What is attention?

 Attention is required to limit entry to a finite

capacity processing system by selecting only a
subset of all available information.

 Attention itself is a resource of limited capacity

which can be divided between tasks.

 Also, processing capacity and attentional

capacity linked to arousal.
 What is attention for?
 Allport (1993) we must ask “what is attention
for?” and accept that “attention” refers to a set
of cognitive/brain processes.

 Selective attention may limit the subset of

objects to be perceived, memories to be
accessed or actions to be executed.

 Attention is also necessary for awareness and

memory. Involved in most tasks.

 To maintain coherence of behaviour.

 Purposes of Attention

 Cope with inherent capacity limitation of brain

 Facilitate stimulus detection
 Facilitate stimulus perception
 Facilitate thinking
 Facilitate memory James
 Recruit relevant processors
 Prepare for action
 Definition of Attention
 “Attention is the process of psychological selectivity by
which we select, from a vast number of potential stimuli,
only those which are related to present interests and
needs” (Ruch, F. L., 1967).

 “Attention is the focusing of perception leading to

heightened awareness of a limited range of stimuli” (Hilgard
et.al., 1975).

 Attention is the tendency of organisms to respond to only a

portion of available stimuli at a given moment” (Bruno, F.
J., 1980).
 Characteristics of Attention

 Readiness
 Bodily adjustment
 Selective reaction
 Specific purpose
 Exploration
 Cognitive process
 Limited span
 Types of Attention
 Voluntary Attention: deliberately
concentrating mental process

 Involuntary Attention: physical properties

of objects determine involuntary attention

 Habitual Attention: attention becomes

more automatic with the passes of time
 Types of Attention

• Selective Attention
Requires monitoring of several channels
of information to perform a single task.

 Examples:
Pilot
Nuclear power operator
Receptionist
 Focused Attention

• A person must attend to one source of

information and ignore all other
sources.

• Examples:
Conversations in noisy
environment
Distracting audience side
conversations
 Divided Attention

• Limited capacity to process information.

• Required to do more than one task at a

time,
also called “time sharing”.

• Two theories regarding information

processing
Single-Resource Theory
Multiple-Resource Theory
 Sustained Attention

• Vigilance - Concerns the ability of

observers
to maintain attention and remain alert
to stimuli
over prolonged periods of time.
 Selective Attention
(Cherry, 1953)

 subjects are presented with 2 or

more stimuli at the same time
• this is called dual-task
performance

 for selective attention tasks,

subjects are instructed to attend to
one input only
 Dichotic listening task

 2 simultaneous messages
• one to each ear
 each message has different meaning
• like listening to TV & radio at same time
 Subjects must shadow one ear only
• repeat the message from one ear only

 Experimenter then asks about unattended ear!

• Listeners extract only certain info from unattended
ear
Early Research Demonstrating
Limited Capacity
 E.C. Cherry (1953)
 Selective auditory
attention
 “Dichotic listening”
procedure
 Subjects given different
inputs to each ear
 Subject asked to report
input from one ear first
then other ear second
 Subjects unable to
report unattended
channel
Early Research Demonstrating
Selection and Inhibition:
The Cocktail Party Effect
 A relevant
conversation can be
attended to in spite
of a welter of
background
conversation…
 But attention can be
captured by some
meaningful stimulus
event occurs in that
background
 Results
 Listeners are good at detecting physical info
• could tell whether unattended channel was…
 Male vs. Female
 Speech vs. Buzzing
 Listeners are bad at detecting meaning info
• couldn’t identify any word/phrase
• couldn’t tell whether unattended channel was…
 Forward vs. Backward
 English vs. German
 So, attention is
• sensitive to physical properties
• insensitive to meaning
 Early Selection Theory
 Broadbent - A Filter Theory of Attention (1958)
• Subject presented with digits simultaneously to each ear.
They are asked to report all of the numbers.
• Tended to report all the information coming to one ear first
then all the information coming to the other ear.
• If required to report digits in the order of arrival they did
better the more time there is between presentations.
Theory claims:
• Sensory channels have an unlimited capacity
• There is a bottleneck allowing only one piece of
information into working memory at a time
• A selective filter allows info from only one
channel at a time.
• Info from unattended channel is completely blocked
• Time is required to switch between channels
 Broadbent’s Filter Model of selective
attention (early selection)

 sensory information (speech) is received in channels

• channel = ear!
 Info goes first to buffer (=temporary store)
 Then filter selects attention (chooses 1 of the channels)
based on physical characteristics (e.g. voice tone)
 This single channel is now the focus of attention
 Ignored (=unattended) info can be recalled from the buffer
but this decays rapidly!

• We can only attend to one channel at a time

• Switching attention between channels requires effort
 Broadbent’s Filter Model

Other channel no longer meaning extracted

processed after filtering… here on after

I
N
P
U
T
sensory filtering short term
buffer (based on memory
physical
characteristics)
 Treisman’s attenuation theory
(early selection)
 a filter attenuates unattended input rather than
“turning it off”
 so non-attended meaning does pass on, but in a
weaker form.
 because it’s weaker (attenuated) we can ignore
it

 stimuli that are emotionally important

(our name) or biologically relevant
(“sex”) are pre-set and so not attenuated
• allows for breakthrough
 Broadbent’s Filter Model

Other channel no longer meaning extracted

processed after filtering… here on after

I
N
P
U
T
sensory filtering short term
buffer (based on memory
physical
characteristics)
 Deutsch & Deutsch

 ALL stimuli are analysed for meaning but only

SOME are selected for response
• unattended message may be recognised but
not pass into awareness
 allows for subliminal perception;
• processing inputs with no conscious
awareness
 Non-attended inputs are noticed only if their
relevance exceeds the relevance of the
attended channel
 allows for breakthrough
 Treisman’s Attenuation Model

Other channel gets meaning extracted

attenuated … here on after

I
N
P
U
T
sensory Attenuator short term
buffer memory
Broadbent STM
Sensory register Selection Filter

Stim#1

Stim#2

Treisman
Sensory register Attenuator STM

Stim#1

Stim#2

Deutsch & Deutsch

STM
Sensory register

Stim#1

Stim#2
 Early or late selection?

 We understand at least some meaning on the

unattended channel. SO Early selection models
are not satisfactory.
 But we still get more meaningful info from the
attended (vs. unattended) channel. SO Late
selection models are not satisfactory
 Perhaps an attenuated theory… But what is
attenuated meaning processing???

Perhaps we need an alternative approach?

 Johnston & Heinz’s Hybrid Model

 We filter out channels either... EASY

 EARLY (based on physical
characteristics)
OR
 LATE (based on meaning)

HARD
 Johnston & Heinz’s Hybrid Model

 Evidence: Dual Task

• Task 1.
Blah blah blah … Attend to this.

Blah blah blah …

 Watch screen. Say when a light appears.
• Task 2.

“NOW!”
 Johnston & Heinz’s Hybrid Model

 Evidence: Dual
Task
• Task 1.
Blah blah blah … Attend to this.

Blah blah blah …

Watch screen. Say when a light appears.
• Task 2.
“NOW!”
 Johnston & Heinz’s Hybrid Model

 Results
• Task 1.

Blah blah blah … Attend to this.

[different voice] …
•Task 2.

“NOW!” FAST
 Johnston & Heinz’s Hybrid Model

 Results
• Task 1.
Blah blah blah …
Attend to this.
[different meaning] …
• Task 2.

“NOW!” SLOW
 Summary – Selective Attention
Unattended channel: we
don’t know meaning

 Early models
• select attention based on physical characteristics
 Late models
• select attention based on meaning
Unattended channel: we
do know meaning
 Attenuation models
• have ‘attenuated’ meaning on unattended channel

 Hybrid model seems most satisfactory

• Can do both early and late selection - Early
selection is easier
 Determinants of Selective Attention
Objective determinants
- Intensity: other things being equal some
thing large attracts attention better than
some thing small

- Change: change in physical properties

- Repetition: a weak stimulus frequently

repeated may be as effective as a strong
one presented once.
 Determinants of Selective Attention

 Color: color objects attracts more

than black and white
 Contrast: a tall person among
average height people will gain more
attention
 Novelty: unfamiliar objects gain
more attention than familiar objects
 Clarity: clear objects are attended
more than unclear objects.
 Determinants of Selective Attention

Subjective determinants
Characteristics of the individual determines
attention
- Interest

- Habit

- Past experience

- Motivation

- emotion
 Divided Attention
Dual Task Performance
Relevant to processing capacity
Interference methodology a useful tool
to determine whether two tasks share
resources
What determines degree of
interference?
Task similarity
Task difficulty
Practice/expertise
Inhibition in Attention
Inhibition in Attention
Name the color from right to left

Red Blue Green Yellow

Yellow Red Blue Green

Blue Yellow Green Red

Read the color from right to left

Red Blue Green Yellow

Yellow Red Blue Green

Blue Yellow Green Red

Name each of these color patches
as quickly as possible…
 Sustained Attention

• Vigilance - Concerns the ability of

observers
to maintain attention and remain alert to
stimuli
over prolonged periods of time.
 Scientific background

I. Studies of vigilance
• detection tasks
• long duration
• scarcely occuring signals

Vigilance decrement :
increase in RT
decrease in accuracy

Since Mackworth ’s clock test (1948)

 Vigilance Factors

SUBJECT
•Personality (extravert vs. introvert)
•Motivation
•Smoking
TASK
•Task duration
•Rest pauses
•Multiple monitors
•Time sharing, bimodal Vigilance Performance
•Incentives
•Knowledge of results
•Practice
•Pacing

ENVIRONMENT
•Noise
•Stimulation level
•Fatigue and sleep deprivation
•Heat and cold
•Time of day
Sustained Attention

 Current interest in sustained attention is rooted

in the problem encountered by technicians in
handling practical information-processing tasks.

 During the world war I British radar observers on

antisubmarine patrol were often required to carry
out their duties in a setting which included long
term monotonous waiting for uncertain stimulus
events.

 Many hours were spent watching for an

occasional blip of light on a radar screen which
signified the presence of a hostile submarine.
Sustained Attention
 Observers were highly motivated to detect those
crucial signals.

 Despite a high level of accuracy early in their

period of watchkeeping, a surprising change took
place in their behavior as time wore on.

 They failed with increasing frequency to notice

the critical blips displayed by their equipment.

 Sustained attention is also relevant to the

medical setting, as for example when
anesthesiologists monitor electronic equipment
displaying a patient’s life signs during prolonged
surgery.
The decrement function
 Controlled laboratory experiment on sustained
attention is started from the work of Norman
Mackworth (1950).
 Mackworth devised a simulated radar display
called the clock test. The black pointer of the
clock moved 0.3 inch to a now position once
every second. Occasionally, it executed a “double
jump” of 0.6 inch which was the critical signal for
detection.
 Subjects were asked to detect the double jump
and press the key, this task lasted for two hours.
The signals occurred infrequently and
aperiodically.
Results

 Missed signals increased sharply from the

first to second half-hour of the session.
 Showed a more gradual decline for the
remainder of the two hour period.
 Decline in performance is complete from
20 to 35 minutes after the initiation of the
vigil.
 50 percent of the loss is completed within
the first 15 minutes.
Factors affecting vigilance
 Mackwort’s study revealed that a number of
psychological processes on different levels. These
factors are sensory, cognitive, and psycho
physiological control decrement function as well
as the overall level of vigilance efficiency.
Sensory factors
 Sensory modalities in which stimuli are delivered
to the observer are important. In sustained
attention tasks involving auditory signals tends to
be superior to that in tasks involving visual
signals.
 It is also found that experience in one sensory
modality can influence subsequent performance
in another modality.
Signal conspicuity
 Amplitude and duration of critical signals
are important factors in their delectability
over prolonged periods.
 Amplitude

 Performance efficiency increases as the

signal to nose ration of critical signals
increases (Loeb and Binford, 1963).
 In the noise, the critical signal (occasional
increment in the loudness) of 2.1, 3.6, or
5.1 dB were presented and subjects were
asked to detect the change in the intensity
of sound.
Results
 Number of missed signal was inversely related to
the intensity of the critical stimuli. Missed signals
were less in 5.1 dB.
 The decrement function was relatively stable
 In the lowest level of critical signal amplitude,
progressive increase overtime in missed signals
was most prominent.
 The number of missed signal was high in the
lowest level of critical signal amplitude.
 Response latency to correct detections decreases
with increments in critical signal amplitude.
Duration
 Vigilance performance is related to the duration
of critical stimuli. signals of the brief duration are
less likely to detected than those which remain
on the observer’s display for longer periods of
time.
Event rate
 In sustained attention tasks, the frequency of
neutral events, or the background event rate, is
an important factor in determining performance
efficiency.
 In an experiment subjects were shown one bar of
light which moved right from its starting point
and returned back. Again it moved about the
same length and returned back to its point of
origin. This dual movement was considered as
one event.
 The critical signal for detection was an increment
of 4 mm in the length of the second deflection
within an event.
 In this experiment two event rates were used. 5
events/min and 30 events/min. equal number
of critical signals were presented at each event
rate.
Results
 The presented of signal correctly detected was
considerably grater at the slow than at the fast
event rate.
 The decrement function occurred predominantly
in the fast event rate condition.
 Background event rate modifies the effects
associated with other stimulus parameters. Event
rate exert a strong influence on the effects of
critical signal amplitude.
 Factors such as sensory modality of signals, their
amplitude and duration, and the background
event can be considered as first order
psychological factors. In these factors some
immediate physical property of the stimulus is
manipulated.
Temporal and spatial uncertainty
 There are second order factors in which
characteristics of the signal which must be
derived by the observer on the basis of
experience with the task. These factors also play
important role in the maintenance of sustained
attention.
Temporal uncertainty
 Signal detection is related with the number of
critical signals within a fixed time period.
Percentage of signal detection increases as
critical signal density increases over a broad
range of values.
 Variation in the intervals of time between critical
signals can induce uncertainty. Signals can be
presented regularly (i.e., interval of 5 minutes)
or irregularly (interval is varied). Regularity will
induce more certainty and high detection of
signals and irregular presentation will induce
more uncertainty and less detection of signals.
Spatial uncertainty
 Varying the probability that signal will appear in
different positions in the monitored display will
lead to high uncertainty and less detection.
Knowledge of result
 Performance efficiency can be substantially
improved by giving the observers information to
how well they are doing. Knowledge of result
works as motivation factors and improves the
performance.