor activation in the
executive control of action. Frontiers in Human Neuroscience, 6, 82.
Mesulam M. (2012). The evolving landscape of human cortical connectivity: facts and
inferences. Neuroimage, 62(4), 21822189.
Moore, J. W., & Fletcher, P. C. (2012) Sense of agency in health and disease: a review of cue
Moore, J. W., Middleton, D., Haggard, P., & Fletcher, P. C. (2012) Exploring implicit and explicit
aspects of sense of agency. Consciousness and Cognition, 21(4), 17481753.
Moore, J. W., Teufel, C., Subramaniam, N., Davis, G., & Fletcher, P. C. (2013). Attribution of
intentional causation influences the perception of observed movements: behavioral evidence and
neural correlates. Frontiers in psychology, 4.
Page 20 of 23
Moro, V., Pernigo, S., Zapparoli, P., Cordioli, Z., & Aglioti, S. M. (2011). Phenomenology and
neural correlates of implicit and emergent motor awareness in patients with anosognosia for
hemiplegia. Behavioural Brain Research, 225(1), 259269.
Nardone, I. B., Ward, R., Fotopoulou, A., & Turnbull, O. E. (2007) Attention and emotion in
anosognosia: evidence of implicit awareness. Neurocase, 13, 438445.
ODoherty, J. P., Dayan, P., Friston, K., Critchley, H., & Dolan, R. J. (2003) Temporal difference
models and reward-related learning in the human brain. Neuron, 38(2), 329337.
Orfei, M. D., Robinson, R. G., Prigatano, G. P., Starkstein, S., Rsch, N., Bria, P., Caltagirone, C.,
& Spalletta, G. (2007). Anosognosia for hemiplegia after stroke is a multifaceted phenomenon: a
systematic review of the literature. Brain, 130(Pt 12), 30753090.
Orfei, M. D., Caltagirone, M., & Spalletta, G. (2009). The evaluation of anosognosia in stroke
patients. Cerebrovascular Disease, 27, 280289.
Pacherie, E. (2007). The anarchic hand syndrome and utilization behavior: a window onto
agentive self-awareness. Functional Neurology, 22(4) 211217.
Ramachandran, V. S. (1995). Anosognosia in pariental lobe syndrome. Consciousness and

Cognition, 4, 2251.
Rao, R. P., & Ballard, D. H. (1999). Predictive coding in the visual cortex: A functional
interpretation of some extra-classical receptive field effects. Nature Neuroscience, 2, 7987.
Riddoch, M. J., Edwards, M. G, Humphreys, G. W., West, R., & Heafield, T. (1998). Visual
affordance direct action: neuropsychological evidence from manual interference. Cognitive
Neuropsychology, 15(6/7/8): 645e683.
Sarrazin, J. C., Cleeremans, A., & Haggard, P. (2008) How do we know what we are doing?
Time, intention and awareness of action. Consciousness and Cognition, 17(3), 602615.
Schaefer, M., Heinze, H. J., & Galazky, I. (2010). Alien hand syndrome: neural correlates of
movements without conscious will. PLoS One, 5(12), e15010.
Schacter, D. L. (1990). Toward a cognitive neuropsychology of awareness: implicit knowledge

and anosognosia. Journal of Clinical and Experimental Neuropsychology, 12, 155178.
Searle, J. R. (1983). Intentionality: an essay in the philosophy of mind. Cambridge: Cambridge

University Press.
Seth, A. K., Suzuki, K., & Critchley, H. D. (2012). An interoceptive predictive coding model of
conscious presence. Frontiers in Psychology, 2(395), 116.
Shergill, S. S., Bays, P. M., Frith, C. D., & Wolpert, D. M. (2003). Two eyes for an eye: the
neuroscience of force escalation. Science, 301(5630), 187.
Page 21 of 23
Shergill, S. S., Samson, G., Bays, P. M., Frith, C. D., & Wolpert, D. M. (2005) Evidence for
sensory prediction deficits in schizophrenia. American Journal of Psychiatry, 162(12), 2384
2386.
Sumner, P., Nachev, P., Morris, P., Peters, A. M., Jackson, S. R., Kennard, C., & Husain, M.
(2007). Human medial frontal cortex mediates unconscious inhibition of voluntary action.
Neuron, 54(5), 697711.
two-step account of agency. Conscious and Cognition, 17(1), 219239.
Todorov, E. (2004). Optimality principles in sensorimotor control. Nature Neuroscience, 7(9),

907915.
Todorov, E., & Jordan, M. I. (2002). Optimal feedback control as a theory of motor coordination.
Tsakiris, M., & Fotopoulou, A. (2008). Is my body the sum of online and offline body
representations? Consciousness and Cognition, 17(4), 13171320.
Vainio, L. (2009). Interrupted object-based updating of reach program leads to a negative

compatibility effect. Journal of Motor Behavior, 41(4), 305315.
Vainio, L., Hammaren, L., Hausen, M., Rekolainen, E., & Riskila, S. (2011) Motor inhibition
associated with the affordance of briefly displayed objects. The Quarterly Journal of
Experimental Psychology, 64(6), 10941110.
Vainio, L., & Mustonen, T. (2011). Mapping the identity of a viewed hand in the motor system:
evidence from stimulus-response compatibility. Journal of Experimental Psychology: Human
Perception and Performance, 37(1), 207221.
Venneri, A., & Shanks, M. F. (2004). Belief and awareness: reflections on a case of persistent
anosognosia. Neuropsychologia, 42, 230238.
Vocat, R., Staub, F., Stroppini, T., & Vuilleumier, P. (2010). Anosognosia for hemiplegia: a
clinical-anatomical prospective study. Brain, 133(Pt 12), 35783597.
Vocat, R., Saj, A., & Vuilleumier, P. (2012). The riddle of anosognosia: does unawareness of
hemiplegia involve a failure to update beliefs? Cortex, 49(7):17711781.
Wegner, D. M, & Wheatley, T. (1999) Apparent mental causation: sources of the experience of
will. American Psychologist, 54(7), 480492.
Wolpert, D. M. & Kawato, M. (1998). Multiple paired forward and inverse models for motor
control. Neural Networks, 11, 13171329.
Page 22 of 23
Page 23 of 23
Action Generation, Intention, and Agency in Motor and Body Awareness
Deficits

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Action Generation, Intention, and Agency in Motor and Body Awareness Deficits
Anna Berti
Francesca Garbarini
Lorenzo Pia
DOI:10.1093/acprof:oso/9780190267278.003.0017

The dramatic impairment of motor and body awareness, often observed in brain-damaged
patients, can shed light on the mechanisms implicated in the emergence of conscious experience
of the self as an acting body. This chapter first reviews evidence that shows, in patients affected
by anosognosia for hemiplegia, how the activation of normal intentional process can give rise to
false beliefs of movement when predictive models of motor behavior cannot be confronted with
sensory feedbacks, due to damage of motor control systems. In brain-damaged patients with
motor and somatosensory deficits, puzzling cognitions and behaviors can sometimes be
observed where patients can misidentify other peoples limbs as their own, showing a
pathological embodiment (E+) of others body parts (a condition called somatoparaphrenia). The
chapter notes that in this condition the sense of agency and the intentional attitude can be
normally transferred to the movements of anothers limbs.
Keywords: impairment, brain damage, anosognosia for hemiplegia
Introduction
Interaction with other individuals and with the environment is mediated by motor actions
through which people try to achieve their goals and purposes. Action is generated through a
chain of neurobiological events that is often not available to consciousness, although we are
usually aware of moving (or not moving) different parts of our body. Therefore, one has the
conscious knowledge that, for instance, ones hand is moving, as well as intentional attitudes,
defined as the felt urge to make a movement that precedes the actual execution of a specific
Page 1 of 16
Deficits
motor act. We also feel a strong sense of controlling our own actions (usually indicated as sense
of agency), and that our own body performs those actions. The normal integration between the
different aspects of conscious motor control and the feeling of body ownership (i.e., the
experience that bodily states are so clearly and inexorably mine) are considered to be crucial
for the building up of a coherent sense of identity. Therefore, it has been proposed that one of
the most fruitful ways to refer to the conscious self is to conceive it as intimately related to the
concept of body as source or power to action, i.e. as the variety of motor potentialities that
define the horizon of the world in which we live (Gallese & Sinigaglia, 2010). Accordingly, in
the present chapter, we shall review some studies that investigate which processes are (p.372)
critical when people have beliefs about moving parts of their bodies, and which kind of body
experiences are involved in shaping and/or modeling the various senses of motor self-awareness
(see also Tsakiris, Chapter 10 of this volume). Neurophysiological and neuropsychological
evidence suggest that the neural bases of the different kinds of self-awareness can be kept
relatively distinct and are discretely organized (Spinazzola et al., 2008). The relation between
them therefore needs to be clarified in order to capture the real nature and structure of the self.
Agency and Intention

The sense of agency can be defined as the feeling that I am the one who generates a willed
action. The relation between agency and intention seems to be strong: the experience of being in
control of ones own movements, and of their effects, is stronger when there is an explicit
intentional selection of a motor action than when the same movements are externally generated
by a mechanical device or by a transcranial magnetic stimulation (TMS)induced stimulation of
the motor cortex (Haggard & Clark 2003). For instance, in a recent study, healthy subjects
pressed a button, which caused a tone to occur on some trials. Their task was to judge the time
they became aware of either pressing the key, or of hearing the tone. When the key press
caused the tone, participants judged their key press as occurring later, and the tone earlier,
than when the two events occurred alone. The effect did not occur when an involuntary
movement (caused by stimulating the motor cortex using TMS) was followed by a tone, or when
subjects judged the timing of two causally related external sensory events. This temporal
attraction strongly suggests that there is a tight link between our experience of agency and
motor intentionality. Other evidence confirming how the sense of agency is modulated by motor
intentionality comes from brain-damaged patients affected by an unusual form of limb
misattribution and will be discussed later in this chapter (see E+ patients).
Intention and Motor Awareness

The relation between conscious intention and motor awareness has been studied in different
areas of psychology and cognitive neuroscience. In a seminal study, Libet and coworkers (Libet
et al., 1983) showed that although the awareness of the intention to move (the W judgment)
anticipates the actual movement by about 200 milliseconds, it follows by hundreds of
milliseconds (instead of preceding) the electrophysiological preparatory activity related to
movement (readiness potential), usually registered on the scalp over the supplementary motor
area. These results strongly suggested that the conscious experience of (p.373) intentionality is
a consequence of brain activity related to the process of programming and selecting the correct
movements for action (Haggard & Eimer, 1999; Libet et al., 1983). Libet and coworkers also
asked their subjects to estimate the time at which they became aware of executing a voluntary
movement (M judgment, or motor awareness). Participants timing of the M moment followed
Page 2 of 16
Deficits
the timing of the W moment, as expected. But, surprisingly, it precedes, instead of follows, the
actual initiation of the movement by 50 to 80 milliseconds. This outcome suggests that motor
awareness emerges before any sensory feedback reaches the brain, thus showing that the
feeling that a movement is produced is not linked in an absolute way to the feedback coming
from the moving muscles and joints. Although counterintuitive, these results are consistent with
other studies that showed that the sensations associated with the actual execution of movements
could be unnecessary for the construction of movement awareness. Similarly, Fourneret and
Jeannerod (Fourneret & Jeannerod, 1998) concluded that we are aware of the movement we
intend to perform, rather than of the movement we actually produce. Therefore, although
proprioception and vision are fundamental aspects of our capacity of judging the course and the
consequences of a motor event, motor awareness is somehow independent from their
operations, but is strictly related to intentionality. Blakemore and colleagues (Blakemore &
Frith, 2003) suggested that motor awareness must correspond to some neural signal that (a)
precedes the movement, (b) follows the development of a conscious intention, and (c) is formed
prior to the processing of sensory feedback. They proposed a model of how the control of motor
systems relates to various forms of awareness (the comparator model; Haggard, 2005; Wolpert
et al., 1995). According to the model, once the appropriate motor commands are selected for the
execution of the appropriate action, a prediction of the sensory consequences of the movement
is formed and would be compared with the feedbacks associated with the actual execution of the
intended movement. This prediction is the signal upon which motor awareness is constructed.
Note that because this signal precedes, instead of follows, sensory feedback, it is the ideal
candidate for being the basis for Libets participants M judgment. Moreover, it is also
consistent with Fourneret and Jeannerods (1998) conclusion. Data on brain-damaged patients
confirm this model and shall be discussed later in the chapter.
Motor Cognition and Body Ownership

As mentioned earlier, a fundamental component of the conscious feeling of moving is the
sensation that we are acting with part of our own body (p.374) (see also Tsakiris, Chapter 10 of
this volume). Indeed, normal functioning of processes underlying intentional actions and the
veridical construction of motor awareness entail a normal sense of body ownership (Tsakiris &
Haggard, 2005). The processes that are involved in the construction of normal body awareness
are, however, far from being clear. It has been suggested that several different components
contribute to our feeling of body ownership. According to some views, conscious, top-down,
cognitive representation that also incorporates semantic knowledge about the body are thought
to interact with unconscious, bottom-up, dynamic representation that relies on proprioceptive
information from the muscles, joints, and skin (Head & Holmes, 1911). This latter representation
is used to control posture and to enable and shape motor actions. The coherent sense of body
ownership that would emerge from these representations can be, however, affected by different
experimental interventions that manipulate the sensory-motor state of the body. For example,
synchronous touches of a visible rubber hand and the participants hand (hidden from sight)
produce the sensation of feeling the touch on the rubber hand, as well as a compelling feeling of
ownership of that hand (the rubber hand illusion; e.g., (Botvinick & Cohen, 1998; Costantini &
Haggard, 2007; Ehrsson et al., 2004; Farn et al., 2000; Longo et al., 2008; Tsakiris & Haggard,
2005). This illusion exploits the brains predilection for integrating congruent tactile, visual, and
proprioceptive stimuli, even when they are in contrast with one another. Recent studies also
Page 3 of 16
Deficits
demonstrated that we may actively incorporate tools in our body representation, so that the tool
changes both our perception of body extension (Sposito et al., 2012) and our mapping of the
external world (Berti & Frassinetti, 2000; Longo & Lourenco, 2006). Interestingly, these studies
showed that tool embodiment was present only when the tool was used for performing a
voluntary action, suggesting that only intentional tool use shapes ones own body schema.
Accordingly, we shall show the flip side of this principle and present evidence that altered
feeling of body ownership directly affects motor cognition. The studies that we shall review in
this chapter aim at clarifying the relation between agency, intention, and body ownership from
the point of view of pathological behavior, observed in brain-damaged patients with dramatic
impairment of motor and body awareness.
Motor Awareness Deficits and Their Relation with Agency and Intention
In this part of the chapter, the relation between motor awareness, motor intention, and agency
will be discussed referring to recent experimental studies (p.375) in brain-damaged patients
affected by anosognosia for hemiplegia (AHP) and motor neglect who, in different ways, deny
their contralesional motor disorders.
An important contribution for the understanding of the neural bases of motor awareness comes
from the study of a pathological condition in which movement awareness is dramatically
impaired. AHP patients suffering from right-brain damage develop a paresis of the left side of
the body but obstinately deny their motor deficit, and when asked to move their paralyzed limb
they claim to have performed the action required by the examiner (see Bottini et al., 2010;
Fotopoulou, 2012; Orfei et al., 2007; Pia et al., 2004; Spinazzola, Bellan, Pia, & Berti, 2014). It
has been proposed that AHP might be explained as a domain-specific disorder of motor control
(Berti & Pia, 2006; Gold et al., 1994; Jenkinson & Fotopoulou, 2010; Spinazzola et al., 2008).
Several imaging studies of intact brains show that the cortical network for conscious awareness
of action overlaps with that for control of movement (e.g., (Desmurget & Sirigu, 2009).
Accordingly, it has been demonstrated that AHP follows from brain damage to the same cortical
network responsible for motor monitoring. This network is located in the lateral premotor and
insular cortex (Berti et al., 2005; Fotopoulou et al., 2010; Garbarini et al., 2013b; Karnath et al.,
2005; Moro et al., 2011; Pia et al., 2013b; Vocat et al., 2010). Consequently, the well-established
framework of a forward model of normal motor control (Blakemore & Frith, 2003; Wolpert et al.,
1995) has been employed to predict the pattern of intact and impaired neurocognitive
mechanisms, pinpointing the distorted motor awareness of AHP patients. As already mentioned,
the model posits that, when a subject has the intention to move and the appropriate motor
commands are selected and sent to the appropriate motor areas, a prediction (forward model) of
the sensory consequences of the movement itself is formed based on the efference copy of the
programmed motor act. This would be subsequently matched (by a comparator system) with the
actual sensory feedback (see also Gold et al., 1994). The efference copy is the signal from which
motor awareness is constructed. This model has two important implications. First, motor
awareness would, counterintuitively, precede movement execution, instead of following it. This
entails that whenever a sensory prediction is formed, motor awareness emerges before the
availability of any sensory feedback. Second, motor awareness is evaluated against the sensory
feedback by the operation of the comparator system that, among other functions, can
differentiate between movement/no-movement conditions. Within this framework, it has been
Page 4 of 16
Deficits
proposed that, in AHP patients, damage to the comparator processes would alter the monitoring
of voluntary actions, thus impairing the ability to distinguish between movement and no-
movement states. The (non-veridical) feeling of movement in AHP would then arise from an
intact (p.376) motor intentionality, assisted by the normal activity of the brain structures that
implement the intention-programming system (Berti & Pia, 2006; Berti et al., 2007; Garbarini et
al., 2012; Pia et al., 2013b; Spinazzola et al., 2008). Several studies clearly demonstrated
preserved movement intentionality in AHP patients; first, AHP patients may show normal
proximal muscle electrical activity in the affected side when they believe they are moving the
plegic limb (Berti et al., 2007; Hildebrandt & Zieger, 1995). Second, their subjective experience
of willed actions is strictly related to their preserved intentionality. For instance, AHP patients
falsely detect the movement of their plegic arm when they intend to move it, but do not detect
movement when they do not have intention (Fotopoulou et al., 2008). Recently, we directly
demonstrated that the motor programs for the affected limbs of AHP patients are identical to
those that govern normal movement execution, implying no deficit in generating motor
intentions (Garbarini et al., 2012; Pia et al., 2013b). We did this by taking advantage of the
difficulties of bimanual motor tasks. When both hands move simultaneously, strong coupling
effects arise and neither hand is able to perform independent actions successfully (see Swinnen,
2002, for a review). This conflict stems largely from internal motor programming rather than
from the on-line feedback coming from movement execution (Drewing et al., 2004; Franz &
Ramachandran, 1998; Spencer et al., 2005; but see Mechsner et al., 2001, for a different point of
view). We expected that, in AHP patients, these bimanual interference effects should be present,
because motor programs and sensory predictions are both present, despite the absence of
movement and sensory feedback for the affected hand. Using bimanual motor tasks, in which
AHP patients were asked to simultaneously perform movements with both hands, we found that
the movements of the intact hand were influenced by the intended movements of the paralyzed
hand, although these movements were, of course, not actually executed. This influence produced
both spatial (Garbarini et al., 2012) and temporal (Pia et al., 2013b) coupling effects,
comparable to those found in healthy subjects actually performing bimanual tasks. Examples of
spatial bimanual coupling, for both healthy subjects and AHP patients, are shown in Figure 17.1
A and B. Recently, it has been showed that the spared intention-programming system can affect
other distal kinematic parameters of the healty hand as grip aperture (Piedimonte, in press).
These findings in AHP patients suggest that their motor awareness may be constructed from a
normal intentional process, even in absence of movement execution.
Page 5 of 16
Deficits
An intriguing contrast to AHP was provided

by patients without hemiplegia who were
affected by the distinct neurological
syndrome of motor neglect (MN). Indeed,
MN can, in some ways, be considered the
neurological opposite of AHP. Sometimes
described as pseudo-hemiplegia (Laplane
& Degos, (p.377) (p.378) 1983), and often
interpreted as the consequence of a damage
to intentional motor circuits (Berti et al.,
2007; Gold et al., 1994), MN is
characterized by underutilization of the
contralesional limb in the presence of
normal strength, reflexes, and sensibility
and thus preserved potential for actual
movement on the affected side. Crucially,
when MN patients are asked to perform
bimanual movements, they only perform ipsilesional hand movements (Laplane & Degos, 1983),
even though (unlike AHP cases) they are actually capable of moving the contralesional hand.
However, unlike AHP patients or normal controls, MN cases did not show any coupling effect
(Garbarini et al., 2012). This pattern was also obtained with HP cases (without anosognosia),
even though MN patients differ in being physically capable of moving the contralesional limb
(see Figure 17.1 C and D). The MN cases provide an interesting contrast to the AHP cases. The
former are non-plegic but apparently lacking intention/planning, whereas the latter are plegic
but still maintain intentions/plans for the affected hand.
Although AHP and MN are different in terms of motor intention and motor planning, it has been
recently demonstrated that they are both characterized by lack of awareness for the motor
impairment (Garbarini et al., 2013a). This finding can be explained within the computational
model of motor control described earlier. More precisely, denial behavior in AHP patients may
be due to direct damage to the comparator system, which cannot detect the mismatch between
the intended but unexecuted action. The evident feeling of movement that AHP patients
(erroneously) experience would then arise from intact structures implementing motor
intentionality. Conversely, a deficit at the intention-programming level can explain MN. Here,
brain damage causes an inability to form motor intentions; if the intention to move is defective,
motor planning is prevented. Because the comparator, although working properly, does not
receive any information about movement planning, it cannot interpret the lack of movements as
aberrant (Gold et al., 1994). Therefore, MN patients are completely unaware of the lack of any
execution, not because of a direct damage to the comparator system, as it has been proposed for
AHP patients, but because they cannot discover the abnormality of their behavior, since they do
not attempt to make any movement.
In summary, the evidence described here, in AHP patients, suggests that conscious intention to
move can be experienced without actual movement. The cerebral areas that seem to be involved
in this process are the mesial-frontal and posterior-parietal areas (spared in AHP patients) for
the intentional component of the motor act, and the premotor and insular cortices (damaged in
Page 6 of 16
Deficits
AHP patents) for the process of construction

of motor awareness. It seems logical that Figure 17.1 A modified version of a Circles-
Lines task (Franz et al., 1991) was used to
the sense of agency, the feeling that I am
investigate the spatial bimanual coupling.
the one who is controlling the movements, Participants were asked to draw circles and
should be related to the presence of an lines, either performing unimanual drawing
ongoing action (p.379) (Haggard, 2005). movementsvertical lines with the right
(intact) handor bimanual movements
However, pathological conditions in which vertical lines with the right hand and
the presence of actions is dissociated from simultaneously, circles with the left
both the sense of agency and conscious (paralyzed) hand. In normal subjects, we
know that the trajectory of the right hand
intention have been described. For instance,
drawing lines tends to assume an oval shape
in the anarchic hand syndrome, patients (i.e., more spatial error) when the other
perform actions coherent in respect to hand must concurrently draw circles,
environmental stimuli that do not compared to unimanual conditions,
indicating that the motor programs for the
correspond to the patients conscious free
hand drawing circles can affect the motor
will (see Rowe & Wolpe, Chapter 18 of this programs of the hand drawing lines, to
volume). These patients have intact motor produce a bimanual interference effect
awareness and motor function, but they feel (Franz et al., 1991). We predicted that, if
patients with anosognosia for hemiplegia
that it is not their self that is in charge of (AHP) do intend and plan movements with
the decision of moving. In contrast, their paralyzed hand, the lines drawn by the
anosognosic patients not only believe that intact hand should become more oval when
the patients are requested to draw a circle
their left limbs can move, but they are
concurrently with their plegic hand, as found
always convinced that they are the ones in normal subjects. Different results were
who move the arm. In other words, they expected in hemiplegic patients fully aware
never doubt their authorship once they have of their motor deficits (hemiplegic patients
without AHP) and in patients without
programmed a movement, despite the
hemiplegia who do not intend/plan
impossibility of performing it. Moreover, contralesional movements in bimanual
they seem to have intact, although situations (motor neglect [MN] patients). In
decontextualized, motor intention: they try Figure 17.1, examples of subjects right-hand
trajectory in the crucial bimanual Circle-Line
to make movements with the paralyzed condition are shown. Note the increased
limbs, both spontaneously and under ovalization for healthy controls (A) and for
request. This picture suggests that, hemiplegic patients with AHP (B), but not for
although the sense of agency can be present hemiplegic patients without AHP (C) or
patients with MN (D).
without any ongoing motor activity, it is
tightly linked to the feeling of intention to
move, even when that feeling is constructed
on pathological beliefs.
Body Awareness Deficits and Their Relation with Agency and Intention
In this section of the chapter, the relation between motor awareness, motor intention, and
agency will be discussed, referring to recent experiments carried out in some patients who,
despite being aware of their contralesional sensory-motor deficits, believe that someone elses
limb is attached to them.
In right-braindamaged patients with motor and somatosensory impairments, a body awareness

deficit can sometimes be observed (a disturbance called somatoparaphrenia). Patients may feel
a sense of strangeness toward contralesional limbs that may be felt as separated from the
Page 7 of 16
Deficits
patients body. The more frequent manifestation of this disorder is characterized by a sense of
disownership, which is the delusional belief that the contralesional limbs do not belong to ones
own body but to another person (Jenkinson, Haggard, Ferreira, Fotopoulou, 2013; Vallar &
Ronchi, 2009). Recently, an anatomical account of somatoparaphrenia as delusion of
disownership has been proposed (Gandola et al., 2012), suggesting a crucial involvement of
white matter and subcortical gray structures (thalamus, basal ganglia, and amygdala). Recently
we observed the case of a somatoparaphrenic patient who attributed his contralesional arm to
his brother. He claimed that when the nurses took the blood from his arm they were actually
taking the blood from his brothers arm and expressed his unhappiness about the nurses
needlessly intervening to take blood from his brothers arm, when his brother was demonstrably
healthy. Once the somatoparaphrenia recovered, the patient not only regained consciousness of
his own (p.380) left arm, but also remembered his delusional behavior. Crucially, he reported
having previously held the false belief that his left arm belonged to his brother (unpublished
case).
The possibility that an opposite behavior exists, that is, patients who misidentify other peoples
limbs as if they were their own, has been rarely considered. However, in recent studies
(Garbarini & Pia, 2013; Garbarini et al., 2013b; Garbarini, Fornia, et al., 2014; Garbarini,
Fossataro, et al., 2015; Pia et al., 2013a), we observed this behavior in a sample of hemiplegic
and/or hemianesthesic patients. While they did not explicitly deny that their contralesional (left)
limbs belonged to themselves (as in the somatoparaphrenic delusion of disownership), they
claimed that the examiners left hand was their own whenever it was positioned, in egocentric
coordinates, on the table next to their real left hand. This delusion of ownership, although
resembling the rubber hand illusion (Botvinick & Cohen, 1998), was spontaneous and not
induced by any experimental procedure. Patients treated and cared for the experimenters left
arm as if it were their own, showing a consistent embodiment of an alien hand in their own body
schema. Because of this pathological embodiment, we named them E+ patients. Interestingly,
this phenomenon occurs not only with a static alien hand but also when movements are present,
that is, when the examiner moved his or her left hand, patients claimed that they were moving
their own (paralyzed) hand.
It is worth noting that, as mentioned earlier, at the time of testing E+ patients did not show any
explicit form of disownership, never spontaneously reporting delusional beliefs about the
contralesional body parts. Accordingly, when only their left hand was present, they correctly
identified it as their own left hand. However, when both the own and the alien hands were
present, not only did they misidentify the alien hand as their own, but they also identified their
own left hand as alien, affirming that it belonged to someone else, thus showing an explicit
sense of disownership. This suggests that the two delusional behaviors (disownership of the own
hand and ownership of an alien hand) may coexist in the same patients, though directed at
different objects (see also Kammers et al., 2011; Moseley et al., 2008, for similar results on the
rubber hand illusion). The fact that sometimes the disownership behavior is immediately evident
while another time it emerges only as a consequence of the alien hand misattribution may
suggest that these two forms of body unawareness lay on a continuum, possibly characterizing
different phases (acute/sub-acute/chronic) of the disease or different intensity of the deficit.
Furthermore, the E+ patients lesion pattern seems consistent with that described in previous
Page 8 of 16
Deficits
studies on neural correlates of the delusion of disownership (Gandola et al., 2012; Garbarini et
al., 2013b).
(p.381) Starting from the clinical description of these cases, different questions can be
addressed: Can we really incorporate into our body schema the body parts of others, altering
our sense of body ownership? In other words, is my body schema fixed and immutable and
does it correspond only to me, or can part of others bodies become part of my sensory-motor
representations? And if the answer to the last question is positive, does this incorporation
influence how we consciously perceive our body? Body image and body perception would
therefore be extended to others body parts, so that the stimulation of embodied alien hand may
elicit subjective sensation specific to our own body. Furthermore, to what extent, given the tight
link between body and motor representations, does an altered sense of body ownership affect
patients intentional attitude and their motor awareness and sense of agency? To answer these
questions, we conducted a series of experiments, some of them still in progress, aiming at
verifying the consequences of the altered body awareness on the patients sensory-motor
parameters.
We investigated the somatosensory domain (Pia et al., 2013a; Garbarini et al., 2014), reasoning
that the pathological embodiment observed in E+ patients is an ideal condition to examine
whether tactile/pain sensations can be transferred to an alien arm subjectively experienced as
own. Patients (with and without the delusion) and healthy controls were tested with a pinprick
protocol to assess pain perception. In the own hand condition, participants placed their arms on
a table and the hand dorsum (either of the right or of the left hand) was stimulated. In the alien
hand condition, the co-experimenters left or right arm was placed alongside the participants
left or right arm, respectively, and the left or right co-experimenters hand dorsum was
stimulated. In both conditions, participants had to rate the perceived sensation on a Likert scale.
We hypothesized that, if the false belief of owning an alien arm is not a mere verbal
confabulation but, rather, the result of a profound embodying mechanism that affects the E+
patients sensory processing, then this should paradoxically produce a feeling of pain not only in
the own condition, when the stimuli are actually applied on the patients own hand, but also in
the alien condition, when stimuli are applied onto the co-experimenters left hand. The results of
a first behavioral experiment (Pia et al., 2013) confirmed these predictions, showing that a body
part of another individual can become, in a pathological condition, so deeply embedded in ones
own somatosensory representation to have a consistent effect on the subjective sensation of
pain. Furthermore, in a second experiment where the skin conductance was recorded during
noxious stimulations (either to the own or the alien hand), we demonstrated that the alien
(embodied) hand can be able to elicit physiological reactions specific to the own hand (Garbarini
et al., 2014).
(p.382) In these patients, we also investigated the motor domain (Garbarini et al., 2013b;
Garbarini et al., 2015) and we asked whether, once an alien hand is embodied into the patients
body schema, its representation can affect motor production and motor control as if it actually
belonged to the patients. E+ patients were asked to execute a modified version of a bimanual
Circles-Lines task (Franz & Ramachandran, 1998; Garbarini et al., 2012; Garbarini et al., 2013;
Garbarini et al., 2015; Piedimonte et al., 2014) in which they had to draw lines with their intact
hand while watching an alien hand performing circles, either in an egocentric position, that is,
Page 9 of 16
Deficits
congruent with the position of the patients left hand, or in an allocentric non-congruent
position, that is, positioned opposite the patient. As in the previous experiment, the crucial
aspect was that in the congruent condition E+ patients misidentified the alien hand as their
own, while in the non-congruent condition patients recognize the alien hand as belonging to the
co-experimenter. We hypothesized that, if the delusion of ownership arises from an abnormal
embodiment process that automatically triggers the intention-programming processes for the
own hand, when the alien hand drew circles in the egocentric condition, the lines drawn by the
E+ patients intact hand should become ovalized (coupling effect), as in normal individuals
actually performing the bimanual task (see Figure 17.1). The E+ patients results verified these
predictions, showing a clear coupling effect in the alien congruent (egocentric) condition, in
which patients claimed having performed circles with their own left hand. It is important to note
that, in the same condition, neither healthy controls nor hemiplegic patients without
embodiment showed any coupling effect. This suggests that simply looking at a hand drawing
circles is not sufficient to induce line ovalization.
In a different experiment (Garbarini et al., 2015), always related to the motor domain, we
investigated in E+ patients the effect of tool-use training on length representation of their
contralesional forearm. We know from the literature that an active tool-use can reshape ones
own body schema, extending peripersonal space and modulating the representation of related
body parts (e.g., Sposito et al., 2012). In our task, an alien hand performed the tool-use training,
acting either in a body-congruent position (aligned with the patients shoulder; where the
pathological embodiment occurs) or in a no-congruent position (misaligned with the patients
where the pathological embodiment does not occur). Coherently, only in the body-congruent
condition, when patients were convinced to perform the tool-use training with their own
paralyzed arm, a significant overestimation effect was found.
These findings clearly showed that a profoundly altered sense of body ownership affects both
motor awareness (E+ patients, usually aware of their motor impairment, were convinced that
their left hand was moving) and sense of agency (E+ patients ascribed the alien movements to
themselves), by directly (p.383) modulating either the action execution (E+ patients showed an
interference/coupling effect very similar to those found in healthy subjects actually performing
bimanual Circles-Lines task) or the body length representation (E+ patients showed an
overestimation effect very similar to those found in healthy subjects actually performing the
tool-use training). It is important to emphasize that E+ patients, although similar to anosognosic
patients in relation to the presence of hemiplegia, are, however, behaviorally different. Indeed,
in everyday situations they acknowledge their motor deficit, therefore showing normal motor
awareness and motor monitoring. However, the presence of an alien hand in an egocentric
position crucially affects their sensory-motor processes and motor consciousness. It is still to be
clarified how this can happen. In the rubber hand illusion the discrepancy between visual,
tactile, and proprioceptive input is resolved by the brain with the transient incorporation of the
rubber hand into the (normal) subjects body schema. We may speculate that when the
representation of the contralesional hand is partially impaired or made fragile by the brain
damage, as in E+ patients, the brain solves the incoherence of the mutilated body
representation by automatically incorporating an alien hand (when positioned in the egocentric
space) so to regain consistency and functionality. As a consequence, the alien hand becomes
part of the patients body representation, and starts to have an impact on the patients sensory-
Page 10 of 16
Deficits
motor processes. This, in turn, modulates motor awareness and the sense of agency. Note that
the alien hand embodiment is fully dependent on seeing the alien hand. However, vision is not
fundamental per se, but is so only if the alien hand is in a position that is congruent with the
patients body schema. Only in this condition does the image of the hand trigger a
reconstruction of the patients body representation. The Circles-Lines task, employed in this
experiment, allowed us to describe to what extent this modulation is effective. Our results
showed that body parts belonging to other people can be so fully incorporated into ones own
body schema as to imply sensory-motor consequences on ones own action execution.
Conclusion
In this chapter we have discussed the relation between different components of motor cognition,
capitalizing on recent findings obtained in brain-damaged patents. Pathology, integrated with
data from healthy subjects, is fundamental for breaking with the accepted depiction of some
mental processes. In particular, the traditional views on agency, intention, and motor awareness
are challenged by the patients behavior described in the previous paragraphs. First, it has been
shown that under particular conditions, as occurs in anosognosic patients, the presence of
normal motor intentionality is sufficient to construct (p.384) motor awareness for a given
action in absence of any real movement execution. This confirms that full consciousness of a
volitional motor act implies the activation of a cascade of motor events that construct motor
awareness based on a combination of intention and prediction. Previous studies in healthy
subjects had already raised this possibility. However, our results offer strong further support
that we are aware of what we intend to do instead of what we really do (see also Fourneret &
Jeannerod, 1998) Specifically, we found counterintuitive but objective consequences for the
motor behavior of the unaffected hand due to programmed but not executed movement of the
paralyzed hand (in AHP patients). In the AHP patients, the sense of agency seems to be
normal, or at least conforms with the new body condition, insofar as the movement that they
believe was executed is recognized as their own, that is, related to their own will. It is difficult to
identify the precise neural signal denoting the sense of agency. However, it seems to be
constructed from a normal integration of both intentionality and motor awareness, despite
absence of actual movement: I have the intention to move, I feel that I move, therefore I am the
one who is controlling the movements.
Even more intriguing is the relation between voluntary action, body sensation, and body
ownership, which is at the core of humans sense of self. Here we described a pathological
condition in which brain-damaged patients, under particular constraints, automatically
embodied other peoples arms. Whatever anomaly leads to such a profound alteration of the
sense of body ownership may be the crucial aspect of this false belief and its objective effects on
the sensory-motor awareness of the patients. In absence of any kind of voluntary action,
humans body ownership is built up through the interaction between afferences (e.g.,
proprioception and vision) and pre-existing representations of the body. This integration, in
health, allows people to distinguish their own bodies from those of others (and from the
surrounding world). When a voluntary action is performed, efferent information is added to this
process. Hence, the match between efferences, afferences, and pre-existing body
representations allows discrimination between our own actions and the actions of other people.
Both afferences integration and efferent processes are affected in E+ patients, who, once they
have embodied the alien arm, feel sensation on it (in absence of any real stimulation or
Page 11 of 16
Deficits
afference) and feel that they move it (in absence of any real movement of their own limb, again
in absence of any afferent feedback). Viewing the embodied alien arm moving seems to trigger
the same motor cascade that leads from conscious intention to motor awareness, in healthy
participants and in AHP patients. The system seems to respond to the moving alien arm as if the
movement were initiated by the patients themselves. Once the cascade is triggered, the sense of
agency is affected as well because E+ patients ascribe the observed movement to their own will.
Again, (p.385) the sense of agency seems to be the result of a negotiation between intention
and motor awareness: I have the intention to move, I feel that I move, therefore I am the one
who is controlling the movements.
References
Berti, A., & Frassinetti F. (2000). When far becomes near: remapping of space by tool use.
Journal of Cognitive Neuroscience, 12(3), 415420.
Berti, A., & Pia, L. (2006). Understanding motor awareness through normal and pathological
behavior. Current Directions in Psychological Science, 15(5), 245250.
Berti, A., et al. (2007). Motor awareness and motor intention in anosognosia for hemiplegia. In
P. Haggard, Y. Rossetti, & M. Kawato (Eds.), Sensorimotor foundations of higher cognition (pp.
1738). Oxford: Oxford University Press.
Berti, A., et al. (2005). Shared cortical anatomy for motor awareness and motor control. Science,
309(5733), 488491.
Blakemore, S. J., & Frith, C. (2003). Self-awareness and action. Current Opinion in
Neurobiology, 13(2), 219224.
Bottini, G., et al. (2010). Anosognosia for hemiplegia and models of motor control: insights from
lesional data. In G. P. Prigatano (Ed.), The Study of Anosognosia (pp. 363379). Oxford: Oxford
University Press.
Botvinick, M., & Cohen, J. (1998). Rubber hands feel touch that eyes see. Nature, 391(6669),
756.
Costantini, M., & Haggard, P. (2007). The rubber hand illusion: sensitivity and reference frame
for body ownership. Consciousness and Cognition, 16(2), 229240.
Desmurget, M., & Sirigu, A. (2009). A parietal-premotor network for movement intention and
motor awareness. Trends in Cognitive Sciences, 13(10), 411419.
Drewing, K., et al. (2004). Timing of bimanual movements and deafferentation: implications for
the role of sensory movement effects. Experimental Brain Research, 158(1), 5057.
Ehrsson, H. H., Spence, C., & Passingham, R. E. (2004). Thats my hand! Activity in premotor
cortex reflects feeling of ownership of a limb. Science, 305(5685), 875877.
Page 12 of 16
Deficits
Farn, A., et al. (2000). Left tactile extinction following visual stimulation of a rubber hand.
Brain, 123(11), 23502360.
Fotopoulou, A. (2012). Illusions and delusions in anosognosia for hemiplegia: from motor
predictions to prior beliefs. Brain, 135, 13441346.
Fotopoulou, A., Tsakiris, M., Haggard, P., Vagopoulou, A., Rudd, A., & Kopelman, M. (2008). The
role of motor intention in motor awareness: An experimental study on anosognosia for
hemiplegia. Brain, 131, 34323442.
Fotopoulou, A., et al. (2010). Implicit awareness in anosognosia for hemiplegia: unconscious
interference without conscious re-representation. Brain, 133(12), 35643577.
normal subjects. Neuropsychologia, 36(11), 11331140.
Franz, E. A., & Ramachandran, V. S. (1998). Bimanual coupling in amputees with phantom
limbs. Nature Neuroscience, 1(6), 443444.
Franz, E. A., Zelaznik, H. N., & McCabe, G. (1991). Spatial topological constraints in a bimanual
task. Acta Psychologica (Amsterdam), 77(2), 137151.
Gallese, V., & Sinigaglia, C. (2010). The bodily self as power for action. Neuropsychologia, 48(3),
746755.
Gandola, M., et al. (2012). An anatomical account of somatoparaphrenia. Cortex, 48(9), 1165
1178.
Garbarini, F., & Pia, L. (2013). Bimanual coupling paradigm as an effective tool to investigate
productive behaviors in motor and body awareness impairments. Frontiers in Human
Garbarini, F., et al. (2013a). Dissociations and similarities in motor intention and motor
awareness: the case of anosognosia for hemiplegia and motor neglect. Journal of Neurology,
Neurosurgery and Psychiatry, 84(4), 416419.
Garbarini, F., et al. (2013b). Embodiment of an alien hand interferes with intact-hand
movements. Current Biology, 23(2), R57R58.
Garbarini, F., et al. (2012). Moving a paralysed hand: bimanual coupling effect in patients with
anosognosia for hemiplegia. Brain, 135(5), 14861497.
Garbarini, F., Fornia, L., Fossataro, C., Pia, L., Gindri, P., & Berti, A. (2014). Embodiment of
others hands elicits arousal responses similar to ones own hands. Current Biology, 24(16),
R7389.
Page 13 of 16
Deficits
Garbarini, F., Dagata, F., Piedimonte, A., Sacco, K., Rabuffetti, M., Tam, F., Cauda, F., Pia, L.,
Geminiani, G., Duca, S., Graham, S., & Berti, A. (2013). Drawing lines while imagining circles:
neural basis of the bimanual coupling effect during motor execution and motor imagery.
NeuroImage, 88, 100122.
Garbarini, F., Rabuffetti, M., Piedimonte, A., Solito, G., & Berti, A. (2015). Bimanual coupling
effects during arm immobilization and passive movements. Human Mov Science, 41, 114126.
Garbarini, F., Fossataro, C., Berti, A., Gindri, P., Romano, D., Pia, L., Della Gatta, F., Maravita,
A., & Neppi-Mdona, M. (2015). When Your Arm Becomes Mine: Pathological Embodiment of
Alien Limbs Using Tools Modulates Own Body Representation. Neuropsychologia, 70, 402413.
Gold, M., et al. (1994). Anosognosia for hemiplegia: an electrophysiologic investigation of the
feed-forward hypothesis. Neurology, 44(10), 18041808.
Haggard, P. (2005). Conscious intention and motor cognition. Trends in Cognitive Sciences, 9(6),
290295.
voluntary movements. Experimental Brain Research, 126(1), 128133.
Haggard, P., & Clark, S. (2003). Intentional action: conscious experience and neural prediction.
Consciousness and Cognition, 12(4), 695707.
Head, H., & Holmes, J. (1911). Sensory disturbances from cerebral lesions. Brain, 34, 102254.
Hildebrandt, H., & Zieger, A. (1995). Unconscious activation of motor responses in a hemiplegic
patient with anosognosia and neglect. European Archives of Psychiatry and Clinical
Neuroscience, 246(1), 5359.
Jenkinson, P. M., & Fotopoulou, A. (2010). Motor awareness in anosognosia for hemiplegia:
experiments at last! Experimental Brain Research, 204(3), 295304.
Jenkinson, P. M., Haggard, P., Ferreira, N. C., & Fotopoulou, A. (2013). Body ownership and
attention in the mirror: Insights from somatoparaphrenia and the rubber hand illusion.
Neuropsychologia, 51, 14531462.
Kammers, M. P., Rose, K., & Haggard, P. (2011). Feeling numb: temperature, but not thermal
pain, modulates feeling of body ownership. Neuropsychologia, 49(5), 13161321.
Karnath, H. O., et al. (2005). Normalized perfusion MRI to identify common areas of dysfunction:
patients with basal ganglia neglect. Brain, 128(Pt 10), 24622469.
Laplane, D., & Degos, J. D. (1983). Motor neglect. Journal of Neurology, Neurosurgery and
Psychiatry, 46(2), 152158.
Libet, B., et al. (1983). Time of conscious intention to act in relation to onset of cerebral activity
(readiness-potential): the unconscious initiation of a freely voluntary act. Brain, 106, 623642.
Page 14 of 16
Deficits
Longo, M. R., & Lourenco, S. F. (2006). On the nature of near space: effects of tool use and the
transition to far space. Neuropsychologia, 44(6), 977981.
Longo, M. R., et al. (2008). What is embodiment? A psychometric approach. Cognition, 107(3),
978998.
Mechsner, F., et al. (2001). Perceptual basis of bimanual coordination. Nature, 414(6859), 69
73.
Moro, V., et al. (2011). Phenomenology and neural correlates of implicit and emergent motor
awareness in patients with anosognosia for hemiplegia. Behavioural Brain Research, 225(1),
259269.
Moseley, G. L., et al. (2008). Psychologically induced cooling of a specific body part caused by
the illusory ownership of an artificial counterpart. Proceedings of the National Academy of
Sciences of the United States of America, 105(35), 1316913173.
Orfei, M. D., et al. (2007). Anosognosia for hemiplegia after stroke is a multifaceted
phenomenon: a systematic review of the literature. Brain, 130(12), 30753090.
Pia, L., et al. (2004). The anatomy of anosognosia for hemiplegia: a meta-analysis. Cortex, 40(2),
367377.
Pia, L., et al. (2013a). Pain and body awareness: evidence from brain-damaged patients with
delusional body ownership. Frontiers in Human Neuroscience, 7, 298.
Pia, L., et al. (2013b). Temporal coupling due to illusory movements in bimanual actions:
Evidence from anosognosia for hemiplegia. Cortex, 49(6), 16941703.
Piedimonte, A., Garbarini, F., Rabuffetti, M., Pia, L., & Berti, A. (2014). Executed and Imagined
Bimanual Movements: A Study Across Different Ages. Developmental Psychology, 50(4), 1073
1080.
Piedimonte, A., Garbarini, F., Rabuffetti, M., Pia, L., Montesano, A., Ferrarin, M., & Berti, A. (in
press). Invisible Grasps: Grip Interference in Anosognosia for Hemiplegia. Neuropsychology.
Spencer, R. M., et al. (2005). Bimanual coordination during rhythmic movements in the absence
of somatosensory feedback. Journal of Neurophysiology, 94(4), 29012910.
Spinazzola, L., et al. (2008). Modular structure of awareness for sensorimotor disorders:
evidence from anosognosia for hemiplegia and anosognosia for hemianaesthesia.
Neuropsychologia, 46(3), 915926.
Spinazzola, L., Bellan, G., Pia, L., & Berti, A. (2014). Unawareness of Motor and Somatosensory
deficits after stroke (UMAS): Validity and Reliability of UMAS Questionnaire. Applied Psychology
Bullettin, 62(270), 3758.
Page 15 of 16
Deficits
Sposito, A., et al. (2012). Extension of perceived arm length following tool-use: clues to plasticity
of body metrics. Neuropsychologia, 50(9), 21872194.
Swinnen, S. P. (2002). Intermanual coordination: from behavioural principles to neural-network

interactions. Nature Reviews Neuroscience, 3(5), 348359.
Tsakiris, M., & Haggard, P. (2005). The rubber hand illusion revisited: visuotactile integration
and self-attribution. Journal of Experimental Psychology Human Perception & Performance,
31(1), 8091.
Vallar, G., & Ronchi, R. (2009). Somatoparaphrenia: a body delusion. A review of the
neuropsychological literature. Experimental Brain Research, 192(3), 533551.
Vocat, R., et al. (2010). Anosognosia for hemiplegia: a clinical-anatomical prospective study.
Brain, 133(12), 35783597.
Page 16 of 16
Disorders of Volition from Neurological Disease

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
Altered Awareness of Action in Neurological Disorders
James B. Rowe
Noham Wolpe
DOI:10.1093/acprof:oso/9780190267278.003.0018

The ability to act voluntarily is fundamentally important to humans, yet it can be impaired by
many neurological conditions. These can lead to several disorders of volition, affecting the
control of voluntary action and the sense of agency. This chapter considers three different
groups of neurological disease: movement disorders, dementia, and focal brain lesions. These
conditions can cause involuntary movements, apraxia, automatic behaviors, alien limb, or
anarchic handall reflecting different abnormalities in the awareness and control of actions.
Historically, these clinical phenomena have enabled a detailed mapping of the functional
anatomy of volitional control. This chapter shows how recent developments in cognitive
neuroscience have been used to study agency in neurological conditions. It also illustrates how
brain imaging has provided new insights into awareness of, and attention to, action in
neurological disease.
Keywords: agency, neurological disease, anarchic hand, alien limb, apraxia, movement disorders
Introduction
Neurological and psychiatric disorders of the brain are extremely common (Fineberg et al.,
2013). Many neurological disorders have been associated with abnormalities of volition,
awareness of action, and agency (Box 18.1). Although common disorders, such as stroke and
Parkinsons disease, can affect the sense of agency, it is paradoxically the rarer disorders with
Page 1 of 26
the extreme phenotypes that have historically been used to characterize the phenomenology and
functional anatomy of volition.
Neurological disorders cause a wide variety of behavioral and perceptual changes that speak
directly to the psychological and neural mechanisms of the sense of agency, the experience of
intention, and volitional control. Interestingly, neurological disorders do not generally abolish a
patients concept of himself or herself as a long-term individual, distinct from others. In other
words, the patient retains a sense of herself as an actor: an agent in the world who feels capable
in principle of initiating some voluntary action, no matter how difficult certain actions are. In
neurological disorders, it is uncommon for patients to report passivity phenomena, existential
depersonalization, self-misidentification, and personal reduplicative paramnesias.
Box 18.1 Neurological Disorders Associated with Disordered Volition or

Agency
Neurodegenerative
Corticobasal degeneration
Frontotemporal dementia (Picks disease)
Parkinsons disease
Alzheimers disease
Progressive supranuclear palsy
Huntingtons disease
Neurodevelopmental
Tourette syndrome
Tic disorders
Epilepsy
Complex partial seizures
Neuropsychiatric
Psychogenic movement disorders
Focal brain lesions
Brain tumors
Trauma and resection
Stroke
Page 2 of 26
Neuroimmunological
Multiple sclerosis
Antibody mediated encephalitis
(p.390) The neurological disorders of agency include hyperkinetic states (e.g., alien limb or
anarchic hand), hypokinetic states (e.g., apathy or paralysis), and integrative deficits that are
specific to certain commands or contexts (e.g., apraxia). Added to this clinical diversity has been
a wide variation in ontology, which has unfortunately been associated with inconsistent
taxonomy, poor operationalization of terms, and an uncertain or inconsistent nosology of
disorders of agency. However, despite clear differences in etiology and neuropathology, these
disorders have convergent clinical features, anatomy, physiology, and pharmacology, which we
discuss in this chapter.
In this chapter, we begin with a description of common syndromes. We then consider recent
insights from cognitive and computational neuroscience approaches before showing how brain-
imaging methods have helped to define the functional anatomy of disorders of agency (see
Figure 18.1). In the final section, we look ahead to the prospect of interdisciplinary and
integrative understanding of disorders of agency, in support of more effective and rational
therapies. (p.391)
Page 3 of 26
Figure 18.1 Summary of methodologies and

key anatomical regions related to
neurological disorders of agency.
Neurological disorders affecting the
supplementary motor area (SMA), pre-SMA,
motor cortex (M1), premotor cortex (PM),
lateral prefrontal cortex (PFC), striatum and
posterior parietal cortex (PPC) can all alter
the sense of agency, and volitional control of
action. Clockwise from top left: functional
brain imaging (Schrag et al., 2013),
delineation of structural lesions, focal
neuropathology, motor control theory,
Bayesian integration and hierarchical active
inference (Edwards et al., 2013), Libet
clocks and structural connectomics (Wolpe
et al., 2014).
Page 4 of 26
(p.392) The Main Disorders of

Agency
Alien and Anarchic Limb Syndromes and
Related Phenomena
The alien limb provides an archetypal
disorder of agency. The clinical reality of
alien limb disorders is very different from
the images portrayed in Hollywood films
such as Mad Love (1935), Dr Strangelove
(1968), or The Hand (1981). Box 18.2
presents three short case descriptions,
illustrating the variety and key features of
alien limbs. The patients can be aware of
their own limb (usually a hand, occasionally Figure 18.1 Summary of methodologies and
key anatomical regions related to
leg or face) behaving in a (semi-)purposeful
neurological disorders of agency.
manner, but the actions themselves are Neurological disorders affecting the
either independent of their stated supplementary motor area (SMA), pre-SMA,
intentions, or even contrary to their motor cortex (M1), premotor cortex (PM),
lateral prefrontal cortex (PFC), striatum and
intentions. The affected limb is in this sense posterior parietal cortex (PPC) can all alter
not under full voluntary control, and the the sense of agency, and volitional control of
actions it makes are associated with a action. Clockwise from top left: functional
diminished sense of agency. brain imaging (Schrag et al., 2013),
delineation of structural lesions, focal
neuropathology, motor control theory,
Bayesian integration and hierarchical active
inference (Edwards et al., 2013), Libet
clocks and structural connectomics (Wolpe
et al., 2014).
Box 18.2 Case History Examples of Alien Limb
Case 1: A 68-year-old woman, two years into the course of corticobasal degeneration,
noticed that her left hand would mirror the right hand unintentionally, and levitate
or reach to her face. It would then reach out to grasp hold of nearby objects, and she
could not willingly let go. It would stroke her leg, furniture, and occasionally other
people nearby. Other clinical features during the course of her illness included
apraxia, dystonia, and aphasia, supporting the diagnosis of corticobasal
degeneration.
Case 2: A 67-year-old woman complained that her right hand would not let go of
objects. Her right hand became functionally useless for activities of daily living, but
it continued to reach out, grasp, and refuse to let go of objects. Sometimes it would
Page 5 of 26
place inedible items such as dog food or coins in her mouth unless she removed
herself from within reach of these items. She remained aware of what her hand was
doing but was unable to stop it. She denied the sensation that another agent was
controlling her hand, even though it does not feel part of me. Other clinical
features during the course of her illness included apraxia, myoclonus, asymmetric
akinetic rigidity, and aphasia, supporting the diagnosis of corticobasal degeneration.
Case 3: A 66-year-old man in residential nursing care for poor mobility was described
by nursing staff as deliberately behaving mischievously. He would grasp door
handles and furniture when staff tried to move him past in a wheelchair, and he
would reach out to touch other peoples buttocks or breasts, and might straighten his
leg to trip passersby. In clinic, he undertook pickpocketing when I stood close by.
He was aware of the ongoing actions, was aware of the verbal instructions to stop
during the behavior, and despite his own distress and embarrassment, he reported
being unable to stop the actions. Postmortem examination confirmed corticobasal
degeneration.
The patient may lack awareness of the action until it is pointed out to him, or he may be aware
from the outset, but in both cases is unable to stop the action. The sense of agency may,
however, be retained by the patient for other actions. (p.393) Typical reactions include not only
frustration, surprise, and denial of ownership of the limb itself (Biran & Chatterjee, 2004), but
also sometimes embarrassment or laughter. It is worth noting that such cases confirm that
visual or proprioceptive awareness of the ongoing action, with acknowledgment that the limb is
part of the patients body (objective ownership), is not sufficient for a subjective sense of
ownership of the limb and its actions, or agency. There is clearly a potential for uncoupling the
awareness of the action from the sense of agency for that action.
Alien limb phenomena can be persistent or transient following focal lesions of the frontal and
parietal lobes, and corpus callosum from stroke or a tumor (Biran & Chatterjee, 2004; Doody &
Jankovic, 1992; Fisher, 2000). Alien limb phenomena are also part of the clinical diagnostic
criteria for the neurodegenerative corticobasal syndrome (Mathew, Bak, & Hodges, 2012;
Armstrong et al., 2013), although they are poorly defined and operationalized in this context.
This syndrome can be caused by corticobasal degeneration, Alzheimers disease, and
frontotemporal dementia, among other aetiologies. The cases in Box 18.2 are examples from our
patients with corticobasal degeneration or a neurodegenerative corticobasal syndrome.
When reviewing alien limb symptoms and signs, one must also consider several related
phenomena: anarchic hand, dystonias, primitive reflexes, and utilization behaviors. These
clinical phenomena may overlap spatiotemporally with an alien limb, or may occur at separate
time points in a given patient, or may be completely dissociated between patients.
Unfortunately, the literature is not consistent in the definitions and distinctions. We next briefly
review these related phenomena.
Page 6 of 26
The anarchic hand syndrome is sometimes considered as synonymous with alien limb syndrome,
but other authors regard it as a distinct entity. We begin by discussing the proposed distinctive
features, before challenging this distinction based on evidence from the corticobasal syndrome.
An anarchic hand may undertake a wide range of complex actions. The pickpocketing in Case 3
or feeding non-foodstuffs in Case 2 (see Box 18.2) are examples. The anarchic hand movements
are unintended, or contrary to intentions, but appear to be to some purpose (cf. Della Sala et al.,
1998). Inter-manual interference from the anarchic hand to the normal hand (Riddoch et al.,
1998) and affordance errors (Chainay & Humphreys, 2003; Humphreys & Forde, 1998) may also
occur. While the anarchic hand has been characterized by actions without intentions, this also
applies to alien limb phenomena.
To distinguish an alien limb from an anarchic hand, some authors have emphasized the need for
an alien quality in the patients reports. For example, in one of the earliest accounts, Golstein
(1908) reported a patient with unilateral repetitive grasping following corpus callosotomy, with
the sensation that (p.394) an alternate agent was responsible: There must be an evil spirit in
the hand (Es muss wohl ein bser Geist in der Hand sein). However, this might reflect a post
hoc attribution of external agency to the actions, especially where the actions were harmful or
antisocial (including choking, self-harm and manual groping).
Brion and Jedynak used the term la main trangre (the foreign hand) to capture the
foreignness or alien nature of alien hand syndrome, in their cases following callosal lesions
(Brion & Jedynak, 1972). Others have tried to draw a distinction between disowning the hand
(an alien limb) and disowning the actions of the hand (an anarchic hand), in which the disowned
actions do not by definition have a sense of agency (Marcel, 2003). A further distinction has
been made on the basis of the apparent concern or indifference of the patient to the behavior:
that patients with an anarchic hand are to some extent distressed by the actions that it makes or
the consequences of those actions.
Fisher retained the overarching term alien limb phenomena for both anarchic and alien hand
syndromes (Fisher, 2000) but reflected the issues raised earlier in his two subgroups: (1)
Complex, unwilled motor acts, including intermanual conflict, mirror movements, interference,
and the pushing aside of the directed limb by the autonomous limb. This offending hand is
usually involuntarily recruited to tasks that the patient intends to perform unimanually with the
other hand or undertakes an incorrect action when desired to act in concert with the other hand.
These actions are sometimes substitutions for, or additions to, intended acts rather than
spontaneous behaviors at rest, and are therefore related to apraxias (see discussion later in the
chapter). (2) Simple, unwilled, quasi-reflex actions. These include autonomous reaching,
grasping, and utilization behavior, automatic limb withdrawal, or levitation. Aspects of
utilization behaviors are also included, but as Biran points out, they may occur in sequence,
producing behaviors that appear to be complex, goal-directed, and autonomous (Biran et al.,
2006).
In a prospective study of 30 patients with neurodegenerative corticobasal syndrome, 15/30 of

patients reported the experience of feeling that they did not control their own hand (Lewis-
Smith & Rowe, subm). Intermanual conflict was reported by 5/30, although subtle mirror
movements were commonly observed by the neurologists. Unintentional and antagonistic
Page 7 of 26
reaching and grasping behaviors were reported by 12/30 and 10/30 patients, respectively.
Interestingly, only 5/30 reported that the hand felt as if it belonged to someone else or was
controlled by someone else: indeed, patients were aware of the logical inconsistency between
their sense that the offending hand was not controlled by them, and the lack of a sense that
someone else controlled it.
These observations suggest that the distinction between an anarchic hand and an alien hand is
not absolute in a cross-sectional population, and may be (p.395) dynamic over the course of
illness. Without prejudicing descriptive accounts of cases, or their value to theoretical
considerations of agency, the evidence from corticobasal syndrome is that a phenomenological
distinction between alien and anarchic hands should not be enforced.
Another example for a possible overlap with the clinical features of alien limb can be caused by
dystonia: sustained or episodic muscle contractions that lead to either static abnormal postures
or dynamic twisting and slowly repetitive movements. For example, as described in Case 1 in
Box 18.2, arm elevation may also arise from dystonia. Unlike an alien limb, dystonia lacks
purposefulness, and may occur without awareness, especially if it is out of vision, not
unbalancing, and not painful. However, disorders of the basal ganglia with alien limb
phenomena are also often associated with dystonia (Riley et al., 1990) and levitation may also
occur with parietal infarcts (Denny-Brown, Meyer, & Horenstein, 1952), making the differential
diagnosis sometimes challenging.
Primitive reflexes, including grasp, are found not only in healthy infants, but in adult patients
after many multifocal injuries or degenerative disorders affecting the frontal lobes. These
movements are simple, rather than complex or sequential, and occur in response to objects
placed in or near the hand or to simply touching the hand. The grasp merely holds the object,
and may turn the object over but goes no further in manipulating or using the object. Unlike an
action, a grasp reflex thus lacks a goal or sense of agency. However, a strong grasp may be
unable to be voluntarily released by the patient, and perhaps because of its inconvenience for
caregivers, it is often misjudged to be willful.
In all three clinical cases in Box 18.2, the affected hand would reach out and grasp objects,
including furniture and people. Depending on the object or body part, the action may appear to
be a utilization behavior (Lhermitte, 1983) or a disinhibited object affordance (McBride et al.,
2013). A critical aspect of these movements is that they are in response to environmental stimuli
within intimate or personal space. With utilization behaviors, the action itself is appropriate to
the object (e.g., putting on a pair of presented spectacles), even if the timing and context are not
(e.g., one is already wearing two pairs). Utilization behaviors may retain a degree of voluntary
control, and a failure to recognize the inappropriate nature of the action in the given context.
Such behaviors might therefore be regarded as impulsive rather than non-agentive (Shallice et
al., 1989) and vary according to attention or distraction.
Moore and Puri (2012) have proposed that the distinction between alien limb and utilization
behaviors lies in the presence of intermanual conflict versus cooperative actions, and that alien
hands tend to be left-lateralized. These observations provide useful clinical guides, but may not
be easy to implement in all cases. In addition, disinhibited, complex, reactive, and quasi-
Page 8 of 26
autonomous (p.396) movements have received many terms, including pseudospontaneous

movements, Nachgreifen (after-grasping) (Schuster, 1923), magnetic apraxia (Denny-
Brown, 1958), manual grasping behavior (Lhermitte, 1983), the groping-grasping
reaction (Magnani et al., 1987), visual groping (Yagiuchi et al., 1987), and tactile mitgehen
(manual pursuit) (Fitzgerald et al., 2007). While the cases may have had differences, there is a
phenomenological core throughout the detailed descriptions of these cases, which we suggest
results from shared underlying mechanisms.
Apraxia
Praxis refers to a set of higher-order motor faculties that enables the performance of purposeful
and complex voluntary actions in response to either internal goals or external stimuli. The
impairment of praxis, called apraxia, or (less commonly) dyspraxia, leads to errors in the
timing, conformation, and execution of movements, in the absence of elementary motor or
sensory deficits. The affected movements are often of the upper limbs, but apraxia also affects
lower limbs and gait, speech, and orobuccal movements. Adequate memory and comprehension
of the task at hand are implicit in the diagnosis of apraxia.
The historical terminology of apraxia is complex. Some terms reflect cognitive models
(ideational and ideomotor apraxias), others the functional precipitants (dressing, gait, and motor
apraxias) and others are contextually or anatomically restricted (constructional, reflexive,
dynamic, graphic, and optic apraxias). Patients are more inclined to use adjectives like clumsy
or useless or not working, while recognizing that sensation and strength are retained.
Perceptual, conceptual, and production deficits can all contribute to apraxia (Stamenova, Roy, &
Black, 2009). The most widely recognized subtypes are ideomotor and ideational apraxia.
Ideomotor apraxia refers to deficits in the translation of learned visuokinaesthetic engrams
into movements. In contrast, ideational apraxia refers to a conceptual deficit, losing the linkage
between objects or tools and their respective actions, such that the visuokinaesthetic engrams
cannot be created (Stamenova, Roy, & Black, 2009; Pramstaller & Marsden, 1996).
Apraxia can arise from damage or degeneration of parietal and frontal cortex, overlapping
extensively with regions associated with alien limb phenomena. To what extent might alien limb
and apraxia be causally related, rather than just correlated phenomena? Later sections of this
chapter will review the evidence for a common disorder of internal models of action and a
failure to integrate sensory and volitional motor signals, in both apraxia and alien limb.
(p.397) Alien limb phenomena have sometimes been attributed to apraxia, with descriptive
(and perhaps confusing) terms used to reinforce the link, such as diagonistic, magnetic, and
repellent dyspraxia (Fisher, 2000; Josephs et al., 2004). In our study of corticobasal syndrome
(Lewis-Smith & Rowe, subm), we examined limb and orobuccal apraxia on a 020-point scale
(based on neurological examination) as well as the range of alien limb phenomena reported by
patients (using a 14-point questionnaire). We included praxis for transitive and intransitive
mime, and copying of unimanual, bimanual, and orobuccal gestures. We found that 28/30
patients had abnormal praxis on one or more tasks, with deficits evident on approximately half
of all tasks. Intriguingly, there was no correlation between the severity of apraxia and the
number of alien limb phenomena reported, whether using all alien limb phenomena or subsets of
Page 9 of 26
motor and apraxic alien limb questions. This suggests that apraxia and alien limb phenomena
can be dissociated, despite common occurrence in a cross-sectional population of corticobasal
syndrome, and despite similar anatomical correlates in focal lesion studies (see discussion later
in this chapter).
Psychogenic Movement Disorders

Psychogenic movement disorders (PMDs) are a major group of movement disorders that are
attributed to a psychological or psychiatric disturbance, in the absence of structural, metabolic,
developmental, or degenerative neurological disease. The majority of cases are characterized by
involuntary movements. Although common (Wiliams, Ford, & Fahn, 1995; Factor, Podskalny, &
Molho, 1995; Stone, Warlow, & Sharpe, 2010), they are difficult to diagnose or treat effectively
(Ellenstein, Kranick, & Hallett, 2011; Schrag et al., 2004; van Rijn, van Hilten, & van Dijk, 2009;
Edwards et al., 2011). The following discussion excludes the minority of cases arising from
malingering or factitious disorders (Ellenstein, Kranick, & Hallett, 2011; Lang & Voon, 2011).
One faces again a complex, inconsistent, and often poorly operationalized set of terms for PMDs.
Some are descriptive (functional disorders, medically unexplained symptoms), reflecting the lack
of a known and well-defined psychological trigger, remaining agnostic on etiology or
pathophysiology of PMDs (Edwards, Lang, & Bhatia, 2012). Other terms presume an etiology
(psychogenic, hysteria, non-organic) or a mechanism (dissociative, conversion disorder). For
simplicity, we stick to PMD, while recognizing that psychogenicity is debatable (Stone, 2011).
The clinical manifestations are also diverse, including tremor (about half of patients with PMDs),
dystonia, myoclonus, gait disturbance, chorea, and paralysis. They may present with features
that are inconsistent with anatomy (e.g., (p.398) sensory loss that is incompatible with either
central somatotopy or peripheral dermotomes) or subject to distraction or entrainment (Kenney
et al., 2007). These clinical features are beginning to be supported by new laboratory tests
(Schwingenschuh et al., 2011) and brain-imaging studies (Schrag et al., 2013), but the diagnosis
remains essentially clinical.
Tic Disorders
There is general agreement about the terminology of tics: they are sudden, stereotyped,
repetitive, but non-rhythmical movements involving discrete muscle groups. They may be
transient simple actions (e.g., blink) or vocalizations (e.g., throat clearing), or more complex
sequences of movements and speech (e.g., palilalia), occurring in isolation or combination.
Multiple tics are the core feature of Tourette syndrome, but developmental and acquired
neurological disorders can both lead to tics. Tics occupy a gray area between voluntary and
involuntary actions: they are temporarily suppressible but ultimately irresistible. Indeed, they
may be experienced as a voluntary response to an escalating premonitory urge (with motor
and sensory components) that eventually overwhelms any intention to suppress the movement.
Emerging Concepts in Voluntary Control: Chronometric Studies in Neurological

Patients
In recent decades, new experimental paradigms from cognitive neuroscience and motor control
theory have been developed to investigate agency or, more generally, the awareness of action.
These quantitative tasks go beyond the use of subjective, introspective measures of ones own
Page 10 of 26
volition - an advantage that has made such an approach particularly appealing for studying
disease states, wherein such insights and self-monitoring can themselves be impaired (de Lange,
Roelofs, & Toni, 2007). In this section, we review the major psychology paradigms and
computational methods that have also been applied to patients with neurological disorders.
The Libet Task in Neurological Disorders

The seminal work of Libet and colleagues (1983) is described at length elsewhere in this book. In
the paradigm, subjects choose when to press a button, while attending a clock face. Using the
clock, they can report the time at which they felt an urge to move (called W judgment) or the
time they pressed a button (M judgment). Typically, the W judgment precedes the time (p.
399) of movement by 200400 ms, and the M judgment precedes muscle activity by ~50 ms.
Lau et al. (2007) suggested that these chronometric measures rely on a weighted cue
integration of different sources of information. The weight of each source depends on its
reliability, which is inversely related to the degree of noise or variability of that source. Lau et
al. (2007) further suggested that the neural information for the awareness of intentions and
actions could come from both sensory feedback of the moving body part and the readiness
potential (the strong negative potential measured by EEG before a voluntary movement
[Kornhuber & Deecke, 1965]). Distinct preparatory processes may be encoded in the readiness
potential, such as the processing of movement schemas or the efference copy of motor
commands, which can be used to predict the sensory effect of ones action (Von Holst, 1955).
Neuronal activity before an action could thus represent both preparation for action and activity
reflecting the predicted (simulated) action outcome. In neurological disorders, changes in these
processes are likely to cause shifts in W and M judgments.
Sirigu et al. (2004) examined patients with lesions of the angular gyrus and cerebellum.
Cerebellar lesions did not affect W or M judgments. Importantly, the largest variability in time
estimates in the study was in the W judgment of cerebellar patients, and it is difficult to exclude
a contribution of the cerebellum to the awareness of action based on this study alone. In
contrast, parietal damage significantly delayed the perception of intention in the W judgment
(Sirigu et al., 1996). This delay could reflect an over-reliance on later action preparatory
processes due to deficits in the early representation of movement schemas, consistent with the
role of the posterior parietal cortex suggested previously.
Delayed W judgment was also found in patients with Tourette syndrome (Moretto et al., 2011),
in proportion to disease severity. The authors interpreted this abnormal sense of volition as a
result of dissociations between motor intentions and sensory feedback. The continuous effort to
inhibit tics and the frequency of involuntary movements could reduce intracortical excitability
prior to movement (Heise et al., 2010) and diminish awareness of preparatory processes. This
interpretation fits well with the cue integration account: as earlier preparatory processes
become unreliable, the experience of intention to act is captured by processes just prior to the
movement, only after the action is fully specified and is about to occur.
The W judgment is also delayed in PMD patients (Edwards et al., 2011). PMD, however, causes a
small positive shift in the M judgment, such that the difference between the two judgments was
not significantly different from zero. This suggests that the normal temporal distinction between
Page 11 of 26
intentions and actions is impaired in PMDs. The blurring of intentions and actions can (p.400)
be reflected in the patients misperception of their psychogenic actions as involuntary, despite
neurophysiological origins similar to voluntary actions (Schrag et al., 2013).
Intentional Binding in Neurological Disorders

In the intentional binding paradigm, subjects use a Libet clock to report either the time of
an action or a sensory event, such as a tone. When the action and tone are coupled, subjects
perceive their action as occurring later in time and the tone as occurring earlier in time. This
temporal attraction - binding - does not occur for involuntary or passive actions (Haggard, 2002;
Engbert, Wohlschlager, & Haggard, 2008). The paradigm overcomes several confounding
factors of individual differences in strategy and bias in time estimates that have undermined
other chronometric studies of volition. Binding has therefore been used as a quantitative
measure in the study of agency in health and disease (Moore & Fletcher, 2012).
In the healthy population, Moore and Haggard (2008) have shown that action binding depends
on a predictive process (modulated by the probability of the tone following the action) as well as
an inferential process. Further, both the predictive and inferential processes depend on the
learned contiguity of the action and its effect (Moore, Wegner, & Haggard, 2009). In accordance
with the cue integration theory, we showed that action binding is modulated by the reliability of
the tone event (Wolpe et al., 2013) under postdictive inferences. Tone binding, on the other
hand, is more likely supported by prediction processes, rather than cue integration (Wolpe et al.,
2013).
Wolpe et al. (2014) recently used the binding paradigm to study 10 patients with alien limb and
apraxia from a corticobasal syndrome (CBS). Tone binding was normal in CBS, suggesting
preservation of sensorimotor prediction for awareness of action in CBS. In contrast, there was a
specific increase in binding of action in the more-affected hand. Binding was normal in the less-
affected hand, providing a crucial internal control condition. Moreover, action binding was
associated with the severity and range of alien limb and apraxia. The substantial increase in
action binding suggests high uncertainty in the perception of time of action, with over-reliance
on the tone for the perception of ones own action. Supporting this interpretation, the precision
of time estimates in baseline conditions correlated with action binding, as predicted by the cue
integration account (Wolpe et al., 2013). We proposed that in CBS the volitional signals that
drive internally generated actions (and suppress actions triggered by the environment) are
imprecise.
Kranick and colleagues (2013) studied patients with psychogenic movement disorders, and
found no difference in action binding between PMD patients (p.401) and controls. Tone
binding, however, was reduced in PMD patients relative to controls. As tone binding relies
mainly on a prediction process (Wolpe et al., 2013), reduced tone binding points to deficits in
sensorimotor prediction, which are developed further in active inference models of PMDs
(Mehta et al., 2013; Voon et al., 2010). Moreover, Moore et al. (2010) reported that Parkinsons
disease did not change either action or tone binding, ON or OFF dopaminergic medication, or
relative to matched controls. However, overall binding, that is, the sum of action and tone
binding, differed between patients ON and OFF medication. The authors suggested that the
increase reflects an increased sense of agency in patients ON dopaminergic medication.
Page 12 of 26
Interestingly, the precision of the action and tone reports were similar between and within
groups (ON vs. OFF), so changes in binding are not likely to result from differential weighting of
action and tone cues in action binding. Dopaminergic medication may instead alter the
prediction processes associating a voluntary action with its sensory effect Moore et al. (2010).
Taken together, the advantages of binding over other tasks make it an invaluable tool for the
study of normal and abnormal volition. Binding studies suggest specific changes in mechanisms
of motor control and awareness of action for several major neurological disorders: impairments
in volitional signals that drive voluntary behavior in CBS; deficits in sensorimotor prediction in
PMD; and an enhancement of action-effect association by dopaminergic medication in PD.
Motor Control and Volition in Neurological Disorders

One of the main assumptions implied in many of the previous studies reviewed above states that
awareness of action arises from distinct processes of motor control. It is proposed that to
optimize motor control, the brain internally represents the dynamics of ones own body and its
interaction with the external world (Wolpert, 1997; Wolpert & Ghahramani, 2000). A comparison
between the current state of the body with the desired state is used by an inverse model to
generate the appropriate motor command for movement. An efference copy of the motor
command is used by a forward model to predict the sensory consequence of ones action, which
is integrated with the actual sensory feedback from the moving body part to generate an optimal
state estimate (Wolpert, Ghahramani, & Jordan, 1995). The estimated state is used to update the
motor command, in order to achieve the desired goal.
Awareness of action has been suggested to arise from the comparison between the predicted
and actual sensory feedback (Frith, Blakemore, & Wolpert, 2000). Deficits in this comparator
have been suggested to underlie abnormalities in the awareness of action (Blakemore, Wolpert,
& Frith, 2002). (p.402) Motor control theory could therefore provide a useful framework to
examine the sense of agency in neurological disorders. Using experimental tools that probe
mechanisms of motor control can thus enhance the investigation of awareness of action.
An important example for this approach is sensorimotor attenuation: the reduction in the
perceived intensity of the consequences of ones own voluntary actions relative to externally
caused sensations (Shergill et al., 2003). Attenuation relies on accurate sensorimotor prediction
(Bays, Wolpert, & Flanagan, 2005) and its integration with sensory feedback. A robust method
to investigate sensorimotor attenuation is a force-matching task. In this task, subjects are
usually asked to reproduce forces applied to their left index finger by a lever attached to a
torque motor (Shergill et al., 2003). Subjects reproduce the force by pressing on the lever with
their right index finger. Typically, the reproduced forces are larger than the forces that are
actually applied. Sensorimotor attenuation has been suggested to support awareness of action,
by facilitating the difference in sensory salience between internally and externally triggered
actions.
Parees and colleagues have used this task to examine sensorimotor attenuation in PMD patients
(Pares et al., 2014). Patients produced greater force when directly matching a stimulus,
compared to healthy controls. That is, they showed reduced sensorimotor attenuation. In other
words, in PMD patients, the perception of the sensory consequences of their own actions is
inappropriately accurate. The perception of ones own action is thereby more similar to that of
Page 13 of 26
external events, which might lead to the perception of movements as involuntary and not self-
caused in PMD patients (Parees et al., 2012), without an appropriate sense of agency.
Internal predictions of outcomes can also be tested using ballistic movements, with manipulated
visual feedback through a mirrored computer screen. Subjects perception of the position of
their action outcome adapts to gradual deviations (Synofzik, Thier, & Lindner, 2006). With focal
or degenerative cerebellar lesions, perceptual abilities were intact in control conditions
(Synofzik, Vosgerau, & Newen, 2008). However, after adaptation to feedback manipulations,
cerebellar patients required visual feedback to maintain the adaptation. The results suggest that
patients with cerebellum lesions have specific impairments in the error-based adaptation of their
internal predictions (Synofzik, Vosgerau, & Newen, 2008). This accords with normative data on
the role of the cerebellum in internal models (Wolpert, Miall, & Kawato, 1998). Therefore,
although awareness of actions seemed to be preserved in cerebellar patients, deficits may
emerge when patients are required to update internal models in the absence of visual feedback.
Thus these patients might be affected when predictions of the results of ones own actions
require adjustments, for example with aging.
(p.403) Predictive Coding and Active Inference: A New Approach to Neurological Disorders of
Agency
Previous sections have emphasized the importance of sensory predictions for motor control and
the sense of agency. The failure to predict, or failure to respond to prediction errors, is a
common feature across several disorders of agency. Predictions need not be restricted to the
outcomes of ones actions. For example, in the behavioral variant of frontotemporal dementia,
characterized by environmental dependency (impulsivity, utilization, grasp), apathy, and loss of
insight, even the prediction of simple sensory stimuli is impaired (Hughes, Ghosh, & Rowe,
2013; Hughes & Rowe, 2013). Where the stimuli are auditory, there is reduced frontotemporal
connectivity in the beta frequency band, consistent with a loss of predictive top-down
influences from frontal to temporal cortex. A widespread reorganization with de-differentiation
of cortical networks occurs, for the prediction and response to unexpected events (Hughes,
Ghosh, & Rowe, 2013; Hughes & Rowe, 2013).
The overarching role of prediction by the brain, to minimize surprise, has been formalized in the
free energy principle (Friston et al., 2010). Surprise in this context amounts to unexpected
sensations (sensory inputs and perceptual inferences) that have not been predicted, including
those that arise from voluntary action. Perception on this account emerges from the adjustment
of predictions by upper levels of a neuronal hierarchy, so as to explain away sensory samples
(or perceptual inferences) from lower levels. Predicted sensations provided by backward
projections from higher to lower levels are compared with the sensory belief in the lower level,
that is, the probabilistic representation of the causes of sensation. A discrepancy constitutes a
prediction error, which is projected forward to the higher level, which can adjust its predictions
so as to minimize the future prediction error. The integration of beliefs and sensory evidences is
precision-weighted (similar to that described earlier for cue integration in intentional binding or
in the perception of time of action in the Libet task). The precision of the prediction error in
each level (a function of post-synaptic gain) is thus important for determining the integration of
prior beliefs and sensory evidence.
Page 14 of 26
These principles have been recently extended to voluntary actions in the active inference
theory (Friston et al., 2012; Friston, 2103). In the motor system, minimizing prediction errors
can be achieved by adjusting the sensory data through movement: movements fulfill the
proprioceptive predictions. In other words, movement is specified in terms of the expected
sensation.
The theory of active inference has been used to explain abnormal awareness of action in patient
populations. For example, PMD has been suggested to result from an abnormally high precision
of the prior beliefs in an (p.404) intermediate level of the cortical hierarchy, for example the
premotor cortex or SMA (Edwards, Lang, & Bhatia, 2012). The effect of this abnormally precise
belief propagates down the hierarchy, even to the level of the spinal cord, where it induces
abnormal movements through the reflex arcs. In parallel, the abnormal precise prediction errors
are propagated up to higher intentional levels in the hierarchy, such as the pre-SMA. As the
relative precision of representations in the higher levels is reduced, prediction errors in the
intermediate levels overwhelm the high-level priors by precision-weighted integration. The
discrepancy between high intentional levels that do not predict movements and the abnormally
precise intermediate levels leading to movements causes the movements to be interpreted as
involuntary, independent of ones own volition (Edwards, Lang, & Bhatia, 2012).
There are fundamental differences between active inference and motor control theory. For
example, in optimal motor control, a voluntary movement is specified according to a desired
state of the system, whereas by active inference, movement is specified according to the
expected sensory effects. In addition, with optimal motor control, the central nervous system
uses internal forward and inverse models to simulate the dynamics of the action in the
environment. In active inference, however, the brain uses generative models that converge on
the best representation of the statistical properties of the sensory input. Under the active
inference models, agency constitutes higher levels belief of the hidden cause of sensory
events.
It is too soon to know whether the active inference model will adequately capture the full range
of disorders of agency in other neurological disorders. Nonetheless, it is an attractive framework
to try to explain diverse phenomena from alien limb and environmental dependences (utilization
behaviors, anarchic hand) to tics and PMDs. Importantly, it offers a clear mechanism to
integrate psychophysical observations with the functional anatomy of volition, from lesion
studies and brain imaging, to which we turn next.
The Anatomy of Volition: Evidence from Focal Neurological Lesions

The first theories of the neural substrates for agency and awareness and control of action were
the result of studies of patients with focal lesions (Geschwind, 1965a). Liepmanns seminal
studies of apraxia following parietal damage suggested that the parietal cortex contains
spatiotemporal schemas for movement. Parietal lesions cause deficits in the generation of these
schemas in ideational apraxia, and a disruption in the flow of these schemas to the frontal motor
areas that execute them in ideomotor apraxia (Liepmann, 1920). Geschwind then proposed a
disconnection syndrome for apraxia, wherein Wernickes (p.405) area for processing motor
commands is disconnected from frontal premotor and motor areas that execute actions. For
performing movement with the non-dominant side, information has to flow via the corpus
Page 15 of 26
callosum to the non-dominant hemisphere: callosal lesions could therefore lead to apraxia in the
non-dominant side (Geschwind, 1965a, 1965b).
The posterior parietal cortex, corpus callosum, and medial frontal cortex are hot spots, where
focal damage can lead to alien limb (reviewed by Scepkowski and Cronin-Golomb, 2003). Based
on the different phenomenology and pathophysiology of the disorder of alien limb, several
authors have proposed systematic structure-function relations. For example, it has been
proposed that medial frontal lesions can cause alien limb in the dominant hand, and can be
expressed in uncontrolled reflexive behavior. In contrast, callosal lesions causing alien limb
result in deficits in inter-manual control (Feinberg et al., 1992). Others have proposed that alien
limb can be distinguished from anarchic limb by the functional anatomy of the lesions, for
example anarchic limb can result from frontal damage and can be expressed in semi-purposeful
movements independent of ones own volition, whereas alien limb can result from posterior
damage with associated loss of the sense of ownership of ones hand (Della Sala et al., 1998). As
suggested previously, we argue that the distinction between these two clinical phenomena is
sometimes challenging, particularly in the context of corticobasal syndrome.
Focal lesion cases have also implicated a similar set of underlying regions in utilization
behaviors. Lhermitte (1983) suggested that frontal areas are responsible for inhibition over the
parietal cortex, so that frontal lesions lead to a release of parietal activity and an over-reliance
on environmental stimuli to trigger actions. In other words, environmental dependency may
arise from a failure to inhibit the externally triggered motor schema (Shallice et al., 1989).
Subcortical areas, such as the thalamus, have also been associated with utilization behavior
(Eslinger et al., 1991). This might be due to subcortical routes for frontoparietal interactions, or
striatothalamic regulation of frontal cortical activity during voluntary behavior (Bhatia et al.,
1994). Moreover, the role of the medial frontal cortex, and SMA in particular, is highlighted by
lesion studies of alien limb (Goldberg, Mayer, & Toglia, 1981) and deficits in the inhibition of
automatic externally triggered action schemas (Sumner et al., 2007). In a case of alien limb
studied by McBride et al. (2013), the reactions to objects that afforded an action by either hand,
and the effects of backward masking on action priming, revealed an exaggerated object
affordance and deficit in the automatic inhibition responses.
Taken together, the studies of patients with focal lesions suggested intriguing accounts for the
neural correlates of voluntary action and the normal sense of agency: the posterior parietal
cortex can be a critical hub for the processing (p.406) of spatiotemporal movement engrams.
These motor plans might be controlled and selected by the frontal cortex according to ones
current goal. The medial frontal cortex, and particularly the pre-SMA, might be a critical relay
center between these areas, which controls the balance between internal, volitionally controlled
motor plans, and external automatic schemas.
Such case-based analyses have been highly influential, but there has been a subtle shift in
emphasis, from phenomenological assessments toward anatomical localization. With the
development of neuroimaging techniques and cognitive neuroscience paradigms (Rowe &
Siebner, 2012), better lesion definition and functional mapping in health are now possible.
Page 16 of 26
Combining Neuroimaging for the Study of Abnormal Agency in Neurological

Disorders: Uncovering the Key Roles of the SMA and Pre-SMA
Although lesion studies have the advantage of demonstrating whether a region is necessary for
agency, they are restricted to regional rather than system-level inferences. In contrast, human
brain imaging reveals widely distributed networks related to agency, even in the presence of
redundancy or parallel systems (see Figure 18.1).
Wolpe et al. (2014), for example, studied patients with alien limb and apraxia resulting from the
neurodegenerative corticobasal syndrome. They showed how the conjoint analysis of three key
experimental components leads to a clear and integrated model of volition, in terms of both
cognitive processes and neural systems supporting a sense of agency. The three components
were (1) a quantitative and objective measure of awareness of action, such as intentional
binding (above); (2) a mechanistic account that draws on motor control theory (above); and (3)
multimodal brain imaging to investigate brain structure and function.
Patients with CBS showed increased binding of the perceived time of actions toward their
effects (intentional binding for action), correlated with the severity of alien limb and apraxia
from structured clinical assessment (Wolpe et al., 2014). Moreover, individual differences
correlated with gray matter volume in pre-supplementary motor area (pre-SMA), and changes in
the integrity of the white matter tracts from medial to lateral prefrontal cortex, and anterior
corpus callosum. They also observed changes in functional connectivity at rest between the pre-
supplementary motor area and prefrontal cortex that were proportional to behavioral measures
of the loss of sense of agency in CBS. The changes in structural and functional connectivity of
the pre-SMA are proposed to lead to alien limb and apraxia from reduced precision of the
internal (p.407) volitional signals for actions, linking the clinical phenomenology to motor
control theory (Wolpe et al., 2012).
Functional brain imaging has been especially helpful to elucidate the mechanisms of PMDs
(Mehta, Rowe, & Schrag, 2013), especially in conjunction with electromyography (Brown &
Thompson, 2001). For example, with PMD-dystonia, EMG confirms pathological co-contractions
during observed rest periods that are not seen with genetically determined dystonia (Mehta et
al., 2013). With PMD-dystonia (Esposito et al., 2009), back averaging of the
electroencephalogram reveals a slow cortical potential beginning ~1 s before the myoclonic
jerk. This contrasts with the brief cortical discharge about 20 ms before the peripheral EMG
bursts. In addition, the slow cortical potential in PMD-myoclonus resembles closely the
Bereitschafts potential that is characteristic of volitional movements in health. This suggests
that the PMD movement is mediated by the same underlying central and peripheral mechanisms
as voluntary actions. What distinguishes patients with PMD, however, is that they have lost the
sense that they control their actions, or have lost either the ability to voluntarily initiate an
action or to inhibit it.
PMD patients also show abnormal action preparation activity in the SMA, an area that is
believed to play a critical role in sensorimotor prediction (Wolpe et al., 2014; Voon et al., 2011).
As SMA activity contributes to the early component of the readiness potential (see earlier
discussion), it would be expected that for perception of ones own intentions and actions, PMD
patients would rely less on such imprecise early signals, and instead excessively rely on later
Page 17 of 26
processes adjacent to the movement. On the Libet task, this would result in positive shifts
observed for W and M judgments (Edwards et al., 2011).
The psychogenic origins of PMD remain controversial and even questionable in many cases
(Stone & Edwards, 2011). However, Voon et al. (2010) used fMRI to study motor responses to
the gender of affective faces. In the amygdala, PMD patients showed heightened responses even
to happy faces, not just fearful faces. More important for understanding the pathogenesis of
PMDs, the patients showed increased connectivity between the amygdala and the SMA,
suggesting a greater influence of limbic regions on motor preparatory cortex during states of
emotional arousal. Agency per se was not manipulated or measured in that study, but in a
follow-up study of voluntary actions comparing internally chosen versus externally specified
actions (Lang & Voon, 2011), patients showed reduced activity of the SMA and reduced
connectivity of the SMA with lateral prefrontal cortex. The importance of this interaction
between the dorsolateral prefrontal cortex and pre-SMA for voluntary actions is underscored by
the changes in this connection in two degenerative disorders associated with changes in
voluntary actions and altered sense of (p.408) agency: Parkinsons disease (Rowe et al., 2001;
Rowe et al., 2010) and corticobasal degeneration (Wolpe et al., 2014).
Changes in the activation of the dorsolateral prefrontal cortex at rest or in motor tasks have
sometimes been considered as a marker psychogenicity, specific to PMDs. However, Schrag et
al. (2013) showed this to be false, using PET to compare regional cerebral blood flow (coupled to
brain activity) in patients with PMD-dystonia and similar dystonia caused by a genetic mutation
(DYT1 gene). Both groups had similar increases in the dorsolateral prefrontal cortex, relative to
controls. However, perfusion patterns elsewhere did clearly distinguish the groups: patients
with PMD-dystonia had increased blood flow in the cerebellum and basal ganglia, with
decreases in the primary motor cortex. The opposite pattern occurred in patients with genetic
dystonia. Interestingly, the blood flow changes occurred at rest and during two active motor
conditions, such that standard fMRI methods would have probably missed the distinctive
features of the PMD. This study, supported by long-term follow-up of the PMD cases and EMG
monitoring, convincingly shows that there is a distinctive and regionally specific neurometabolic
signature to PMD, in addition to transdiagnostic changes in prefrontal cortex that may relate to
increased awareness and attention to action or motor outcomes (Rowe et al., 2002).
The SMA and prefrontal cortex have also been implicated in the generation and control of tics. It
is tempting to consider tic disorders as a failure of voluntary control (Channon et al., 2009),
supported by neurophysiological evidence of increased cortical excitability at rest and reduced
intra-cortical inhibition (Heise et al., 2010), as well as a delayed awareness of the intention to
make a voluntary movement (i.e., to not tic) (Moretto, Katschnig, Bhatia, & Haggard, 2011).
However, there is also evidence for enhanced cognitive and motor control in tic (Jung et al.,
2013) and enhanced activity and structural connectivity of prefrontal cortical areas. Moreover,
the medial frontal cortex, including the SMA, has been linked to the generation of tics during
event-related functional brain imaging (Bohlhalter et al., 2006). Conversely, repetitive
transcranial magnetic stimulation (rTMS) of this area can suppress tics for months, in
association with increased motor cortical excitability thresholds (Le et al., 2013).
Page 18 of 26
In this section, we have seen how the combination of functional imaging, neurophysiology, and
behavioral phenomenology has recurrently highlighted the importance of the SMA and pre-SMA
in generating and being aware of voluntary actions. Activity of these two adjacent regions and
their connectivity with prefrontal cortex and amygdala form the kernel of a neural network for
agency. The imaging data build on focal lesion studies of the SMA as a cause for example of
alien limb, as discussed earlier in this chapter (p.409) (McBride et al., 2013; Goldberg, Mayer,
& Toglia, 1981). However, through imaging of neurological disorders of agency, from CBS and
Parkinsons to PMDs and tics, one can both identify widely distributed brain networks in
disorders (Voon et al., 2010), and refer directly to normative studies of automatic or controlled
actions (Sumner et al., 2007; Rowe et al., 2002).
Future Directions
The development of neuroimaging and computational methods to study volition in the context of
neurological disorders has built upon a long tradition of neuropsychology. Progress has been
made in terms of the cognitive processes related to volition, and the selective impact of disease
on these processes, owing to advances in cognitive neuroscience. Progress has also been made
in understanding the neural correlates of volitional disorders, including localization of functional
abnormalities (e.g., the SMA and pre-SMA) and identification of disordered brain network
connectivity (both disconnection and hyperconnectivity states).
Although we do not yet have an established unifying neurocognitive framework for volition and
the impact of neurological disorders on volition, there is the realistic prospect of such
unification. The concept of active inference, for example, is ambitious and effective in explaining
volitional disorders in terms of hierarchical beliefs, prediction, and inference, integrating both
anatomical and pharmacological evidence.
A challenge for the future is to translate this understanding of neurocognitive systems for
volition into more effective therapies for patients with neurological disorders of volition. Future
therapies might be cognitive, for example using cognitive strategies, attention, or training to re-
appraise the sense of agency, or to enhance the efficacy of underlying neural networks. New
therapies might also include transcranial magnetic stimulation or transcranial direct current
stimulation, to modulate cortical excitability or inhibition. These neurophysiological
interventions can selectively modulate and enhance cognitive and motor functions in other
neurological disorders, such as stroke (OShea et al., 2013), and modulate agency in health
(Desmurget et al., 2009). In addition, pharmacological treatments may alleviate disordered
agency, with growing evidence of a role for dopamine in the sense of agency and the control of
voluntary actions.
The confluence of psychiatry and neurology will continue to be an important factor in

understanding disorders of agency. The complexity and inconsistency of the clinical ontologies
have to some extent been helped by the introduction of cognitive neuroscience methods, in the
clinic and laboratory, providing an (p.410) integrated approach to volitional disorders. The
prospect of evidence-based effective therapies for these puzzling but distressing phenomena is
therefore better than it has ever been.
References
Page 19 of 26
Armstrong, M. J., et al. (2013). Criteria for the diagnosis of corticobasal degeneration.
Neurology, 80(5), 496503.
Bays, P. M., Wolpert, D. M., & Flanagan, J. R. (2005). Perception of the consequences of self-
action is temporally tuned and event driven. Currents in Biology, 15(12), 11251128.
Bhatia, K. P., et al. (1994). Vascular chorea: case report with pathology. Movement Disorders,
9(4), 447450.
Biran, I., & Chatterjee, A. (2004). Alien hand syndrome. Archives of Neurology, 61(2), 292294.
Biran, I., et al. (2006). The alien hand syndrome: what makes the alien hand alien? Cognitive
Neuropsychology, 23(4), 563582.
Blakemore, S. J., Wolpert, D. M. & Frith, C. D. (2002). Abnormalities in the awareness of action.
Bohlhalter, S., et al. (2006). Neural correlates of tic generation in Tourette syndrome: an event-
related functional MRI study. Brain, 129(Pt 8), 20292037.
Brion, S., & Jedynak, C. P. (1972). [Disorders of interhemispheric transfer (callosal

disonnection): 3 cases of tumor of the corpus callosum. The strange hand sign]. Reviews in
Neurology (Paris), 126(4), 257266.
Brown, P., & Thompson, P. D. (2001). Electrophysiological aids to the diagnosis of psychogenic
jerks, spasms, and tremor. Movement Disorders, 16(4), 595599.
Chainay, H., & Humphreys, G. W. (2003). Ideomotor and ideational apraxia in corticobasal
degeneration: a case study. Neurocase, 9(2), 177186.
Channon, S., et al. (2009). Tourettes syndrome (TS): inhibitory performance in adults with
uncomplicated TS. Neuropsychology, 23(3), 359366.
de Lange, F. P., Roelofs, K., & Toni, I. (2007). Increased self-monitoring during imagined
movements in conversion paralysis. Neuropsychologia, 45(9), 20512058.
Della Sala, S., et al. (1998). Dissociation between recency and span: neuropsychological and
experimental evidence. Neuropsychology, 12(4), 533545.
Denny-Brown, D., Meyer, J. S., & Horenstein, S. (1952). The significance of perceptual rivalry
resulting from parietal lesion. Brain, 75(4), 433471.
Denny-Brown, D. (1958). The nature of apraxia. Journal of Nervous and Mental Disease, 126(1),
932.
Desmurget, M., et al. (2009). Movement intention after parietal cortex stimulation in humans.
Science, 324(5928), 811813.
Page 20 of 26
Doody, R. S., & Jankovic, J. (1992). The alien hand and related signs. Journal of Neurology,
Neurosurgery, and Psychiatry, 55(9), 806810.
Edwards, M. J., et al. (2011a). Abnormal sense of intention preceding voluntary movement in
patients with psychogenic tremor. Neuropsychologia, 49(9), 27912793.
Edwards, M. J., et al. (2011b). Limb amputations in fixed dystonia: a form of body integrity
identity disorder? Movement Disorders, 26(8), 14101414.
Edwards, M. J., Lang, A. E., & Bhatia, K. P. (2012). Stereotypies: a critical appraisal and
suggestion of a clinically useful definition. Movement Disorders, 27(2), 179185.
Ellenstein, A., Kranick, S. M., & Hallett, M. (2011). An update on psychogenic movement
disorders. Current Neurology and Neuroscience Reports, 11(4), 396403.
Engbert, K., Wohlschlager, A., & Haggard, P. (2008). Who is causing what? The sense of agency
is relational and efferent-triggered. Cognition, 107(2), 693704.
Eslinger, P. J., et al. (1991). Frontal lobe utilization behavior associated with paramedian
thalamic infarction. Neurology 41(3), 450452.
Esposito, M., et al. (2009). Idiopathic spinal myoclonus: a clinical and neurophysiological
assessment of a movement disorder of uncertain origin. Movement Disorders, 24(16), 2344
2349.
Factor, S. A., Podskalny, G. D., & Molho, E. S. (1995). Psychogenic movement disorders:
frequency, clinical profile, and characteristics. Journal of Neurology, Neurosurgery, &
Psychiatry, 59(4), 406412.
Feinberg, T. E., et al. (1992). Two alien hand syndromes. Neurology, 42(1), 1924.
Fineberg, N. A., et al. (2013). The size, burden and cost of disorders of the brain in the UK.
Journal of Psychopharmacology, 27(9), 761770.
Fisher, C. M. (2000). Alien hand phenomena: a review with the addition of six personal cases.
Canadian Journal of Neurological Sciences, 27(3), 192203.
Fitzgerald, D. B., et al. (2007). Asymmetrical alien hands in corticobasal degeneration.

Movement Disorders, 22(4), 581584.
Friston, K. J., et al. (2010). Action and behavior: a free-energy formulation. Biological
Cybernetics, 102(3), 227260.
Friston, K. J., et al. (2012). Dopamine, affordance and active inference. PLoS Computational
Biology, 8(1), e1002327.
Friston, K. (2013). Active inference and free energy. Behavioral and Brain Sciences, 36(3), 212
213.
Page 21 of 26
control of action. Philosophical Transactions of the Royal Socoety of London B: Biological
Sciences, 355(1404), 17711788.
Geschwind, N. (1965a). Disconnexion syndromes in animals and man. II. Brain, 88(3), 585644.
Geschwind, N. (1965b). Disconnexion syndromes in animals and man. I. Brain, 88(2), 237294.
Goldberg, G., Mayer, N. H., & Toglia, J. U. (1981). Medial frontal cortex infarction and the alien
hand sign. Archives of Neurology, 38(11), 683686.
Goldstein, K. (1908). Zur Lehre von der motorischen Apraxie. Journal fur Psychoogie. und
Neurologie. (Lpz.), XI, 169187, 270283.
Haggard, P. (2002). Voluntary action and conscious awareness. Nature Neuroscience, 5, 382
385.
Heise, K. F., et al. (2010). Altered modulation of intracortical excitability during movement
preparation in Gilles de la Tourette syndrome. Brain, 133(Pt 2), 580590.
Hughes, L. E., Ghosh, B. C., & Rowe, J. B. (2013). Reorganisation of brain networks in
frontotemporal dementia and progressive supranuclear palsy. Neuroimage Clinical, 2, 459468.
Hughes, L. E., & Rowe, J. B. (2013). The impact of neurodegeneration on network connectivity: a
study of change detection in frontotemporal dementia. Journal of Cognitive Neuroscience, 25(5),
802813.
Humphreys, G. W., & Forde, E. M. E. (1998). Disorder action schema and action dysorganization
syndrome. Cognitive Neuropsychology, 15, 771811.
Josephs, K. A., et al. (2004). Frontotemporal lobar degeneration and ubiquitin

immunohistochemistry. Neuropathology and Applied Neurobiology, 30(4), 369373.
Jung, J., et al. (2010). Cognitive control over motor output in Tourette syndrome. Neuroscience
and Biobehavioral Reviews, 37(6), 10161025.
Kenney, C., et al. (2007). Distinguishing psychogenic and essential tremor. Journal of
Neurological Sciences 263(12), 9499.
Kornhuber, H. H., & Deecke, L. (1965). [Changes in the brain potential in voluntary movements
and passive movements in man: readiness potential and reafferent potentials]. Pflugers Archiv
fur die gesamte Physiologie des Menschen und der Tiere, 284, 117.
Lang, A. E., &Voon, V. (2011). Psychogenic movement disorders: past developments, current
status, and future directions. Movement Disorders, 26(6), 11751186.
Page 22 of 26
Le, K., et al. (2013). Transcranial magnetic stimulation at 1 Hertz improves clinical symptoms in
children with Tourette syndrome for at least 6 months. Journal of Clinical Neuroscience, 20(2),
257262.
Lhermitte, F. (1983). Utilisation behaviour and its relation to lesions of the frontal lobes. Brain,
106, 237255.
Libet, B., et al. (1983). Time of conscious intention to act in relation to onset of cerebral activity
(readiness-potential): the unconscious initiation of a freely voluntary act. Brain, 106(Pt 3), 623
642.
Marcel, A. (2003). The sense of agency: awareness and ownership of action. In J. Roessler & N.
Eilam (Eds.), Agency and self-awareness: issues in philosophy and psychology (p. 77). New York:
Oxford University Press.
Magnani, G., et al. (1987). Involuntary grasping and groping responses to space-related visual
stimuli. Movement Disorders, 2(1), 923.
Mathew, R., Bak, T. H., & Hodges, J. R. (2012). Diagnostic criteria for corticobasal syndrome: a
comparative study. Journal of Neurology, Neurosurgery, and Psychiatry, 83(4), 405410.
McBride, J., et al. (2013). Exaggerated object affordance and absent automatic inhibition in alien
hand syndrome. Cortex, 49(8), 20402054.
Mehta, A. R., Rowe, J. B., & Schrag, A. E. (2013). Imaging psychogenic movement disorders.
Current Neurology and Neuroscience Reports, 13(11), 402.
Mehta, A. R., et al. (2013). Coactivation sign in fixed dystonia. Parkinsonism and Related
Disorders, 19(4), 474476.
Moore, J., et al. (2010). Dopaminergic medication boosts actioneffect binding in Parkinsons
disease. Neuropsychologia, 48, 11251132.
Moore, D., & Puri, B. (2012). Textbook of clinical neuropsychiatry and behavioural neuroscience
(3rd ed.). Abingdon, UK: Taylor & Francis.
Moretto, G., et al. (2011). Delayed experience of volition in Gilles de la Tourette syndrome.
Journal of Neurology, Neurosurgery, and Psychiatry, 82(12), 13241327.
Moretto, G., Katschnig, P., Bhatia, K. P., & Haggard, P. (2011). Delayed experience of volition in
Gilles de la Tourette syndrome. Journal of Neurology, Neurosurgery, and Psychiatry, 82(12),
13241327.
Page 23 of 26
OShea, J., et al. (2013). Predicting behavioural response to TDCS in chronic motor stroke.
Neuroimage, 15;85(Pt 3), 924933.
Parees, I., et al. (2012). Believing is perceiving: mismatch between self-report and actigraphy in
psychogenic tremor. Brain, 135(Pt 1), 117123.
Pares, I., Brown, H., Nuruki, A., Adams, R. A., Davare, M., Bhatia, K. P., Friston, K., & Edwards,
M. J. (2014). Loss of sensory attenuation in patients with functional (psychogenic) movement
disorders. Brain, 137(Pt 11), 29162921.
Pramstaller, P. P., & Marsden, C. D. (1996). The basal ganglia and apraxia. Brain, 119(Pt 1),
319340.
Riddoch, M. J., et al. (1998). Visual affordances direct action: neuropsychological evidence from
manual interference. Cognitive Neuropsychology, 15(68), 645683.
Riley, D. E., et al. (1990). Cortical-basal ganglionic degeneration. Neurology, 40(8), 12031212.
Rowe, J. B., et al. (2001). Attention to action in Parkinsons disease: impaired effective
connectivity among frontal cortical regions. Brain, 125(Pt 2), 276278.
Rowe, J., et al. (2002). Attention to action: specific modulation of corticocortical interactions in
humans. Neuroimage, 17(2), 988998.
Rowe, J. B., et al. (2010). Dynamic causal modelling of effective connectivity from fMRI: are
results reproducible and sensitive to Parkinsons disease and its treatment? Neuroimage, 52(3),
101526.
Rowe, J. B., & Siebner, H. R. (2012). The motor system and its disorders. Neuroimage, 61(2),
464477.
Schrag, A. E., et al. (2004). The syndrome of fixed dystonia: an evaluation of 103 patients. Brain,
127(Pt 10), 23602372.
Schrag, A. E., et al. (2013). The functional neuroimaging correlates of psychogenic versus
organic dystonia. Brain, 136(3), 770781.
Schuster, P., (1923). Zwangsgreifen und Nachgreifen, zwei posthemiplegische

Bewegungstorungen. Zeitschrift fur die gesamte. Neurologie und Psychiatrie 81, 586609.
Scepkowski, L. A., & Cronin-Golomb. A. (2003). The alien hand: cases, categorizations, and
anatomical correlates. Behavioral and Cognitive Neuroscience Reviews, 2(4), 261277.
Schwingenschuh, P., et al. (2011). Moving toward laboratory-supported criteria for

psychogenic tremor. Movement Disorders, 26(14), 25092515.
Shallice, T., et al. (1989). The origins of utilization behaviour. Brain, 112(Pt 6), 15871598.
Page 24 of 26
Shergill, S. S., et al. (2003). Two eyes for an eye: the neuroscience of force escalation. Science,
301(5630), 187.
Sirigu, A., et al. (1996). The mental representation of hand movements after parietal cortex
damage. Science, 273(5281), 15641568.
Sirigu, A., et al. (2004). Altered awareness of voluntary action after damage to the parietal
cortex. Nature Neuroscience, 7(1), 8084.
Stamenova, V., Roy, E. A., & Black, S. E. (2009). A model-based approach to understanding
apraxia in corticobasal syndrome. Neuropsychology Review, 19(1), 4763.
Stone, J., Warlow, C., & Sharpe, M. (2010). The symptom of functional weakness: a controlled
study of 107 patients. Brain, 133(Pt 5), 15371551.
Stone, J. (2011). How psychogenic are psychogenic movement disorders? Movement

Disorders, 26, 17871788.
Sumner, P., et al. (2007). Human medial frontal cortex mediates unconscious inhibition of
voluntary action. Neuron, 54(5), 697711.
Synofzik, M., Thier, P., & Lindner, A. (2006). Internalizing agency of self-action: perception of
ones own hand movements depends on an adaptable prediction about the sensory action
outcome. Journal of Neurophysiology, 96(3), 15921601.
Synofzik, M., Vosgerau, G., & Newen, A. (2008). I move, therefore I am: a new theoretical
framework to investigate agency and ownership. Consciousness and Cognition, 17(2), 411424.
van Rijn, M. A., van Hilten, J. J., & van Dijk, J. G. (2009). Spatiotemporal integration of sensory
stimuli in complex regional pain syndrome and dystonia. Journal of Neural Transmission, 116(5),
559565.
Von Holst, E. (1955). [New thoughts and experiments with sensomobility]. Acta Neurovegetativa
(Wien), 12(4), 337345.
Voon, V., et al. (2010a). Impulsive choice and response in dopamine agonist-related impulse
control behaviors. Psychopharmacology (Berlin), 207(4), 645659.
Voon, V., et al. (2010b). Emotional stimuli and motor conversion disorder. Brain, 133(Pt 5),
15261536.
Voon, V., et al. (2011). Dopamine agonists and risk: impulse control disorders in Parkinsons
disease. Brain, 134(Pt 5), 14381446.
Williams, D. T., Ford, B., & Fahn, S. (1995). Phenomenology and psychopathology related to
psychogenic movement disorders. Advances in Neurology, 65, 231257.
Page 25 of 26
Wolpe, N., et al. (2012). Structure and function changes in the medial frontal cortex underlie
impairments in voluntary control in corticobasal degeneration. Dementia and Geriatric Cognitive
Disorders, 33, 131131.
Wolpe, N., et al. (2013). Cue integration and the perception of action in intentional binding.
Experimental Brain Research, 229(3), 467474.
Wolpe, N., et al. (2014). The medial frontal-prefrontal network for altered awareness of action in
corticobasal syndrome. Brain, 137(Pt 1), 208220.
Wolpert, D. M. (1997). Computational approaches to motor control. Trends in Cognitive

Sciences, 1(6), 209216.
Wolpert, D. M., Miall, R. C., & Kawato, M. (1998). Internal models in the cerebellum. Trends in
Wolpert, D. M., & Ghahramani, Z. (2000). Computational principles of movement neuroscience.

Nature Neuroscience, 3(Suppl), 12121217.
Yagiuchi, T., et al. (1987). The groping phenomena in a case of Alzheimer type dementia. No To
Shinkei, 39(1), 7176.
Page 26 of 26
(p.415)
Index

The Sense of Agency


Print ISBN-13: 9780190267278
DOI: 10.1093/acprof:oso/9780190267278.001.0001
(p.415) Index
Haggard and Eitam/The Sense of Agency/9780190267278 subject index July 2
Action generation. see also voluntary action

active inference models, 35535859, 36566, 401, 403404
anarchic hand syndrome, 379
anosognosia for hemiplegia (see anosognosia for hemiplegia)
body awareness deficits, 37983
body ownership, 37374, 382
Circles-Lines task, 376, 377f, 383
comparator model, 373, 375, 4012
coupling effects, 37678, 377f, 382
efference copy, 375, 399, 401
focal neurological lesions studies, 4046
free energy principle, 22627, 35354, 403
intentions, 37273, 378, 38384, 4034
Libet task, 39899, 403
M judgment, 373, 398400
motor awareness, 37374, 382
motor cortex stimulation study, 372
motor neglect syndrome, 37678, 377f
neurological disorders, 4012
predictive coding, 35254, 361, 36566, 4034
proprioception, 37374, 383
rubber hand illusion, 240, 374, 380, 383
sense of agency, 37172
sensory feedback, 37273, 375, 399, 402
tool embodiment, 374
Page 1 of 20
(p.415)
Index
tool-use study, 382

W judgment, 37273, 398400
Action Identification Theory, 15, 2045
Active inference models, 35535859, 36566, 401, 403404
Affordances
anarchic hand syndrome, 355, 36364
if-then planning, 78
intentions, 98, 1078, 108f
sense of agency, 355, 36364
Agency
authorship judgments, xiiixiv
capacity for in social institutions, xixii
causes, consequences of, xiixiv
cognitive basis of, 910, xii
comparator model (see comparator model)
deviant views of, xii
motivation in, xivxv
self-efficacy in, xv
Agentive experiences
action identification, 15
attention, 68, 1516, 21n4
cognitive reconstruction model, 811
content shifts, 1118, 13f
cue integration, 1115, 13f, 1921
elements of, 45, 89, 2021
evolution of, 34
factors modulating, 1416, 2021
(p.416) flow-chart modeling, 1213, 13f
goals, actions in, 56
granularity, 12, 20
hierarchy, 1214, 13f, 1719
intentional binding, 1011, 20, 50n5, 23839, 271, 272, 298, 349, 400401
motor prediction model, 811, 13, 199200
passivity, 6
priming studies, 910, 21n5
sense of agency in, 5
sensorimotor representations, 1617, 21
short-term memories, 1315
skill level factors, 1516
sources, strength variations, 1820
sources of, 811
tense-shifts, 12
Page 2 of 20
(p.415)
Index
willed action, 351

Alien limb syndrome, 38182, 39296, 4057
Ambient echolalia, 26
Amygdala, 154, 156, 158f, 379
Anarchic hand syndrome
affordances, 355, 36364
body ownership, 246, 247n1, 379
characterization, 39296
consciousness retention, 26
etiology, epidemiology, 36364, 405
fMRI studies, 1034, 363
information processing in, 38
intentions, 365
neurocognitive mechanisms, 36465
priming studies, 364
sense of agency in, 347, 36265, 379, xii
spontaneous neural activity, 109
supplementary motor cortex, 36465
voluntary action, 36465, 379
Anosognosia for hemiplegia
action generation, 37479, 377f, 38385
agentive experiences, 9
body ownership, 236, 24446, 38485
sense of agency, 347, 35962, 38384
Anterior cingulate cortex, 102, 152, 154
Anticipatory agency, xiii
Anti-saccade task, 34
Apparent mental causation model, 9, 31415, 315f, 318
Apraxias, 392, 394, 39697, 400, 4047
Associative learning models, 310, 316
Associative-propositional evaluation (APE) model, 12930
Automaticity
conditional, 7374
effortless control, 4243
intentions, 12021
strategic, 6671, 8586
voluntary action, 2526
Automatic word reading, 33, 34
Awareness, 12830, 18589. see also conciousness
Bayess Theorem, 35152
Behavior. see also action generation; voluntary action
automatic vs. controlled, 13132
Page 3 of 20
(p.415)
Index
internal/external cues affecting, 12628

planned behavior theory, 11718, 136
Behavioral intention. see intention
Body awareness deficits, 37983
Body ownership
action generation, 37374, 382
additive model, 23643
anarchic hand syndrome, 246, 247n1, 379 (see also anarchic hand syndrome)
BA6, BA44, 24446
brain activations, 24246
concepts, definitions, 23536
consciousness, 240, 243, 24647
control, 236, 239, 242
fMRI studies, 241, 243, 246
independence model, 237, 24045
insula, 240, 246
intentions, 238
neural mechanisms, 24047
outcome predictions, 242
posterior cingulate, 243
(p.417) precuneus, 243
prefrontal cortex, 241, 405, 4078
premotor cortex, 240, 243
pre-SMA, 24344, 246
proprioception, 241
questionnaire data, 243
rubber hand illusion, 240, 374, 380, 383
schizophrenia, 236 (see also schizophrenia)
sensory attenuation effects, 50n5, 239
somatoparaphrenia, 236, 246, 37983
spatial body-representation effects, 23940
superior frontal gyrus, 243
temporal attraction effects, 1011, 20, 50n5, 23839, 271, 272, 298, 349, 400401
temporoparietal cortex, 241, 246
voluntary action, 23745 (see also controlled action; volition)
Canonical neurons, 1069, 108f
Checking behavior, 209, 229
Choice. see also volition
as control, 145, 15155, 153f
free-choice paradigm, 14849
influences on value of, 15861
no-choice condition, 157
Page 4 of 20
(p.415)
Index
tyranny of, 15960

Civilization and Its Discontents (Freud), 333
Cognitive reconstruction model, 811
Comparator model. see also efference copy; forward model
action generation, 373, 375, 4012
adjustments requirement, 29293
affective cues in, 29091
agency feelings vs. judgment, 29394
efference copy-based internal predictions, 29192
helping hands experiment, 29192
inferences, 29394
intentions, 116
limitations, 29094
misguided agency beliefs, 293
motor simulation, 292
outcome predictions, 22426, 225f, 29192
overview, 290, 291f, xiii
sense of agency, 31314, 314f, 31617, 35256
voluntary action, 40, 51n7 (see also voluntary action)
Computational brain theory, xv
Conflict Monitoring Theory, 18081, 190n1
Conflicts. see goal conflicts
Consciousness. see also volition; voluntary action
action options, 3638, 31921
action-related urges, 3335
affinity-based systems, 4849
attention and, 68, 21n4
automatisms, 2528
concepts, definitions, 50n1
control-truth relationship, 33435
neuroimaging data, 50n4, 243
self-consciousness, 240, 243, 24647 (see also body ownership)
sense of agency, 4041, 25152, 261, 27576, 276f, 34750
subjectivity, 4648
suppression of action, 3840
Control. see also response selection; volition
amygdala, 154, 156, 158f, 379
anatomical areas affecting, diagram, 391f
anterior cingulate cortex, 102, 152, 154
behavioral, 147, 15158, 153f, 158f
Page 5 of 20
(p.415)
Index
blood-oxygen-level-dependent responses, 152, 154, 157

body ownership, 236, 239, 242
capacity to choose, 14748, 15152, 164
caudate, 150, 158f, 159
choice, influences on value of, 15861
choice as, 145, 15155, 153f
(p.418) competence/effectance, 14647
cortical-striatal network, 14955, 153f
cue effects, 15254, 153f
decision-making strategies, 16061
dopaminergic modulation, 149, 27275, 401
effortless, 4243
emotions, 15558, 158f, 162
enhancement as therapy, 163
feedback, 26677, 26768f, 27172
fMRI studies, 150, 153f, 155, 162
free-choice paradigm, 14849
health benefits in understanding of, 16164
illusion of, 148, 160, 162
inferences, 200
insula, 156, 159
lack of control, 124
learned helplessness, 16364
Libet task, 39899
medial prefrontal cortex, 149, 151, 158, 158f, 159, 164
motivation from ( see response selection)
negative context effects, 15558, 158f
neural mechanisms, 14546, 14951, 158f
no-choice condition, 157
nucleus accumbens, 150, 158f
orbitofrontal cortex, 149, 151
outcome predictions, 16364, 21819
perception of control theories, 14649
preference changes, 14849
prefrontal cortical, 149, 15657, 158f
putamen, 150, 158f
reinforcement learning theory, 147
rewards system, 14954, 153f, 158f
rodent studies, 16364
self-efficacy, 147, 162
self-serving bias, 159
sensorimotor attenuation, 157
Page 6 of 20
(p.415)
Index
social relationships, 160

striatum, 14958, 153f, 158f, 161, 162 (see also striatum)
thalamus, 100, 361, 363, 379, 405
tyranny of choice, 15960
uncertainty, 15255
willingness to pay, 15455
Control effectiveness
assessment, 33940
children studies, 33334
commitment, 337
concepts, definitions, 328, 33132
conflicts, 33233
constructive alternativism, 336
contra-freeloading, 331
distancing, 33536
drive theories, 329
feedback, 33738
goal theories, 329
going in right direction, 33840
hedonic principle, 329, 330
locomotion, 33940
mental transformation, 334
motivation, 329, 330, 33233
reality principle, 333
reduction, repression, 33435
rumination, 335
selection, 33637
self-control mechanisms, 33236
self-regulation, control model of, 33740
sense of agency in, 32728, 34041
simulated reality, 330, 332
strength, 331
third-person perspective, 33536
truth effectiveness, 32832
value effectiveness, 32829
verbalization, 335
willpower, 33334
Controlled action, 7879, 4089
Cortical-striatal network, 14955, 153f
Corticobasal syndrome, 39297, 400, 4068. see also anarchic hand syndrome
Cue effects, integration
agentive experiences, 1115, 13f, 1921
Page 7 of 20
(p.415)
Index
control, 15254, 153f

inferences, 208
(p.419) optimal, 296302
response selection, 26869, 268f
theory, 400
Delusions of influence. see schizophrenia
Delusions of reference, 300
Dopaminergic modulation, 149, 27275, 401
Dystonia, 392, 395, 407, 408
ECOG (electrocorticography), 97, 99
EEG (electroencephalography), 97
Efference copy, 375, 399, 401
EMG (electromyogram), 98
Emotions, control of, 15558, 158f, 162
ERD/ERS (event-related desychronization/synchronization), 97, 101
Feedback
affect control, 33738
outcome, control vs., 26669, 26768f, 33738
response selection, 26677, 26768f, 276f
sensory in action generation, 37273, 375, 399, 402
social control, 27780
Feed-forward hypotheses, 360. see also comparator model; efference copy
Flanker task, 34, 18283
FMRI, 97
Forward models, 81, 360
Free energy principle, 22627, 35354, 403
Goal conflicts
awareness, 18589
behavioral variance, 18788
behavior facilitation, 18485
cognitive control activation, 18085
concepts, definitions, 17879
Conflict Monitoring Theory, 18081, 190n1
conflict resolution, 18990
decision duration, 18788
detrimental effects of, 17980
Flanker task, 34, 18283
information processing, 18284
intentions, conflicting, 12324
mindset, 18384
monitoring, adaptation processes, 182
Page 8 of 20
(p.415)
Index
non-conscious processes, 18589, 190n2

personal significance effects, 18283
positive effects of, 18485
separate experiments paradigm, 186
social dilemma paradigm, 188
stimulus response compatibility tasks, 182, 185
Stroop task, 181
in voluntary action, 3032, 3638, 4748
Ideomotor theory of action control, 3032, 31013, 312f, 31617, 319, 404
If-then planning
action consequences mental representation, 8182
action-effect principle, 7778
action execution components, 73, 85
affordances, 78
anticipated behavioral outcomes, 7778, 8183
Brodmann area 10, 7071
cognitive load, 6667
conditional automaticity, 7374
conscious awareness bypass, 6970
critical situation perceptual simulation, 7475
direct motor priming, 82, 84
dual-task paradigm, 6667
forward simulations, low-level, 81
goal intentions, 6465
goal state, 7374
habitual behavior, 72
higher-order goals, 83
hindsight reconstruction, integration, 83
ideo-motor principle, 7778
(p.420) implementation intentions, 6465
implementation intentions mechanisms, 7179
intended behavioral outcome priming, 82
location congruency effects, 6768
motor simulation, 7576
objective agency, 64
on-line guidance, 78
overview, 6364, 8486
pattern-overlap principle, 72
self-efficacy, 7981
self-regulation, 8081
shooter paradigm, 6869
Page 9 of 20
(p.415)
Index
Simon task, 6768

stereotypes, implicit, 6869
strategic automaticity, 6671, 8586
subjective agency, 64
temptations, 67
unintended behaviors, 82
unwanted behaviors, 67, 83
Implementation intentions, 6465, 7179
Inferences
action identification level, 15, 2045
agency as integration product, 208
agency disruptions, 20910
boundary conditions, 2025
causality perceptions, 2035
checking behavior, 209, 229
comparator model, 29394
control, 200
cue effects, 208
EEG studies, 21112
fMRI studies, 21011
goal vs. primes comparison, 2058
neural substrates, 21012
optimal integration accounts, 208
outcome predictions, 199204, 208, 22627
self-agency implicit route, 200
sense of agency, 2012, 352, 35355, 361
social-psychological models, 201
unconscious, 33
volition, 199
Insula, 156, 159, 240, 246
Intention
in volition, 9596, 115
Intentional binding, 1011, 20, 50n5, 23839, 271, 272, 298, 349, 400401
Intentions
action generation, 37273, 378, 38384, 4034
affordances, 98, 1078, 108f
anarchic hand syndrome, 36465 ( see also anarchic hand syndrome)
anterior cingulate cortex, 102, 152, 154
approach motivations, 13435
Page 10 of 20
(p.415)
Index
associative-propositional evaluation (APE) model, 12930

attentional bias, 12627
attitude-behavior relationships, 12526, 133
automaticity, 12021
avoidance motivations, 13435
awareness, 12830
behavior, automatic vs. controlled, 13132
behavior, internal/external cues affecting, 12628
beliefs, attitudes, 11516, 12834
body ownership, 238
braking mechanism, 1079, 108f (see also veto function)
canonical neurons, 1069, 108f
cognitive depletion, 132
cognitive heuristics, errors, biases, 11920
conflicting, 12324
context incongruity, 12426
deliberative inferential processes, 12930
(p.421) EEG measures, 135
elderly stereotype priming studies, 12728
elements, 95
event-related potentials, 96100, 100f, 1059, 108f
exercise cues study, 127
feeling, electrical stimulation of, 1045
firing rate changes, 1012
forgetting, 122
goals, behaviors activation, 12728
habituation, 12021
hypothetical bias, 12526
implicit cognitions, 116
information-processing continuum, 11819
instability of, 12223
intention-behavior gap, 12126
lack of control, 124
mirror neurons, 1069, 108f
movement, physiological markers, 1034
neural activation model, 1069, 108f
neural measurement tools, 9798
neural underpinnings, 96
outcome predictions, 6667, 95, 102, 105, 11618, 12124, 13136, 21819
parietal regions, 1045
Page 11 of 20
(p.415)
Index
preconscious, 13435
pre-SMA, 102, 1056
racial attitudes studies, 125, 13032, 135
reasoned action models, 11520, 136
self-efficacy, 11718
sense of agency, 31013, 312f, 316, 360, 372
social cognition, 12830, 136
social context, 13334
social desirability, 133
spectral power changes, 101
stereotype studies, 12728, 13335
supplementary motor area, 99106, 100f, 109, 223
task studies, 13435
uncertainty, 133
unconscious processes role, 12122, 12628
visuo-motor neurons, 1069, 108f
I-Spy study, 910, 21n5
LFP (local field potentials), 97
The Matrix, 330, 332
Megalomania, 300
Mental representations, 7475, 79, 8182, 84
Mirror neurons, 1069, 108f
Motor cortex, 391f
Motor neglect syndrome, 37678, 377f
Motor prediction model, 811, 13, 199200. see also comparator model
Multifactorial weighting model, 29496
Myoclonus, 407
Neurodegenerative corticobasal syndrome, 39297, 400, 4068. see also anarchic hand
syndrome
Neurological disorders, 38990, 39091f, 4012. see also specific disorders
Obsessive-compulsive disorder, 22830
Optimal cue integration, 296302. see also multifactorial weighting model
Orbitofrontal cortex, 149, 151
Outcome predictions
action-effect anticipation, 21819
action representations, 1417
apparent mental causation theory, 219
attentional bias, 12627
attitudes, implicit beliefs in, 13134
body ownership, 242
cerebellum, 223
Page 12 of 20
(p.415)
Index
checking behavior, 209, 229

common coding theory, 218
comparator model, 22426, 225f, 29192 (see also comparator model)
control, 16364, 21819
control effectiveness, 331, 332
cue effects, 22526, 228
(p.422) emotion, 23031
forward simulations, 81
habituation, 12021
hierarchy of, 1113, 226
if-then planning, 7475, 79, 8182, 84
inferences, 199204, 208, 22627
intentions, 6667, 95, 102, 105, 11618, 12124, 13136, 21819
memory, short-term, 13, 22829
mental representations, 7475, 79, 8182, 84
mismatch detection, 211
models, 22427, 225f, 23031
neural mechanisms, 102, 105, 109, 15155, 153f, 219, 22223, 222f, 229
obsessive-compulsive disorder, 22830
perceptual intensity, 21920, 221f
predictive coding, 22627
primary motor cortex, 223
principles, xiii
reafferent principle, 224 (see also control)
reasoned action models, 116
schizophrenia, 222, 22425, 22729 (see also schizophrenia)
self-attenuation, 21923, 22122f, 228
sense of agency, 19, 212, 21719, 22627, 31719, 35253
sensorimotor gating, 230
sensory feedback, 402
sensory prediction, control feedback-based, 26972
somatosensory cortex, 22223, 222f, 38182
striatum, 14958, 153f, 158f, 161, 162
supplementary motor area, 99106, 100f, 109, 223
uncertainty, 22630, 362
weighting models, 29596 (see also multifactorial weighting model; optimal cue integration)
Parkinsons disease, 409
Perceived self-agency
action systems, 254, 255
Page 13 of 20
(p.415)
Index
auditory-tactile interaction, 258

birth, perceived self-unity at, 25759
dyadic face-to-face exchanges, 261
feelings, first experiences of, 25254
in infancy, 25960
instrumental learning, 26061
movement differentiation, 25960
neonates, acting/feeling by, 25456
newborns, objective perception in, 25657
object grasping, 256
object tracking, 255
overview, 25152
pleasures, frustrations of, 26061
prerequisites, 252
reflex arcs, 25455
reflex systems, 254
self-specifying information processing, 25960
sucking behaviors, 25556
synesthesia, 25759
Planned behavior theory, 11718, 136
Posteriorparietal cortex, 391f, 4056
Predictions. see outcome predictions
Predictive coding, 35364, 361, 4034
Prefrontal cortex
body ownership, 241, 405, 4078
control, 149, 151, 158, 158f, 159, 164
Prefrontal cortical, 149, 15657, 158f
Premotor cortex, 240, 243, 391f, 404
Pre-SMA
active inference, 4034
anatomy, study methods, 391f
body ownership, 24344, 246
disorders of volition, 4069
imaging studies, 4069
intentions, 102, 1056
Primary motor cortex, 223
Proprioception, 241, 37374, 383
Pseudo-hemiplegia, 37677
(p.423) Psychogenic movement disorders, 397403, 4078
Putamen, 150, 158f
Reasoned action models, 11520, 136
Response selection
action-effect learning, 27172
Page 14 of 20
(p.415)
Index
autism, 27879
caudate nucleus, 274
choosing in feeling of autonomy, 266
control as reward, 27374, 27879
control deprivation, compensation, 27980
control effectiveness, 266
control feedback effects, 26677, 26768f, 276f
control feedback fixed parameters, 27172
as control motivation, 26566, 28081
cue effects, 26869, 268f
cutting behavior, 280
decisions of agency, 270, 27577, 276f, 281
dopamine signaling, 149, 27275, 401
learned helplessness, 27980
motor control system, 2831, 270, 277, 404
outcome effectiveness, 266
outcome feedback, control vs., 26669, 26768f
oxytocin, 27879
reinforcement effects, 26669, 26768f
sensory prediction, control feedback-based, 26972
sensory prediction in, 26971
social control feedback, 27780
stereotypy, 27880
striatum, 27479, 276f
temporal binding effect, 1011, 20, 50n5, 23839, 271, 272, 298, 349, 400401
theory of event coding, 270
Reward system, 14954, 153f, 158f
Rubber hand illusion, 240, 374, 380, 383
Schizophrenia
agency attribution, 301
body ownership, 236
inferences, 20910
outcome predictions, 222, 22425, 22729
weighting models, 299302
Self-control mechanisms, 33236
Self-regulation, control model of, 33740
Sense of agency
action-effect representation specificity, 31617
affordances, 355, 36364
agency judgments, 3089, 32122, 321f
Page 15 of 20
(p.415)
Index
in agentive experiences, 5
in anarchic hand syndrome, 347, 36265, 379, xii (see also anarchic hand syndrome)
anatomical areas affecting, diagram, 391f
apparent mental causation model, 31415, 315f, 318
associative learning models, 310, 316
comparator model, 31314, 314f, 31617, 35256
computational neuroscience model, 35355
concepts of, 3079, 34750
conjoint agency, 309, 32223
consciousness, 4041, 25152, 261, 27576, 276f, 34750
conscious will role, 31921
control, 15861
control effectiveness, 32728, 34041
cue effects, 35052
cultural bias, 309, 32223
executive ignorance, 31112
feed-forward hypotheses, 360
force-matching task, 358
(p.424) ideomotor theory of action control, 3032, 31013, 312f, 31617, 319, 404
if-then planning, 7984
inferences, 2012, 35354, 361, 365
information processing, 30814, 318
intentions, 31013, 312f, 316, 360, 372
inverse model, 35253
lesion-mapping study, 360
mapping effects, 31819
measurement of, 34950, 36566
misattribution, 34849, 357
motor awareness and motivation, 35657
neurocognitive mechanisms, 36065, 394
objective agency, 308
optimal control of action, 35253
outcome predictions, 19, 212, 21719, 22627, 31719, 35253
perceived agency, 319
prediction errors, 35355
religion in, 309
Page 16 of 20
(p.415)
Index

sensory states predictions, 35455
table-turning phenomenon, 356
task-evoked pupillary responses, 319
uncertainty, 317, 35355
volition, 238, 244
in voluntary action, 4044, 31115, 312f, 31415f, 31921
willed action experience, 351
Sensory attenuation, 50n5, 239
Sensory prediction, control feedback-based, 26972
Single unit recordings, 98
Social control feedback, 27780
Social psychology, xv
Somatoparaphrenia, 236, 246, 37983
Somatosensory cortex, 22223, 222f, 38182
Strategic automaticity, 6671, 8586
Striatum
control, 14958, 153f, 158f, 161, 162
response selection, 27479, 276f
Stroop task, 3132, 34, 36, 38, 181
Supplementary motor area (SMA)
active inference, 4034
anarchic hand syndrome, 36465
disorders of volition, 4069
focal neurological lesions studies, 405
intentions, 99106, 100f, 109, 223
outcome predictions, 99106, 100f, 109, 223
Synthetic phenomenology, 47
Temporal attraction effects, 1011, 20, 50n5, 23839, 271, 272, 298, 349, 400401
Temporoparietal cortex, 241, 246
Thalamus, 100, 361, 363, 379, 405
Tic disorders, 398, 408
Tourette syndrome, 103, 399
The Truman Show, 330
Truth effectiveness, 32832
Uncertainty
control, 15255
intentions, 133
prediction errors, 361
Page 17 of 20
(p.415)
Index

Utilization behavior syndrome, 26
Value effectiveness, 32829
Visuo-motor neurons, 1069, 108f
Volition
alien limb syndrome, 38182, 39296, 4057
apraxias, 392, 394, 39697, 400, 4047
as brain function, 46, 49, 244 (see also pre-SMA; supplementary motor area (SMA))
(p.425) disorders of, 389, 39091f (see also schizophrenia)
dystonia, 392, 395, 407, 408
inferences, 199
intention in, 9596, 115
lesion studies, 4046
motor control, 4012
motor intentionality, 38384
psychogenic movement disorders, 397403, 4078
research into, xv
tic disorders, 398, 408
Tourette syndrome, 103, 399
Voluntary action. see also action generation; volition
accumulator models, 45
action conflicts in, 3032, 3638, 4748
action effect representations, 2930
action-related urges, 3335
actions decoupled from consciousness, 2627
activity gestalt account, 51n6
adaptation effects, 38
affinity-based systems, 4849
anarchic hand syndrome, 36465, 379
automatisms, 2526
body ownership, 23745
cascade processing mechanism, 2728, 32
comparator model, 40, 51n7 (see also comparator model)
conscious integration, 3638, 4748, 50n4, 50nn12
continuous flow mechanism, 2728, 32, 4243, 51n6
efference binding, 2728, 3637, 43, 46, 50n5
Page 18 of 20
(p.415)
Index
go signal, 4445
homunculus problem, 4244
ideomotor theory, 3032, 31013, 312f, 31617, 319
inclinations, 39, 50n3
information broadcasting, 48
inhibition, suppression, 31, 3740
intention-outcome mismatches, 4041
memory processing in, 3031
motor control system, 2831, 270, 277, 404
motor equivalence, 29
options generation, 3233
perceptual interference, 27
phenomenology of agency, 41 (see also sense of agency)
postvoluntary attention, 51n6
priming studies, 37
procedural learning, 2829
response interference paradigms, 27
selection of, 2934, 3940, 4344
self-report paradox, 4546
sense of agency in, 4044, 31115, 312f, 31415f, 31921
sensory attenuation, 50n5, 239
Stroop task, 3132, 34, 36, 38
subjectivity, 4648
three-term contingency, 3435
unconscious inferences, 33
unconsciously mediated, 2529, 49
Weighting models. see also multifactorial weighting model; optimal cue integration
affective cues, 3012
agency feelings vs. judgment, 29495
cue effects, 29596
delusions of influence disorders, 299302
(p.426) limitations, 296
multifactorial weighting model, 29496
optimal cue integration, 296302
outcome predictions, 29596
reafferences, 29899
reliability in cue integration, 29798
schizophrenia, 299302
self-world distinction, 29899
Page 19 of 20
(p.415)
Index
Page 20 of 20

(Social Cognition and Social Neuroscience) Patrick Haggard, Baruch Eitam-The Sense of Agency-Oxford University Press (2015)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Social Cognition and Social Neuroscience) Patrick Haggard, Baruch Eitam-The Sense of Agency-Oxford University Press (2015)

Uploaded by

Copyright:

Available Formats

The Sense of Agency

Patrick Haggard and Baruch Eitam

Published to Oxford Scholarship Online:

Patrick Haggard, editor

Part II Determining Authorship

Part III Beyond Authorship

University Press Scholarship Online

The Sense of Agency

Print publication date: 2015

Mahzarin Banaji, John A. Bargh, John Gabrieli, David Hamilton,

Elizabeth A. Phelps, and Yaacov Trope

The New Unconscious

Edited by Ran R. Hassin, James S. Uleman, and John A. Bargh

Oxford Handbook of Human Action

Edited by Eziquiel Morsella, John A. Bargh, and Peter M. Gollwitzer

Social Neuroscience: Toward Understanding the Underpinnings of the Social Mind

Edited by Alexander Todorov, Susan T. Fiske, and Deborah Prentice

Self Control in Society, Mind, and Brain

Edited by Ran R. Hassin, Kevin N. Ochsner, and Yaacov Trope

Oxford Handbook of Social Neuroscience

Edited by Jean Decety and John T. Cacioppo

Attention in a Social World

Beyond Pleasure and Pain

The Sense of Agency

Edited by Patrick Haggard and Baruch Eitam

Oxford University Press is a department of the University of

Oxford is a registered trademark of Oxford University Press

Published in the United States of America by

Oxford University Press 2015

All rights reserved. No part of this publication may be reproduced, stored in

Rights Department, Oxford University Press, at the address above.

You must not circulate this work in any other form

Library of Congress Cataloging-in-Publication Data

University Press Scholarship Online

The Sense of Agency

Print publication date: 2015

Utrecht, the Netherlands

Behavioural and Clinical Neuroscience Institute

University Press Scholarship Online

The Sense of Agency

Print publication date: 2015

Causes and Consequences of Sense of Agency

Agency and Motivation

Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological

Gallup, G. G. (1970). Chimpanzees: self-recognition. Science, 167, 8687.

Morewedge, C. K. (2009). Negativity bias in attribution of external agency. Journal of

University Press Scholarship Online

The Sense of Agency

Print publication date: 2015

The Dynamics of Agentive Experiences

Abstract and Keywords

This chapter is dedicated to the memory of Marc Jeannerod.

I start with a preliminary characterization of agentive experiences in the following section. I

It is difficult, however, to conceive of a sense of agency or a feeling of doing as existing

Could you please stop that?

The Sources of Agentive Experiences

This conceptual distinction between agentive experiences and judgments is echoed

A promising approach is to appeal to a Bayesian integrative framework involving a hierarchy of

Agentive Experience: Content Shifts

In his comments on this flow chart,

It is a controversial issue whether low-level sensorimotor representations can contribute

Sense of Agency: Variations in Sources and Strength

Bayne, T. (2010). Agentive experiences as Pushmi-Pullyu representations. In J. Aguilar, A.

Jeannerod, M. (1997). The cognitive neuroscience of action. Oxford: Blackwell Publishers.