Nature Neuroscience March 2000 PDF

2000 Nature America Inc. http://neurosci.nature.
com
contents
volume 3 no 3 march 2000
http://neurosci.nature.com
Mice lacking NMDA receptor

function in the CA1 region of the
hippocampus have impaired
editorial
nonspatial memory. Exposure to
2000 Nature America Inc. http://neurosci.nature.com
an enriched environment, Mysterianism lite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

however, ameliorates these
learning deficits and increases
spine density in CA1. See pages
205 and 238.
news and views
Predicting perception from population codes . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Jennifer M. Groh
SEE ARTICLE, PAGE 270
ChIPping away at potassium channel regulation . . . . . . . . . . . . . . . . . . . . . . . . . 202

Min Li and John P. Adelman
Toying with memory in the hippocampus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

Howard Eichenbaum and Kristen Harris
Attention - brains at work! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

Place cell firing during Roger B.H. Tootell and Nouchine Hadjikhani
space flight. SEE ARTICLES, PAGES 284 AND 292
Page 209.
Signaling dendritic growth in vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Sandra Aamodt
brief communications
Three-dimensional spatial selectivity of hippocampal neurons during
space flight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
JJ Knierim, BL McNaughton and GR Poe
Predicting perception from

population coding.
Pages 201 and 270.
Nature Neuroscience (ISSN 1097-6256) is published monthly by Nature America Inc., headquartered at 345 Park Avenue South, New York, NY 10010-1707. Editorial Office: 345 Park
Avenue South, New York, NY 10010. Telephone 212 726 9200, Fax (212) 696 9635. North American Advertising: Nature Neuroscience, 345 Park Avenue South, New York, NY 10010-
1707. Telephone (212) 726-9200. Fax (212) 696-9006. European Advertising: Nature Neuroscience, Porters South, Crinan Street, London N1 9SQ. Telephone (0171) 833 4000. Fax
(0171) 843 4596. New subscriptions, renewals, changes of address, back issues, and all customer service questions in North America should be addressed to Nature Neuro-
science Subscription Department, PO Box 5054, Brentwood, TN 37024-5054. Telephone (800) 524-0384, Direct Dial (615) 377 3322, Fax (615) 377 0525. Outside North America:
Nature Neuroscience, Macmillan Magazines Ltd., Houndsmill, Brunel Road, Basingstoke, RG21 6XS, U.K.. Tel: +44-(0)1256-329242. Fax: +44-(0)1256 812358. Email:
subscriptions@nature.com. Annual subscription rates: U.S./Canada: U.S. $595, Canada add 7% for GST (institutional/corporate), U.S. $195, Canada add 7% for GST (individual
making personal payment BN: 14091 1595 RT); U.K./Europe:435 (institutional/corporate), 185 (individual making personal payment), 99 (student); Rest of world (excluding
Japan): 480 (institutional/corporate), 195 (individual making personal payment), 110 (student); Japan: Contact Japan Publications Trading Co. Ltd., 2-1 Sarugaku-cho 1 chome,
Chiyoda-ku, Tokyo 101, Japan, phone (03) 292-3755. Back issues: U.S./Canada, $45, Canada add 7% for GST; Rest of world: surface U.S. $43, air mail U.S. $45. Reprints: Nature
Neuroscience Reprints Department, 345 Park Avenue South, New York, NY 10010-1707. Subscription information is available at the Nature Neuroscience homepage at http://neu-
rosci.nature.com. POSTMASTER: Send address changes to Nature Neuroscience Subscription Department, P.O. Box 5054, Brentwood, TN 37024-5054. Application to mail periodicals
postage rate is paid at New York, NY. Executive Officers of Nature America Inc: Nicholas Byam Shaw, Chairman of the Board; Jan Velterop, President; Edward Valis, Secretary-Trea-
surer. Printed by Publishers Press, Shepherdsville, KY, USA. Copyright 2000 Nature America Inc.
nature neuroscience volume 3 no 3 march 2000 i

contents
articles
NMDA receptor-mediated control of protein synthesis at developing
synapses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
AJ Scheetz, AC Nairn and M Constantine-Paton
Rho GTPases regulate distinct aspects of dendritic arbor growth in

Xenopus central neurons in vivo. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Z Li, LV Aelst and HT Cline
SEE NEWS AND VIEWS, PAGE 208
Mapping shifts in visual
spatial attention. Anatomical and physiological evidence for D1 and D2 dopamine receptor
Pages 206, 284 and 292. colocalization in neostriatal neurons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
O Aizman, H Brismar, P Uhln, E Zettergren, AI Levey, H Forssberg,
P Greengard and A Aperia
Growth cone and dendrite dynamics in zebrafish embryos: early

events in synaptogenesis imaged in vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
JD Jontes, J Buchanan and SJ Smith
Enrichment induces structural changes and recovery from nonspatial

memory deficits in CA1 NMDAR1-knockout mice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
C Rampon, YP Tang, J Goodhouse, E Shimizu, M Kyin and JZ Tsien
Muscles express motor patterns of non-innervating neural networks by

filtering broad-band input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
LG Morris, JB Thuma and SL Hooper
D1 and D2 receptors
in the striatum. Microsaccadic eye movements and firing of single cells in the striate
Page 226. cortex of macaque monkeys. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
SM Conde, SL Macknik and DH Hubel
Lack of cortical contrast gain control in human photosensitive epilepsy . . . . . . . 259

V Porciatti, P Bonanni, A Fiorentini and R Guerrini
Learning to find a shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

M Sigman and CD Gilbert
Seeing multiple directions of motionphysiology and psychophysics . . . . . . . . 270

S Treue, K Hol and HJ Rauber
Imaging synapse
A multimodal cortical network for the detection of changes in the
formation in vivo. sensory environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Page 231. J Downar, AP Crawley, DJ Mikulis and KD Davis
The neural mechanisms of top-down attentional control . . . . . . . . . . . . . . . . . . 284

JB Hopfinger, MH Buonocore and GR Mangun
Voluntary orienting is dissociated from target detection in human

posterior parietal cortex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
M Corbetta, JM Kincade, JM Ollinger, MP McAvoy and GL Shulman
classified advertising
see back pages
nature neuroscience volume 3 no 3 march 2000 ii

editorial
Mysterianism lite
A philosophical view known as mysterianism holds that even biochemistry to computation) tend to go unreported, because they
though there is nothing supernatural about how consciousness aris- are very difficult to explain to lay people.
es from neural activity, the human brain is simply not equipped to The problem goes deeper than this, though. Even where mecha-
understand it. The reason we find the mindbrain problem so baf- nistic explanations of brain function have been possible, they do not
fling, the argument goes, is that humans did not evolve sufficient feel like explanations of mental processes. Consider the paper by
cognitive abilities to solve it, just as armadillos did not evolve the Treue et al. on page 270 of this issue, which presents a striking exam-
ability to understand arithmetic. This argument has been advocated ple of how far our understanding of perception has progressed. Based
by philosophers such as Colin McGinn and cognitive scientists such on knowledge of how motion is represented by populations of neu-
as Steven Pinker. Now it has been taken up by a prominent science rons in the visual cortex, the authors were able to predict an entire-
journalist, John Horgan, whose new book The Undiscovered Mind ly unexpected visual illusion; two different patterns of moving dots
offers a view of brain science that might best be described as mys- are perceptually indistinguishable, apparently because they both
terianism lite. It is not just consciousness that is beyond our grasp, he evoke the same pattern of activity in a cortical area called MT. Of
says; neuroscience as a whole is failing, because the brain is too com- course a graph showing the distribution of neuronal firing rates in
plicated for human understanding. MT doesnt feel like an explanation of perception. But why should
Horgan attracted attention, even notoriety, for his 1996 book The it? The criterion for a good theory is not that it feels right, but that it
End of Science, in which he argued that the age of great scientific dis- can successfully predict unexpected results. If a physical theory of
coveries is coming to an end because most of the big questions have neural processing can predict an unexpected mental phenomenon,
been answered. The brain is an obvious exception, but Horgan now that is surely a substantial achievement.
argues that neuroscience too is reaching its limits, not because it has It goes without saying that Treues study raises many further
succeeded in its aims but because those aims are unachievable. The issueshow is the population activity decoded, what other areas
subtitle of his new book is How the human brain defies replication, are involved in representing the stimuli, and so forthbut there is no
medication, and explanation; its thesis is that the achievements of reason why questions of this type should not eventually be answered.
neuroscience (along with psychology, psychiatry and other related Certainly, it will be a challenge to understand how (say) a moving
areas) are being oversold, that the supposed practical benefits have red bar is perceived as a unitary stimulus if its orientation, motion
been exaggerated, and that the field is now confronting an explana- and color are each represented in different cortical areas. Horgan
tory gap that may never be bridged. Unlike particle physicists or mol- may well be right that existing hypotheses to solve this so-called
ecular biologists, says Horgan, neuroscientists have yet to achieve binding problem (such as synchronous oscillations) will prove incor-
their reductionist epiphany. Instead of finding a great unifying insight, rect. But to deny the possibility of further progress seems perverse. A
they just keep uncovering more and more complexity. Neurosciences deeper understanding of the mechanisms underlying mental process-
progress is really a kind of anti-progress. As researchers learn more es should follow from greater knowledge of anatomical and func-
about the brain, it becomes increasingly difficult to imagine how all tional connectivity, better methods of recording and manipulating
the disparate data can be organized into a cohesive, coherent whole. neural activity, and more realistic computational models, all of which
It is tempting to dismiss this as another example of what Richard should be achievable with enough time and effort. For ethical rea-
Dawkins once called the argument from personal incredulity, but sons, we may never know as much as we would like about human
Horgan is surely not alone in finding neuroscience difficult to brain activity, but it seems reasonable to expect that insights from
approach. The brain is immensely complicated, and in the absence other organisms will some day provide good models for much of
of a grand unifying theory for how it works, researchers tend to study what happens inside our own heads.
very diverse problems that often seem unconnected to each other. The mysterians might turn out to be correct in claiming that we
It is therefore understandable that their achievements do not always will never fully understand how brain activity leads to subjective
seem intellectually satisfying to nonspecialists. experience: why the firing of MT neurons feels like visual motion,
Part of Horgans critique may reflect how neuroscience is report- why dopamine release in the nucleus accumbens feels pleasurable, or
ed in the media. Among the stories that attract the most attention why electrical stimulation of the anterior supplementary motor area
are the identification of genes or brain areas that are associated with feels amusing. But that is very different from claiming that these
particular behaviors (think of fosB, the gene for maternal behav- phenomena can never be explained in physical terms, or that neu-
ior, or the orbitofrontal cortex, the brains moral compass), but roscience is, as Horgan puts it, bumping up against fundamental
typically such findings are only the first steps on a long road toward limits of science. Most neuroscientists, fortunately, take a more opti-
mechanistic understanding. In contrast, many of the most impor- mistic view than Horgan; the explanatory gap may not be closed at
tant mechanistic insights into how the brain works (at all levels, from a single stroke, but it is getting narrower by the day.
nature neuroscience volume 3 no 3 march 2000 199

news and views
Predicting perception from

population codes
Jennifer M. Groh
Treue and colleagues use electrophysiological recordings in monkeys and psychophysical

experiments in humans to suggest that the shape of a population response in a motion sensitive
region of the brain (area MT), rather than the peak of the response, determines motion perception.
When neuroscientists can consistently for each. In contrast, if the center of grav- What do monkeys see when this hap-
predict what people perceive by studying ity is important, then the location and pens? Salzman and colleagues trained
their neural activity, we will have achieved number of peaks should not matter. Under monkeys to indicate the perceived direc-
a remarkable level of understanding of the latter mechanism, subjects would per- tion of motion, and found that they alter-
brain function. A notable advance toward ceive a single stimulus intermediate nated between reporting the real motion
this goal is presented by Treue, Hol and between the two actual stimuli, regardless direction and the stimulation-induced
Rauber 1 in this issue of Nature Neuro- of whether the population response has motion direction, as if perhaps they could
science. These authors have used the two separate peaks of activity. see both and were simply picking one of
response profiles of neurons in a motion- Visual motion processing is one the two on each individual trial3. Howev-
sensitive area of the monkey brain, area domain where these issues have been er, we trained monkeys to track the
MT, to predict how humans will perceive explored fairly extensively. Motion per- motion using eye movements, and found
moving stimuli. ception is thought to rely on the popula- that the animals responded as if they saw
The brain encodes many kinds of sen- tion responses in visual area MT, which is only the vector average of the two direc-
sory stimuli using maps of neurons that specialized for processing moving stimuli tions4. Both of these experiments likely
are tuned to the properties of those stim- and contains a columnar organization for involved a population response composed
uli. How does the neural activity in these motion direction (for review, see ref. 2). of two peaks of activity: the neurons
maps subserve perception and sensory- Because of this topographical organiza- whose preferred direction of motion
guided action? Because neurons are tion, microstimulation can be used to acti- matched the visual stimulus and the neu-
broadly tuned, a single stimulus typically vate a population of neurons with similar rons at the tip of the microstimulating
activates a large population of neurons motion preferences, thereby simulating the electrode. Perceptual judgments were cor-
the so-called population response. Sever- response to real motion. Microstimulation related with the locations of these peaks,
al different theories have been proposed in concert with an actual moving visual whereas eye movements were correlated
for how population responses in turn stimulus is presumed to cause a popula- with the vector average of activity in MT.
mediate perception and action. The most tion response in MT that corresponds to Microstimulation is artificial, of
obvious possibilities are that perceptual two different directions of visual motion course. What happens when real stimuli
outcome is determined either by the peak the actual direction of motion of the visu- moving in two directions are presented?
of the population response or by its center al stimulus and the preferred direction of When two patches of moving dots are
of gravity (also known as the vector aver- the cells being stimulated electrically. superimposed on each other (a situation
age of the response).
When only one stimulus is present, the
peak and the center of gravity of the pop-
ulation response are the same. But what
happens when two stimuli with different
features occur at the same place and time?
Both stimuli influence the population
response, but are they perceived as inde-
pendent? Do they both contribute to
behavioral responses? How do the two
stimuli interact? If the peak of the popu-
lation response is the most important fea-
ture, then both stimuli would be perceived
so long as the two stimuli are sufficiently Bob Crimi
different from one another that the pop-
ulation response contains a separate peak Fig. 1. Electrophysiological recordings in visual area MT of rhesus monkeys by Treue and col-
leagues1 suggest that the population response to a transparent motion stimulus with two compo-
nents separated by 40 degrees is probably the same as the population response to a transparent
Jennifer Groh is at the Department of motion stimulus with three components (+50, 0, 50 degrees). Treue and colleagues predicted
Psychological and Brain Sciences, Center for that human observers would therefore perceive the two stimuli as containing identical motion.
Cognitive Neuroscience, Dartmouth College, This prediction was confirmed: human observers judged that both stimuli contained the same
Hanover, New Hampshire 03755, USA. upward and rightward component, even though in one case this component had an angle of 40
e-mail: jennifer.m.groh@dartmouth.edu degrees and in the other case it had an angle of 50 degrees.

news and views
known as transparent motion), humans population responses as certain two-com- Perhaps the most intriguing aspect of
can perceive the two stimuli as distinct ponent stimuli. For example, the popula- this work is the notion that the shape of the
provided the directions of motion are sep- tion response to a transparent motion population response in MT can be impor-
arated by at least 10 degrees5. Clearly, the stimulus consisting of two components 80 tant for motion perception. There are well-
center of gravity could not subserve this degrees apart should be the same as the defined algorithms for identifying peaks of
perceptor else we would always perceive response to a motion stimulus with 3 activity (winner-take-all), or computing
transparent motion as containing only components each 50 degrees apart (see the center of gravity (for example, via vec-
one component of motion at a direction Fig. 3 of ref. 1). If so, and if motion per- tor averaging) to arrive at a perceptual
intermediate between the actual direc- ception relies on this population of neu- judgment or behavioral response, and it is
tionsbut what about the peak(s) of the rons, then the direction of motion of these comparatively easy to imagine how neur-
population response? Does the popula- two stimuli should be indistinguishable. al circuits might perform these calculations
tion response in MT contain separate They tested this hypothesis in human (for example, see J.M. Groh, Soc. Neurosci.
peaks for each component of a transpar- observers, and found it was indeed the Abstr. 23, 1560, 1997). Yet the findings of
ent motion stimulus? Do these peaks case: these two very different motion stim- Treue and collegues suggest that percep-
merge together into one broad peak at the uli appear perceptually to have the same tion can be affected by details of the shape
point where the two directions are too components (Fig. 1). of the active population, details that are lost
close to be resolved? A number of issues remain to be through either of these calculations. There-
In an elegant series of experiments, resolved. For example, do MT cells actual- fore, we need to explore new algorithms
Treue and colleagues1 tested this hypoth- ly respond identically to the two- and for reading population codes.
esis. Although the responses of MT neu- three-component stimuli? Do the demands
1. Treue, S., Hol, K. & Rauber, H.-J. Nat. Neurosci.
rons to both single and multiple stimuli of the psychophysical task affect how MT 3, 270276 (2000).
have been well characterized (for review, represents motion information? Monkeys 2. Albright, T. D. in Visual Motion and its Role in the
see ref. 6), it is less clear how the popula- can certainly be trained to perform motion Stabilization of Gaze (eds. Miles, F. A. & Wallman,
tion response varies as a function of the tasks like the one used by Treue and col- J.) 177201 (Elsevier, New York, 1993).
relative directions of the components of leagues in humans, but there is reason to 3. Salzman, C. D. & Newsome, W. T. Science 264,
multiple stimuli. Treue and colleagues first think that the task itself might influence 231237 (1994).
studied the responses of monkey MT neu- population responses in MT. In particular, 4. Groh, J.M., Born, R.T. & Newsome, W.T.
rons to transparent motion stimuli. Their previous work by Treue and others has J. Neurosci. 17, 43124330 (1997).
results show that because these neurons demonstrated that when an animal is 5. Mather, G. & Moulden, B. Q. J. Exp. Psychol. 32,
325333 (1980).
are broadly tuned for direction, the pop- attending to only one of two directions of
6. Britten, K. H. & Heuer, H. W. J. Neurosci. 19,
ulations of neurons responding to each motion, neurons in MT represent the 50745084 (1999).
component of motion overlap quite exten- attended direction much more strongly79.
7. Treue, S. & Maunsell, J. H. Nature 382, 539541
sively. For directions separated by less than Thus, if MT neurons were studied while (1996).
about 90 degrees, only a single broad peak monkeys performed the psychophysical 8. Groh, J. M., Seidemann, E. & Newsome, W. T.
exists (although when the directions are task used here in human observers, the Curr. Biol. 6, 14061409 (1996).
farther apart, two separate peaks do presence and/or location of peaks in the 9. Seidemann, E. & Newsome, W. T.
appear). Importantly, this single peak population response might be different. J. Neurophysiol. 81, 17831794 (1999).
occurs in monkey MT even when the
directions are sufficiently different to be
readily distinguishable to human observers
(and presumably to the monkeys).
Thus, the relationship between neur-
al activity and perception of the compo-
ChIPping away at potassium
nents of transparent motion does not
seem to be based on the presence or channel regulation
absence of segregated peaks of activity, as Min Li and John P. Adelman
would have been predicted by algorithms
that identify peaks of activity (for exam-
Kv4 subunits form A-type potassium channels. To replicate
ple, winner-take-all). Rather, the transi-
tion from perception of two directions of native currents, these subunits require additional factors,
transparent motion to perception of a sin- now shown to be a family of calcium-binding proteins.
gle direction of motion must depend on
some as-yet unidentified aspect of the In a recent issue of Nature, Kenneth concerning the molecular identity of A-
shape of the population response in MT. Rhodes and colleagues1 present results type potassium channels. They describe
If the overall shape of the population that resolve long-standing questions the isolation and characterization of a
response is critical to motion perception, family of calcium-binding proteins, the
then Treue and colleagues reasoned that John Adelman is in the Vollum Institute, Oregon KChIPs (K + channel interacting pro-
stimuli that produce population respons- Health Sciences University, 3181 S.W. Sam teins; Fig. 1), that bind to the intracel-
es having the same shape should produce Jackson Park Road, Portland, Oregon 97201- lular amino (N)-terminal domain of
the same percepts. Based on their record- 3098, USA. Min Li is in the Department of cloned Kv4 channels and endow them
ings using two-component stimuli, Treue Physiology, Johns Hopkins University School of with many of the properties ascribed to
and colleagues designed three-component Medicine, Baltimore, Maryland 21205, USA. native A-type potassium channels. Co-
stimuli that should produce the same e-mail: adelman@ohsu.edu expression of the KChIPs and cloned
202 nature neuroscience volume 3 no 3 march 2000

news and views
Kv4 subunits increased current densi- tion 6 , increasing their similarity to immunoprecipitated with Kv4 subunits
ties, shifted the voltage dependence of native channels. In addition, a role for from transfected cells and rat brain
activation and speeded their recovery calcium in modulating A-type channels membrane preparations. Immunocyto-
from inactivation. Mutagenesis experi- has been suggested from recordings of chemistry confirmed that the two pro-
ments that eliminated the ability of the cholinergic neostriatal neurons, where teins are colocalized at cellular and
KChIPs to bind calcium showed that blocking voltage-dependent calcium subcellular levels. Remarkably, co-expres-
their channel-modulating effects require channels with cadmium shifts the volt- sion of KChIPs with Kv4 subunits recon-
calcium binding, although the physical age dependence of A-type current acti- stituted the functional characteristics of
association between the interacting pro- vation and inactivation to more native A-type channels, as well as repli-
teins was calcium independent. depolarized potentials3. Until now, how- cating the effects of co-expressing rat
A-type channels are voltage-depen- ever, the relationship between these brain mRNA on cloned Kv4 subunits.
dent potassium channels that activate in observations has remained obscure. Current densities were increased, the
the subthreshold range of membrane The new report from Rhodes and col- voltage dependence of activation was
potentials and completely inactivate leagues 1 reconciles these differences shifted to more hyperpolarized poten-
during depolarizing pulses, while other between native A-type potassium chan- tials, and the channels recovered from
voltage-dependent potassium channels nels and cloned Kv4 channels at the mol- inactivation much more rapidly.
are just beginning to activate. As a ecular level. Using the intracellular The KChIPs range in size from 216
result, A-type channels influence the N-terminal region of Kv4.3 as bait in a to 256 amino acids. The N-terminal
time required for membrane depolar- yeast two-hybrid hunt through a rat mid- domains, 50 amino acids, vary consid-
ization to reach threshold for action brain cDNA library, they captured and erably, but throughout their carboxyl
potential generation as well as the time characterized three members of a family (C)-terminal domains, they share 70%
between action potential spikes. Thus of Kv4 channel interacting proteins sequence identity. The conserved
they are important determinants of the (KChIPs). The KChIPs interact selective- regions of the molecules contain four
firing frequency in excitable cells such ly with Kv4 subunits, and they are E-F hand calcium-binding motifs. Cal-
as neurons and cardiac myocytes. expressed in tissues that also express Kv4 cium binding to the KChIPs was con-
Sequence homologies among cloned subunits. When co-expressed in heterol- firmed by calcium-dependent mobility
subunits distinguish several subfamilies ogous cells, KChIPs were localized with shifts. Interestingly, a mutant KChIP-1
of voltage-dependent potassium chan- Kv4 subunits. KChIPs were selectively that could not bind calcium still inter-
nels, the Kv subfamilies. The Kv chan-
nels are tetramers, and heteromeric
channels are assembled only from sub-
units in the same subfamily. Heterolo-
gous expression studies show that K+
members of the Kv4 subfamily form K+
channels similar (but not identical) to

native A-type potassium channels 2. In
particular, they demonstrate inactiva-
tion mediated by a specialized intracel- KChIP KChIP
lular domain at the extreme N terminus, KChIP
the ball, which physically occludes the
pore when the channel is open. The
expression profiles of Kv4 subunits in Ca2+
the CNS and heart are consistent with KChIP
this role, and Kv4 subunits have been
identified as the components of A-type
potassium channels in rat neostriatal
cholinergic interneurons 3 and cardiac
ventricular myocytes4. KChIP KChIP
The functional profiles of Kv4 chan-
nels expressed in different heterologous Golgi Ca2+ Ca2+
cells are variable5, suggesting a contri-
bution by factors present in the host
cell. This is further supported by differ-
ences between Kv4 channels expressed
in Xenopus oocytes with or without rat
brain mRNA. Co-expression of the low-
molecular-weight mRNA fraction
Fig. 1. KChIPs are integral components of A-type potassium channels. The KChIPs bind to the
increases current density, most likely
N-terminal domain of Kv4 subunits, close to the membrane, and close to the ball domain that
reflecting an increased number of A- mediates channel inactivation. The physical association between KChIPs and Kv4 subunits does
type channels. This co-expression also not require calcium binding, but the effects on channel gating are calcium dependent. The KChIPs
shifts the activation voltage of the chan- are also present in the Golgi, where the association with Kv4 channels during their biosynthetic
nels to more negative potentials and development may regulate the levels of functional A-type potassium channels present in the sur-
allows faster recovery from inactiva- face membrane.

news and views
acted with Kv4 subunits, suggesting that KChIPs for the various Kv4 subunits? ing associations to members of the same
the association between channels and What other proteins might KChIPs subfamily. The N-terminal region also
KChIPs is calcium independent and may bring into the A-type channel complex? interacts with the Kv proteins, some of
be constitutive. However, the mutant This is the latest in a series of find- which endow inactivation that is mod-
KChIP-1 no longer modulated Kv4 ings suggesting that intracellular calci- ulated by protein kinases13. More recent-
channel function. um signaling can modulate membrane ly, two alternatively spliced proteins,
Sequence alignments show that potential. Fluctuations in intracellular ZIP1 and ZIP2, have been identified
KChIPs belong to the recoverin family calcium levels have long been appreci- that act as molecular bridges, linking the
of calcium-binding proteins that ated as an important modulator of ion subunits to protein kinase C
includes the Drosophila protein fre- channel activity. A generally accepted (ref. 14). At their C termini, many volt-
quenin, which regulates transmitter model posits that second messenger sys- age-dependent potassium channels bind
release, and mammalian homologs tems, such as calcium-sensitive protein to PDZ-containing proteins, such as
NCS-1 (neuronal calcium sensor-1) and kinases or phosphatases, alter the phos- PSD-95/SAP90 family members, to
hippocalcin7. Although KChIPs 1 and 2 phorylation status of the channel, affect- modulate the distribution and surface
have not been previously described, ing channel activity and cellular expression of the potassium channels,
KChIP-3 is identical to DREAM (down- excitability10. These are relatively slow which is affected by the presence of a
stream regulatory element antagonist processes that require calcium interac- subunit at the N-terminal domain in
modulator), which is reported to regu- tion with the signaling molecule and some cases 15 . These PDZ-containing
late transcription in a calcium-depen- subsequent interaction with the ion proteins, in turn, interact with other
dent manner8. KChIP-3 is also identical channel. putative regulatory molecules. The
to calsenilin, a protein that resides in the Recently, several reports have sug- emerging picture suggests that the
ER and Golgi and interacts with the gested a faster calcium signaling process subunits of potassium channels are
C-terminal domains of the presenilins, by demonstrating that calmodulin is embedded in large, multimeric protein
transmembrane proteins found in the constitutively bound to the ion-con- complexes with components that sense a
same subcellular compartments. Muta- ducting subunits of voltage-depen- wide range of metabolic signals. Indeed,
tions in the presenilin genes account for dent calcium channels (VDCCs) and the ZIP proteins, which do not them-
40 percent of early-onset familial small conductance calcium-activated selves interact with the subunits,
Alzheimer disease and sensitize neu- potassium channels (SK channels). In begin to define a larger microdomain, a
ronal cells to apoptosis, possibly by dis- both cases, compelling evidence sup- molecular neighborhood, in which the
rupting intracellular calcium levels9. In ports a model in which calmodulin is an channels reside. The two-hybrid screen
this regard, it is tempting to speculate integral part of the channel complex, can be used with each newly identified
that the increased current density and calcium binding to calmodulin resident to determine the next nearest
(thought to reflect an increased number induces structural alterations in neighbor.
of channels in the membrane) induced calmodulin, which are transduced into
by co-expression of KChIPs with Kv4 conformational changes in the channel 1. An, W. F. et al. Nature 403, 553556 (2000).
subunits may reflect effects on mem- proteins that alter their function 11,12 .
2. Serodio, P., Vega-Saenz de Miera, E. & Rudy,
brane trafficking, which could be relat- Local signaling induced by calcium B. J. Neurophysiol. 75, 21742179 (1996).
ed to the interactions between KChIP-3 entry may be much more rapid than 3. Song, W.-J. et al. J. Neurosci. 18, 31243137
and presenilins in the trans Golgi. second-messenger-mediated processes, (1998).
Taken together, the results suggest and can respond rapidly to discrete, 4. Yeola, S. W. & Snyders, D. J. Cardiovasc. Res.
that KChIPs are integral parts of Kv4 localized alterations in intracellular cal- 33, 540547 (1997).
channels, both in the plasma membrane cium. For example, VDCCs are the like- 5. Petersen, K. R. & Nerbonne, J. M. Pflugers
and during their biosynthetic journey ly source of the calcium ions that would Arch. 437, 381392 (1999).
through the trans Golgi. On elevation of bind to channel-associated calmodulin, 6. Serodio, P., Kentros, C. & Rudy, B. J.
intracellular calcium levels, such as dur- and the precise distance between SK Neurophysiol. 72, 15161529 (1994).
ing an action potential, calcium bind- channels and VDCCs or intracellular 7. Pawlowski, K., Bierzynski, A. & Godzik, A. J.
Mol. Biol. 258, 349366 (1996).
ing to KChIPs induces a conformational calcium release sites strongly affects the
alteration that is rapidly transduced to dynamics of burst frequency. The work 8. Carrion, A. M., Link, W. A., Ledo, F.,
Mellstrom, B. & Naranjo, J. R. Nature 398,
the channel, resulting in altered gating. by Rhodes and colleagues1 suggests that 8084 (1999).
The previous localization of KChIP-3 to KChIPs are an integral component of A- 9. Buxbaum, J. D. et al. Nat. Med. 4,
the trans Golgi suggests that the type potassium channels, acting as 11771181 (1998).
increased current densities may result direct calcium sensors that affect chan- 10. Levitan, I. B. Annu. Rev. Physiol. 56,
from regulated cell-surface expression nel gating properties, analogous to the 193212 (1994).
of the channel complexes. role of calmodulin for SK channels and 11. Zuhlke, R. D., Pitt, G. S., Deisseroth, K.,
Several questions await further VDCCs. Tsien, R. W. & Reuter, H. Nature 399,
159162 (1999).
experiments. For example, just how The KChIPs join an expanding list of
does calcium binding to the KChIP proteins that bind to potassium chan- 12. Keen, J. E. et al. J. Neurosci. 19, 88308838
(1999).
result in altered channel activity? Which nels and influence their activity, but do
13. Sheng, M. & Kim, E. Curr. Opin. Neurobiol.
amino acids in the N-terminal domain not contribute to ion conduction. 6, 602608 (1996).
of Kv4 subunits interact with KChIPs? Among the voltage-dependent potassi- 14. Gong, J. et al. Science 285, 15651569
Does the interacting domain of Kv4 um channels, the intracellular N-termi- (1999).
overlap with binding sites for other pro- nal domain has been shown to mediate 15. Tiffany, A. M. et al. J. Cell Biol. 148, 147158
teins? How selective are the different heteromeric subunit assembly, restrict- (2000).

news and views
without NMDA-receptor-dependent LTP?

Toying with memory in the Does the observed anatomical plasticity
compensate for the impaired functional
hippocampus plasticity, and if so, how? Here we consid-
er two interpretations of Tsien and col-
Howard Eichenbaum and Kristen Harris leagues2 that differ in the critical locus
where increased connectivity ameliorates
Mice lacking NMDA receptors in hippocampal area CA1 are the loss of NMDA-receptor-dependent
LTP in area CA1. First, enhanced intrin-
deficient in spatial memory. They also have nonspatial memory
sic hippocampal connectivity might com-
deficits, which are overcome by environmental enrichment. pensate for the loss of LTP through
NMDA-receptor-independent processes.
Memory being... altogether con- and colleagues2 show that mice with the Second, enhanced connectivity outside
ditioned on [the ability to excite] same mutation are severely impaired the hippocampus, specifically within the
brain-paths, its excellence in a across a broad range of nonspatial learn- neocortex where NMDA-receptor-depen-
given individual will depend part- ing tests. The authors also address Jamess dent plasticity is intact, might compen-
ly on the number and partly on number of paths factor by exposing sate for a dysfunctional hippocampus.
the persistence of these paths. these mice to a complex enriched envi- According to the first possibility, the
(William James1, p. 659) ronment, in which they could presumably hippocampus might use processes that do
establish many diverse associations not require NMDA receptors after expo-
In this two-factor view of the biological through exploration. By electron sure to environmental enrichment. In CA1,
basis of memory, James characterized the microscopy, environmental enrichment LTP can be induced by strong patterned
persistence factor as a physiological prop- was shown to increase the number of stimulation, even when NMDA receptors
erty of ones brain tissue. He envisioned synaptic connections within hippocam- are pharmacologically blocked5. This form
persistence as a native tenacity, except for pal area CA1 in normal mice, and sur- of LTP is NMDA receptor independent.
natural variability among individuals and prisingly also in the mutant mice, even Tsien and colleagues2 show that many new
decline with illness or aging. By contrast, without NMDA receptors. Furthermore, dendritic spines form, and robust synap-
James characterized the number of paths enrichment improved learning perfor- togenesis occurs within CA1 after enrich-
factor as very much modifiable with expe- mance in control mice and almost elimi- ment experience (Fig. 1b) in both mutant
rience. He argued that memory could be nated the memory deficits observed in the and control mice. This synaptogenesis does
improved substantially by establishing a CA1 NMDA receptor-knockout mice. not depend on NMDA receptors because
large network of linkages through which How is it possible that enriched expe- an equal number of new dendritic spines
one could readily associate, and later access, rience can support new memory even and synapses formed in mutant and con-
a new memory. As an example, James
described the college athlete who was a
dunce at his books but could astonish Standard Enriched
with his ability to remember sports statis-
tics precisely because he had worked at cre-
ating a rich knowledge framework for this Cortex
kind of information. James may have been
prescient in proposing that enriching ones
memory network can make up for a lesser
native persistence, as in findings from Joe
Tsien and colleagues2 in the current issue of Hippocampus
Nature Neuroscience.
This report extends a recent study that
used state-of-the-art molecular genetics
to knock out the N-methyl-D-aspartate
(NMDA) receptor beginning at postnatal
weeks three and four selectively within the
CA1 region of the hippocampus3. These
mutant mice lack NMDA-receptor-depen-
Bob Crimi
dent long-term potentiation (LTP), a type
of physiological plasticity thought to be a Fig. 1. Environmental enrichment increases the connections within the hippocampus and neocor-
cellular substrate of memory. Corre- tex. (a) Under control conditions, there are fewer synapses within both the hippocampus and cor-
spondingly, they have severely impaired tex, as well as between these areas. The pathway through the hippocampus could be required to
spatial learning and memory4. Now Tsien connect distinct representations in the neocortex (red and blue), and this capacity could be medi-
ated by strengthening the existing connections within the hippocampus using NMDA receptors
(yellow). (b) After exposure to an enriched environment, more connections are formed within
Howard Eichenbaum is in the Department of both the hippocampus and neocortex, and perhaps between these areas. Strengthening of addi-
Psychology, and Kristen Harris is in the tional non-NMDA receptor connections (orange) within the hippocampus, or between the hip-
Department of Biology, Boston University, pocampus and cortex, may suffice to improve memory. Alternatively, the additional connections
Boston, Massachusetts 02215, USA. within the neocortex (green) may suffice to link distinct neocortical representations and thereby
e-mail: hbe@bu.edu or harrisk@bio.bu.edu short circuit the hippocampal contribution.

news and views
trol mice following environmental enrich- can induce enough connectivity outside hypothesis discussed above would be to
ment. Other studies have shown robust the hippocampus, specifically within the determine whether the CA1-NMDA
synaptogenesis in the adult brain when neocortex. Tsien and colleagues did not knockout mice after enrichment can tol-
synaptic activity is silenced pharmacologi- examine the cortex, but previous evi- erate loss of hippocampal area CA1 and
cally6,7. The new spines form either when dence indicates that enriched experience still enjoy improved learning and memo-
presynaptic release of neurotransmitter is increases intrinsic connectivity within the ry. An early study15 found that enriched
blocked or when postsynaptic glutamate neocortex12. It is clear that memory is not experience reduced, but did not eliminate,
receptors are blocked, and new spines can mediated solely by CA1, or even by the the effects of hippocampal damage on
last for at least eight hours without subse- entire hippocampus alone. Rather, the spatial learning. These findings are con-
quent activation. Furthermore, induction hippocampus is part of a memory system sistent with the possibility that both puta-
of NMDA-receptor-dependent LTP in hip- that prominently involves its bidirection- tive mechanisms contribute to the effects
pocampal area CA1 does not require the al connections with diverse and inter- of enrichment.
formation of new synapses8,9. Together connected regions of the cerebral cortex13
with the findings from Tsien and col- (Fig. 1). Within this system, memories are 1. James, W. The Principles of Psychology (Holt,
leagues, these studies show that dendritic likely stored among large cell assemblies New York, 1890).
spines can form in the mature brain with- widespread across the cortex, and the 2. Rampon, C. et al. Nat. Neurosci. 3, 238244
out NMDA-receptor-dependent processes organization of associations is mediated (2000).
like LTP, and even without synaptic activi- by the formation of links between the cell 3. Tsien, J. Z. et al. Cell 87, 13171326 (1996).
ty. Perhaps the new spine synapses can assemblies10. The role of the hippocam- 4. Tsien, J. Z., Huerta, P. T. & Tonegawa, S. Cell
facilitate NMDA-receptor-independent pus may be to facilitate the consolidation 87, 13271338 (1996).
processes within the hippocampus to of these cortical linkages by storing 5. Morgan, S. L. & Teyler, T. J. J. Neurophysiol. 82,
736740 (1999).
enhance subsequent learning and memo- aspects of new information, or indices
ry in CA1-NMDA knockout mice. pointing to cortical loci of new represen- 6. Bravin, M., Morando, L., Vercelli, A., Rossi, F.
& Strata, P. Proc. Natl. Acad. Sci. USA 96,
How might the enrichment-induced tations, and using these to temporarily 17041709 (1999).
dendritic spines within the hippocampus link otherwise separated cortical memo- 7. Kirov, S. A. & Harris, K. M. Nat. Neurosci. 2,
facilitate learning and memory? Hebb10 ries (Fig. 1a). We know that the role of 878883 (1999).
originally suggested that learning and the hippocampus is temporary because it 8. Muller, D. Rev. Neurosci. 8, 7793 (1997).
memory occurs by strengthening some is not necessary for the recall of long- 9. Sorra, K. E. & Harris, K. M. J. Neurosci. 18,
connections and weakening other, inap- established memories, suggesting that 658671 (1998).
propriate connections. Tsien and collegues eventually new intracortical connections 10. Hebb, D. O. The Organization of Behavior
show that the enrichment effects are spe- form to mediate permanent links14. The (Wiley, New York, 1949).
cific to a particular type of spine synapse, increase in synaptic connectivity in neo- 11. Toni, N., Buchs, P. A., Nikonenko, I., Bron,
causing an increase only in those with a cortex, likely to have occurred as a result C. R. & Muller, D. Nature 402, 421425
(1999).
continuous (that is, non-perforated) of enriched training experience12, might
12. Klintsova, A. V. & Greenough, W. T. Curr.
postsynaptic surface. There was no change be so effective that lasting plasticity with- Opin. Neurobiol. 9, 203208 (1999)
in the frequency of large irregularly shaped in the hippocampus is not required (Fig.
13. Eichenbaum, H. Annu. Rev. Psychol. 48,
synapses, those with perforated postsy- 1b), at least for the relatively simple types 547572 (1997).
naptic surfaces. Thus, the non-perforated of learning examined by Tsien and col- 14. Squire, L. R. & Alvarez, P. Curr. Opin.
synapses might enhance some forms of leagues2. Neurobiol. 5, 169177 (1995).
learning and memory via NMDA-recep- One way to distinguish the cortical 15. Hughes, K. R. Can. J. Psychol. 19, 325332
tor-independent mechanisms. Other stud- hypothesis from the hippocampal (1965).
ies have shown a transient elaboration of
a subset of perforated synapses with
NMDA-receptor-dependent LTP 11. An
open question is whether NMDA-recep-
tor-dependent changes at perforated
synapses might be involved in refinement
Attention - brains at work!
of synaptic connections during more com- Roger B.H. Tootell and Nouchine Hadjikhani
plex learning protocols than those tested
by Tsien and colleagues2. Either way, these Two new studies use event-related fMRI to reveal a network
findings are among the first to demon- of brain regions that are activated during different steps in
strate a possible role for non-perforated the control of visual spatial attention.
synapses in learning and memory. Under-
standing the function of the small non-
perforated synapses is especially important The amount of information that is handle. Much of this information must
because these are normally the most abun- potentially available through our sense therefore be discarded, and the brain
dant synapse type (> 75%) in both hip- organs is far greater than our brains can must select only those stimuli that are
pocampus and neocortex. of greatest relevance for further pro-
The second possible explanation for The authors are at the Magnetic Resonance cessing. Understanding how this occurs
the findings of Tsien and colleagues2 is Imaging Center, Department of Radiology, is a major challenge for cognitive neu-
that the hippocampus can be short-cir- Massachusetts General Hospital, 149 13th Street, roscience, and two papers1,2 in the cur-
cuited altogether during learning and Charlestown, Massaschusetts 02129, USA. rent issue of Nature Neuroscience
memory if environmental enrichment e-mail: tootell@nmr.mgh.harvard.edu provide the most detailed spatio-tem-

news and views
subjects know where to presentation. Unlike block designs, event-

expect the stimulus, related designs can reveal the time course
compared to trials in of the response during an individual trial,
DLPF
SPL IPL
which they do not know making it possible to identify different
or are misdirected. patterns of activation associated with dif-
PS
LO
Attention has some- ferent components of the task.
times been likened to a Both groups used covert attention
Cue
a
spotlight, and func- tasks, thus avoiding any complications
tional neuroimaging due to eye movements. The subjects were
has recently allowed instructed to fixate on the center of a
researchers to see the screen and then shift their focus of atten-
M
beam directly38. Sev- tion to either the left or the right, as indi-
SPL
IPL
eral studies have con- cated by a cue at the fixation point. A few
TPJ
firmed that attention is seconds later, a target appeared either at
VIP
mapped topographical- the cued location (valid cue condition)1,2
ly in all the early or on the opposite side (invalid condi-
Target (valid)
b
(retinotopic) visual tion)2, and subjects had to respond to it.
cortical areas, and that Both groups confirmed that their subjects
when attention is really were making attentional shifts dur-

directed to a particular ing the task. Corbetta et al.2 showed that
M
SPL
location, the part of the their subjects reaction times were faster
IPL
TPJ
cortex that represents after valid than invalid cues, and Hopfin-
VIP
that location becomes ger et al.1 showed that the neural activity
increasingly responsive. evoked by the arrow cue (which, being in
But where and how are the center, could activate both hemi-
Target (invalid)
c attention signals first spheres) was greater in the hemisphere
Bob Crimi
generatedin other that represents the cued side.
Fig. 1. Brain areas activated in a 'naked' eye and brain, from a subject words, how is the Despite their similar techniques, the
who was facing a display screen and doing a covert attention task, spotlight controlled? two studies addressed different questions
similar that used in Hopfinger et al.1 and Corbetta et al.2 (a) A cue Answering this ques- and yielded complementary results.
instructs the subject to attend to a given location: here to the sub- tion is now an impor- Hopfinger et al.1 made few prior assump-
ject's left. Then the attention target appears (here a checkerboard tant goal for the field9. tions, and simply asked which brain
stimulus, but usually a more subtle stimulus change), at either the One problem with regions were activated in response to the
expected location (b) prompted by the 'valid' cue, or at an unex-
most previous neu- cues (reflecting an attentional shift) and
pected location (c), misdirected by the 'invalid' cue in (a). The acti-
vated areas described in Hopfinger et al. and Corbetta et al. include
roimaging studies of which were activated by the subsequent
DLPF (dorsolateral prefrontal cortex), IPL (inferior parietal lobule), attention (as well as target presentation (reflecting processing
LO (lateral occipital region of visual cortex), M (supplementary many other cognitive of the attended stimulus). Cues and tar-
motor region), PS (peri-sylvian), SPL (superior parietal lobule), TPJ processes) is that they gets both activated a number of different
(temporal-parietal junction) and VP (ventral parietal region). have used a block regions; the surprising finding was that
Additional areas were activated but are not visible from this vantage design, in which the there was relatively little overlap between
point (see refs. 1 and 2). High levels of activity are shown in red, and hemodynamic signal is the two sets of responses, suggesting that
lower levels of activity are shown in green. averaged over many the brain structures that control spatial
similar trials. This gen- attention are largely distinct from those
erates a static activation that participate in the processing of the
map that represents the attended stimulus.
poral views yet of the brain structures average activation for a particular task, In a more hypothesis-driven approach,
that control the deployment of visual without giving any information about Corbetta et al.2 tested two specific propos-
spatial attention. the individual steps involved. Yet spatial als regarding the role of parietal cortex in
Our focus of attention is constantly attention is inherently dynamic, and our attention. Based on studies of brain-dam-
shifting, either automaticallyin response brains are constantly choosing new loca- aged patients, it has been suggested that
to an attention-grabbing stimulusor tions of interest, disengaging attention the region around the temporal-parietal
voluntarily. Usually, an attentional shift is and (often) eye position from previously junction (TPJ) is involved in reorienting
followed by an eye movement to the attended locations, and shifting attention attention toward stimuli at unexpected
newly attended location, but it is also pos- and eye position to new targets. It is dif- locations, and that the region around the
sible to attend to a location without ficult to resolve these different steps using intraparietal sulcus (IPs) is involved in vol-
looking at it; we are sometimes forced to a block design. untary orientation and maintenance of
do this during demanding visual tasks The new studies1,2 avoid this problem attention at cued locations. As described
(such as driving on a busy road), where it by using event-related designs. Event- below, their data support both these ideas,
is impossible to fixate all items of interest related fMRI is a relatively new analytical and provide a view of parietal function
simultaneously. In the laboratory, these method, in which the hemodynamic sig- that is complementary to, and largely con-
so-called covert attentional shifts nal is analyzed on a trial-by-trial basis to sistent with, that of the other study.
can be detected because reaction identify patterns that occur at fixed times Both groups agree on the role of a pos-
times are shorter for trials in which after a given event, such as cue or target terior parietal region in and around the

news and views
intraparietal sulcus; this region is activat- could happen except through top-down tion fits very well with the clinical litera-
ed in response to the cue and remains signals, and the challenge now will be to ture on parietal neglect, and the link
active as attention is maintained, but identify the anatomical connections that seems even more compelling given that
shows a reduced response once the target underlie these effects. the other activations seen in these stud-
is presented (regardless of whether it The Corbetta et al. 2 study has some ies were not lateralized.
appears at an expected or unexpected interesting clinical implications. For many In conclusion, these two papers demon-
location)2. Similar responses have been years, it has been known that damage to the strate the power of new imaging techniques
observed in electrophysiological record- right parietal cortex, particularly the tem- to resolve complex cognitive operations
ings from alert monkeys (see ref. 2 for ref- poral-parietal junction, causes a complex into their component steps, and to reveal
erences), and the combined evidence from syndrome known as unilateral visual the neural structures involved in each step.
physiology and neuroimaging strongly neglect (reviewed in ref. 12). Parietal neglect They are likely to stimulate many future
suggests that the posterior parietal cortex patients have problems attending to and studies, and by combining ever-better
is involved in selecting a location and responding to objects on in the left visual imaging methods with other approaches
retaining it in working memory (although field; for instance, they often bump into such as patient studies and physiology of
other areas may also be involvedsee objects on their left, and when asked to non-human primates, we can hope to gain
below). Interestingly, the greater region of draw what they see, they tend to neglect a new depth of understanding of how the
posterior parietal activation may include what is in the left visual field. The syndrome brain controls attention.
the visual cortical area V7, which is retino- has attracted a great deal of interest, not
topically organized (albeit crudely), sug- only for its clinical importance but also 1. Hopfinger, J. B., Buonocore, M. H. & Mangun,
gesting a possible role in the spatial because of its implications for normal per- G. R. Nat. Neurosci. 3, 284291 (2000).
allocation of attention3,9,10. ceptual mechanisms. One interpretation of 2. Corbetta, M., Kincade, J.M., Ollinger, J.M.,
Another popular candidate for storing parietal neglect is that the TPJ is responsible McAvoy, M.P. & Shulman, G.L. Nat. Neurosci. 3,
292297 (2000).
spatial cues in working memory is the for disengaging attention from its present
dorsolateral prefrontal cortex (see ref. 1 focus and redirecting it to a new target. 3. Tootell, R. B. H. et al. Neuron 21, 14091422 (1998).
and references cited therein). Hopfinger Corbetta et al.2 now provide elegant 4. Brefczynski, J. & DeYoe, E. Nat. Neurosci. 2,
370374 (1999).
et al.1 observed activation of this region support for this hypothesis. Unlike other
during the cueing period, but Corbetta parts of the parietal cortex, the TPJ 5. Somers, D. C. et al. Proc. Natl. Acad. Sci. USA
96, 16631668 (1999).
et al.2using a better analytical method showed little or no response to the initial
6. Ghandi, S., Heeger, D. & Boynton, G. Proc.
that avoided prior assumptions about the cue, but it responded strongly to the sub- Natl. Acad. Sci. USA 96, 33143319 (1999).
time course of the hemodynamic sequent presentation of the target. More-
7. Martinez, A. et al. Nat. Neurosci. 2, 364369
responsefound that the activation in the over, the TPJ response was much stronger (1999).
prefrontal cortex was more transient than for invalid than for valid targets, suggest- 8. Kastner, S. et al. Science 282, 108111 (1998).
that observed in the intraparietal cortex. ing that it is specifically involved in reori- 9. Corbetta, M. Proc. Natl. Acad. Sci. USA 95,
Thus, the intraparietal cortex seems to be enting of attention in cases where the 831838 (1998).
the stronger candidate for storing spatial target appears at an unexpected location. 10. Culham, J. C. et al. J. Neurophysiol. 80,
working memories, although it is possi- Finally, the TPJ response was always 26572670 (1998).
ble that both regions are involved, partic- stronger in the right hemisphere than the 11. Desimone, R. & Duncan, J. Annu. Rev.
ularly as they are known to be left, regardless of the side where the tar- Neurosci. 18, 193222 (1995).
interconnected in monkeys (and presum- get was presented. This right lateraliza- 12. Mesulam, M. M. Ann. Neurol. 10, 309325 (1981).
ably in humans too).
Most previous models of visual atten-
tion have assumed that it is controlled by
higher cortical regions, which regulate the Signaling dendritic growth in vivo
processing of sensory inputs in lower
regions via top-down projections. This Small GTPases of the Rho family affect cell morphology by regulating the cytoskeleton,
makes sense because decisions about allo- and they have been implicated in neurite outgrowth. On page 217 of this issue, Holly
cating attention are often based on high- Cline and colleagues (Cold Spring Harbor Laboratory, New York) report that RhoA, Rac
level features that are not represented at and Cdc42 regulate different aspects of dendritic growth in vivo. The authors used
the earliest stages of the cortical hierar- vaccinia virus to express constitutively active or dominant-
chy. However, alternative, bottom-up negative forms of these GTPases in albino Xenopus tadpoles.
models have also been proposed 11 , in Time-lapse imaging of DiI-labeled neurons showed that
which attention arises as an emergent constitutively active Rac and, to a lesser extent, Cdc42
property from competitive interactions increased branch addition and retraction. Activation of
at each level in the hierarchy. The new endogenous RhoA promoted the elongation of existing
findings do not completely exclude the branches. Cline has previously shown that blocking NMDA
latter model, but they are more consistent receptors reduces dendritic growth, and the dominant-
with a top-down model. In particular, negative form of RhoA prevented this effect, suggesting that
both groups found that a cue instructing RhoA may act downstream of NMDA receptors to control
subjects to attend to a particular location dendritic development.
caused increased activation of the corre-
sponding parts of the early retinotopic Sandra Aamodt
visual areas, even before any stimulus
appeared. It is difficult to see how this

Three-dimensional spatial during space flight10, it was of interest to determine whether the
hippocampus can create a stable representation of space during
selectivity of three-dimensional movement in the absence of gravity.
During the Neurolab Space Shuttle mission of April, 1998,
hippocampal neurons ensembles of place cells were recorded from three rats implanted
with a multi-electrode recording array11,12. The rats were trained
during space flight to negotiate a three-dimensional track (the Escher staircase) in
which 3 turns of 90 in yaw were interleaved with 3 turns of 90 in
pitch (Fig. 1). As a result, the rat completed a full circuit of the
James J. Knierim1,2, Bruce L. McNaughton1 and track and returned to its starting location/direction after having
Gina R. Poe1 made only 3 right turns (270 total yaw). The spatial information
provided by external landmarks was thus presumably in conflict
1 University of Arizona, Arizona Research Laboratories Division of Neural with the direction information from HD cells, which under nor-
Systems, Memory & Aging, 384 Life Sciences North Bldg., Tucson, Arizona,
85724, USA
mal conditions require a fourth 90 turn (360 total yaw) to signal
2 a return to the starting direction. Nonetheless, place cells even-
University of Texas-Houston Medical School, Department of Neurobiology &
Anatomy and the W. M. Keck Center for the Neurobiology of Learning and tually demonstrated normal, spatially specific firing properties.
Memory, P.O. Box 20708, Houston, Texas 77225, USA Spatial firing patterns of 16 active place cells from rat 2 were
recorded on the ninth day of flight (FD9; Fig. 2a), which was the
Correspondence should be addressed to B.L.M. (bruce@nsma.arizona.edu)

second day in which the rats had been exposed to the Escher stair-
Place cells of the hippocampus and head-direction (HD) cells case track in flight. For comparison, we show the spatial firing
of the thalamus and limbic cortex derive their spatial and direc- patterns of 12 representative place cells from the same rat record-
tional specificity from a combination of idiothetic (self-motion) ed 4 days before launch as the rat ran clockwise on a flat, rectan-
cues and external landmarks, which normally reinforce each other gular track (Fig. 2b). An index of spatial tuning specificity, which
to generate a robust neural code for location and direction1. In quantifies the amount of information about the rats position
weightlessness, however, three-dimensional navigation can cause transmitted by the firing of a single spike13, did not significantly
the idiothetic and landmark cues to conflict. Nonetheless, neur- differ between cells recorded preflight and those recorded on FD9
al recordings on the space shuttle demonstrated that the hip- (for active cells, defined as having a mean firing rate > 0.05 Hz
pocampus can create a robust spatial code for three orthogonal on the track; mean information per spike s.e. preflight,
surfaces in the weightless environment of space flight. 1.12 0.08 bits, n = 19; FD9, 1.13 0.09 bits, n = 16; not signif-
The firing properties of place cells and HD cells are coupled2, icant by Mann-Whitney). We also determined the spatial tuning
and one function of the HD system may be to orient the cognitive of 21 active cells from rat 1 on FD9 (Fig. 2c). The mean spatial-
map in the hippocampus3. As an animal explores a novel envi- information content for these cells (1.19 0.09) did not signifi-
ronment, vestibular input and other idiothetic cues are thought cantly differ from that rats preflight data (1.37 0.10; n = 28).
to keep the HD system aligned with external landmarks long The mean spatial-information content for all 7 active place
enough for the landmarks to form stable associations with, and cells from rat 3 on the fourth day of flight (FD4) was 1.12 0.27
thereby exert control over, place and HD cells27. In normal grav- bits (Fig. 3a), which was not different from the information con-
ity, HD cells are sensitive to only the horizontal component of tent measured for this rats place cells before the flight
head direction; although changes in pitch and roll attitude are not (1.48 0.11 bits; n = 24). Another recording session followed
signaled directly by these cells8,9, calculation of head direction in immediately, and most cells maintained the same firing fields in
the horizontal plane may involve compensation for such changes. both sessions (Fig. 3b), demonstrating that the spatial tuning
Because the otolith organsnormally, a major source of infor- was stable across different exposures to the track.
mation about static pitch and roll attitudeare rendered useless The rats occasionally turned around on the track and
in zero gravity, the HD system would be deprived of this input in moved counterclockwise for short periods. Some cells demon-
compensating for changes in pitch and roll (although the semi- strated place fields when the rat was moving only counter-
circular canals would still detect angular accelerations in all three clockwise through a single location, thus demonstrating the
dimensions). Three-dimensional navigation in microgravity might direction selectivity that is seen on such tracks under normal
thus lead to inconsistent associations between HD and landmark terrestrial conditions14. Hippocampal EEG activity was also
information and a consequent inconsistency in
the hippocampal place code. In light of the dis- A B
a b
orientation frequently reported by astronauts
Fig. 1. Escher staircase track. (a) To obtain stimula-

tion of the medial forebrain bundle as a reward, the rat
moved along the track by grasping the edges of the
track and propelling itself forward. (b) Normal place
field recorded from rat 3. Red indicates maximal firing
rate (> 5 spikes per s), and blue indicates positions
sampled for which the cell never fired. Locations indi-
cated outside the black outline of the Escher staircase
were sampled when the rats head moved off the track.
Although all statistical analyses included these off-track
data, they were deleted in the remaining figures for
clarity of illustration.

a Rat 2, flight day 9 may require either a period of adaptation to microgravity or

more experience with the environment than is typically
required in normal gravity. It remains to be determined
whether the hippocampal code in microgravity can fully rep-
resent three dimensions, or whether the system adapts by devel-
oping independent two-dimensional representations for each
orthogonal surface. It is also unknown what cues drive the fir-
ing of place cells under these conditions. Although the even-
tual formation of stable fields suggests that the visual
landmarks may be primary, other contributing factors may
b Rat 2, preflight
include adaptive mechanisms that alter the efficacy of idio-
thetic cues during extended exposure to microgravity (for
instance, changes in the vestibular system or learned ability to
path integrate in three dimensions). Despite these unanswered
questions, our recordings of CNS neurons from freely moving
mammals in space demonstrate the feasibility of performing
such complex neurophysiological and behavioral experiments
c Rat 1, flight day 9 in the microgravity environment. Further experimentation in
the international space station may yield a better understand-

ing of the effects of prolonged space flight on various compo-
nents of the nervous system as well as insight into their normal
function on Earth.
ACKNOWLEDGEMENTS
We thank the crew of STS-90 (Scott Altman, Jay Buckey, Alex Dunlap, Kay
Hire, Rick Linnehan, Chiaki Mukai, Jim Pawelczyk, Rick Searfoss and Dave
Fig. 2. Representative place fields. (a) Normal place fields from rat 2 Williams), Bryan Roberts, Lisa Baer, Mike Eodice, Laurie Dubrovin, Tom
recorded on FD9. The number inside each map indicates the firing rate Howerton, Louis Ostrach, Chris Maese, Justine Grove, Ali Werner, Dave
coded by red (for instance, > 1 spike per s for the first cell). Bergner, Tom McCarthy, George Swaiss, Steve Carmen, Jim Cockrell and
(b) Preflight place fields from rat 2 as the rat ran on a rectangular track. others at NASA-Ames. We also thank Casey Stengel (who designed and built
(c) Normal place fields from rat 1 on FD9. the recording system), Krzystof Jagiello (who designed and wrote the data
acquisition software), Kathy Dillon, Shanda Roberts, Shane Smith, Vince
Pawlowski, Carol Barnes, Katalin Gothard, Veronica Fedor-Duys, Mark
Bower, Karen Reinke, Chris Duffield, Luann Snyder, Doug Wellington and
recorded during baseline and behavioral sessions in all rats on others at the University of Arizona. Additionally, we thank Bill Skaggs and
FD4 and FD9. Although the small number of subjects pre- Matt Wilson, who wrote much of the data analysis software, and science and
cluded a statistical analysis, they showed normal theta rhythm engineering support and management teams at Johnson Space Center and
during active locomotion and normal sharp waves and rip- Kennedy Space Center. Supported by grants NAG 2-949 from NASA;
ples15 during quiet inactivity in the sleeping pouch on both NS33471 and NS20331 from NIH; and N0014-98-1-0180 and
flight days. N0014-96-1-1082 from ONR.
It is interesting to note that on the first experience on the
Escher staircase on FD4, firing of place cells showed abnormal
patterns of spatial selectivity that differed between rats 1 and RECEIVED 3 SEPTEMBER 1999; ACCEPTED 1 FEBRUARY 2000
2 (J.J.K., B.L.M. and G.R.P., unpublished observations). Thus
hippocampal cells can form unique, reliable representations of
position on three orthogonal surfaces in microgravity, but they 1. Redish, A. D. Beyond the Cognitive Map (MIT Press, Cambridge,
Massachusetts, 1999).
2. Knierim, J. J., Kudrimoti, H. S. & McNaughton, B. L. J. Neurosci. 15,
a Session2 2
A. Session 16481659 (1995).
3. OKeefe, J. & Nadel, L. The Hippocampus as a Cognitive Map (Clarendon,
Oxford, 1978).
5 1 5 1 1 8 1 4. Muller, R. U. & Kubie, J. L. J. Neurosci. 7, 19511968 (1987).
5. Goodridge, J. P., Dudchenko, P. A., Worboys, K. A., Golob, E. J. & Taube, J. S.
Behav. Neurosci. 112, 749761 (1998).
6. McNaughton, B. L. et al. J. Exp. Biol. 199, 173185 (1996).
b Session
B. Session 33 7. Jeffery, K. J. & OKeefe, J. M. Exp. Brain Res. 127, 151161 (1999).
8. Taube, J. S., Muller, R. U. & Ranck, J. B. Jr. J. Neurosci. 10, 420435 (1990).
9. Taube, J. S. Prog. Neurobiol. 55, 225256 (1998).
2 1 3 1 1 3 1 10. Oman, C. M. in Proceedings of the Symposium on Vestibular Organs and
Altered Force Environment (eds. Igarashi, M. & Nute, K.) 2537 (NASA Space
Biomedical Research Institute, Houston, Texas, 1988).
Fig. 3. Stability of place fields across sessions. (a) Normal place fields from 11. Wilson, M.A. & McNaughton, B. L. Science 261, 10551058 (1993).
rat 3 on FD4 recorded on second exposure to the track. (Data from the 12. Gothard, K. M., Skaggs, W. E., Moore, K. M. & McNaughton, B. L. J. Neurosci.
first exposure were lost because of technical problems.) (b) Place fields for 16, 823854 (1996).
the same seven cells on a third run shortly after the second. Most place 13. Skaggs, W. E., McNaughton, B. L., Wilson, M. A. & Barnes, C. A.
Hippocampus 6, 149172 (1996).
cells maintained the same firing locations, although the fourth cell lost its 14. McNaughton, B. L., Barnes, C. A. & OKeefe, J. Exp. Brain Res. 52, 4149
field and a few other cells (not shown) gained a field in session 3; such (1983).
changes in responses of a minority of place cells are not uncommon. 15. Buzsaki, G. Brain Res. 398, 242252 (1986).

articles
NMDA receptor-mediated control of

protein synthesis at developing synapses
A. J. Scheetz1,4, Angus C. Nairn2 and Martha Constantine-Paton1,3
1 Department of Molecular, Cellular and Developmental Biology, Yale University, Kline Biology Tower, P.O. Box 208103, New Haven, Connecticut 06520-8103, USA
2 Laboratory of Molecular and Cellular Neuroscience, Rockefeller University, Box 296, 1230 York Avenue, New York, New York 10021, USA
3 Present address: Massachusetts Institute of Technology, Department of Biology, Building 68, Rm 380, 77 Massachusetts Ave., Cambridge, Massachusetts 02139-4307, USA
4 Present address: Department of Molecular Biophysics & Biochemistry, Yale University, 333 Cedar St., P.O. Box 208024, New Haven, Connecticut 06520-8024, USA
Correspondence should be addressed to A.J.S. (alfred.scheetz@yale.edu)
We demonstrate a rapid and complex effect of N-methyl-D-aspartate receptor (NMDAR) activation on

synaptic protein synthesis in the superior colliculi of young rats. Within minutes of receptor activation,
translation of alpha Ca2+/calmodulin dependent kinase II (CamK II) was increased, whereas total pro-
tein synthesis was reduced. NMDAR activation also increased phosphorylation of eukaryotic elongation
factor 2 (eEF2), a process known to inhibit protein translation by reducing peptide chain elongation.
Low doses of cycloheximide, which reduce elongation rate independently of eEF2 phosphorylation,
decreased overall protein synthesis but increased CaMK II synthesis. These observations suggest that
regulation of peptide elongation via eEF2 phosphorylation can link NMDAR activation to local
increases in the synthesis of specific proteins during activity-dependent synaptic change.
Activity-dependent synaptic plasticity often involves changing the isolated, functional pre- and postsynaptic elements) prepared
function of a subset of a neurons synapses in response to activa- from the superficial visual layers of the developing rat sSC8. By
tion of neurotransmitter receptors. Dendritic synthesis of specific postnatal day (P) 13, the majority of synapses are formed by reti-
proteins is implicated in many of these modifications1,2. Activity- nal ganglion cell axons on sSC neurons15. This property makes
dependent translation of mRNA into protein in dendrites may the sSC well suited for the preparation of relatively homogeneous
occur during synaptogenesis, when young neurons must selectively synaptic fractions for studies of NMDAR-linked signaling path-
stabilize subsets of inputs and eliminate others based on their pat- ways during this period of synaptogenesis.
tern of activity3. The presence of polyribosomes at young synaps- The effect of NMDAR activation on protein synthesis in sSC
es supports this idea4,5. Furthermore, activation of NMDARs is synaptoneurosomes was examined using a 30-second pulse of
associated with synaptic phosphorylation of eEF2 (ref. 6), and plas- 10 M glutamate and 50 M NMDA (referred to as NMDAR
ticity of young contacts requires activation of NMDARs. stimulation)16. NMDAR stimulation was terminated by the addi-
In the rat retinocollicular pathway, blocking NMDAR activation of AP-5 to a concentration of 120 M. For each interval, we
tion during development disrupts normal synaptogenesis7 and incubated an additional set of samples in 120 M AP-5 for 5 min-
reduces levels of CamK II (ref. 8), a protein widely implicated in utes before NMDAR stimulation and continuously thereafter to
synaptic plasticity9. To examine the role of local protein synthesis serve as the NMDAR-inactivated controls. Protein synthesis in
in this pathway, we used isolated synaptic preparations from the these AP-5 controls showed no consistent variation and remained
superficial layers of the postnatal rat superior colliculus (sSC). relatively constant throughout the incubation period. Newly syn-
We found that brief NMDAR stimulation reduced overall pro- thesized proteins were pulse labeled with 35S-methionine at stag-
tein synthesis but rapidly increased CaMK II synthesis. The fast gered intervals after NMDAR stimulation (Fig. 1a) using a
NMDAR-mediated regulation of protein synthesis was tempo- pulsechase protocol. One-minute pulses of 35S-methionine were
rally correlated with increased eEF2 phosphorylation, a modifi- followed by 10-minute chases in the excess non-radioactive
cation that inhibits protein synthesis by reducing peptide chain methionine to allow complete synthesis of proteins labeled dur-
elongation1013. We also found that inhibition of protein synthe- ing the pulse period. All data on protein synthesis following
sis elongation independent of eEF2 phosphorylation similarly NMDAR activation is presented as a percentage of synthesis mea-
increased CaMK II synthesis. Taken together, these results sug- sured in these matched AP-5 control samples.
gest that the dendritic translation of CaMK II transcripts is dif- Overall protein synthesis was maximally decreased within five
ferentially increased by transient blockade of elongation and that minutes of NMDAR stimulation (Fig. 1b). Protein synthesis lev-
phosphorylation of eEF2 represents a rapid, local and selective els subsequently increased above baseline and remained elevated
mechanism that may increase CaMK II synthesis in response from 15 minutes after stimulation until the end of the experi-
to NMDAR activation at developing synaptic contacts. ment. To monitor the effects of prolonged incubation and AP-5
exposure independent of any exposure to the NMDA-stimula-
RESULTS tion cocktail, some samples received either no treatment or
To study the effects of NMDAR activation on synaptic protein AP-5 treatment for one hour. No changes in protein synthesis
synthesis, we used synaptoneurosomes14 (fractions enriched in were observed in these samples (Fig. 1b, boxes).

articles
35S-methionine
Fig. 1. NMDAR activation dynamically regulates
a NMDAR activation AP-5 Cold methionine
protein synthesis in synaptic preparations.
Unstimulated (a) Schematic description of the pulse-labeling
AP-5 Alone procedure used to label newly synthesized pro-
tein in synaptoneurosomes. All samples (17)
3 min. were incubated for the same total time. At the
5 min. designated times, samples received NMDAR acti-
Sample sets
10 min. Time between vation (top) or AP-5 for five min before NMDAR
stimulation and
15 min. 35S-methionine activation (bottom). At the designated times after
30 min. pulse NMDAR stimulation, both pairs of samples were
45 min. pulse-labeled with 35S-methionine for one min
60 min. (black) followed by a ten-min chase with non-
radioactive methionine (gray) For 60 minutes pre-
Total duration, 75 min ceding the 35S-methionine pulse in control
experiments, samples either were untreated or
(percent of value at incubation onset)

were treated with AP-5. (b) NMDAR activation
incorporation into
b resulted in a rapid decrease in synaptic protein
35S-methionine incorporation
protein (percent AP-5 control)
synthesis followed by a prolonged increase in

overall protein synthesis. Data are means and
standard errors from five independent determina-
tions. Filled and open boxes represent the means

and standard errors for control for either pro-
longed AP-5 exposure (n = 7) or unstimulated
35S-methionine
incubation (n = 9), respectively. Data for these

samples are expressed as a percentage of 35S-
methionine incorporation into freshly prepared
samples. (c) Two-dimensional gels from samples
at different times after NMDAR activation. The
numbers in (b) correspond to the panel numbers
Interval between NMDA stimulation
and 35S-methionine pulse
(c). Panel 1, unstimulated; panel 2, 5 min NMDAR
stimulation; panel 3, 60 min NMDAR stimulation.
c Examples of proteins whose synthesis was rela-
tively unaffected by NMDAR stimulation are
marked with rectangles. Circles mark examples of
proteins whose synthesis was dramatically upreg-
ulated by NMDAR activation. These data are rep-
resentative of two independent determinations
for each time point. The isoelectric focusing range
was between pI 4 (left) and 7 (right). Molecular
mass standards are 205 kDa, 116 kDa, 70 kDa, 43
kDa, 36 kDa, 18 kDa and 7.5 kDa.
To resolve a large number of proteins, we used two-dimen- vation on CaMK II synthesis. The level of newly synthesized
sional gel electrophoresis. Under basal conditions, only a few pro- CaMK II was measured by densitometry after immunoprecip-
teins were 35S-methionine labeled (Fig. 1c, panel 1). During the itation and subsequent gel electrophoresis. We detected two radi-
period of maximal inhibition of protein synthesis, incorporation olabeled bands corresponding to molecular weights of 50 and
of 35S-methionine into these proteins was qualitatively reduced 48 kDa. In a western blot, the 50 kDa protein comigrated with
(Fig. 1c, panel 2). A larger array of proteins in samples were 35S- the CaMK II protein detected with a different monoclonal anti-
methionine labeled following 60 minutes of NMDAR stimula- body (Fig. 2b, inset). The other band may correspond to a mid-
tion. Although we cannot rule out the possibility that brain-specific CaMK II isoform 18. An analysis of variance
posttranslational modifications account for some of the observed showed that NMDAR stimulation produced statistically signifi-
changes in distribution of spots, it seems probable that these cant changes in CaMK II synthesis (p < 0.05, F6 = 15.627). In
changes were primarily due to changes in synthesis of individual contrast to overall protein synthesis levels, 3 minutes after
proteins for two reasons. First, the two-dimensional gel elec- NMDAR stimulation CaMK II synthesis was increased (p < 0.05
trophoresis analysis correlates well with our measures of NMDAR- on a planned post-hoc comparison). Following this rapid increase,
stimulated changes in overall protein synthesis. Second, we find 15 minutes after the cessation of NMDAR activation, CaMK II
no evidence of large-scale changes in protein phosphorylation synthesis was reduced to 40% of the AP-5 control level and sub-
that could account for dramatic shifts in protein migration with sequently showed a longer-latency increase that lasted for the
NMDAR stimulation16. We conclude that, even though each of duration of the experiment (Fig. 2a). Western blots of material
the 35S-methionine-labeled spots observed 60 minutes after from samples taken five minutes after stimulation of NMDARs
NMDAR stimulation does not necessarily reflect new translation, (sample set #3 in Fig. 1a) and immunoprecipitated 16.5 minutes
35S-methionine incorporation into postsynaptic protein shows a after NMDAR stimulation confirmed an increase in total synap-
complex response following brief NMDAR activation. tic CaMK II levels (Fig. 2b). Thus, the synthesis as well as the
The mRNA for CaMK II is present at high concentrations overall expression level of CaMK II in sSC synaptoneurosomes
in dendrites17. We therefore examined the effect of NMDAR acti- rapidly increased in response to brief NMDAR stimulation. Both

articles
Fig. 2. Alpha CaMK II synthesis is increased by

a b Auto Blot NMDAR activation. (a) We observed a signifi-
cant increase in 35S-methionine incorporation
into CaMK II 3 min after termination of
NMDAR stimulation (n = 5, p < 0.05, post-hoc
incorporation into
planned comparison). A second phase of

CaMK II synthesis was observed between 30
(percent AP-5 control)
and 60 min after NMDAR activation (n = 3).

CaMKII protein
The numbers correspond to autoradiographs

shown in the inset (1, unstimulated; 2, three
minutes after NMDAR stimulation; 3, five min
Optical density
after NMDAR stimulation). (b) Steady-state
35S-methionine
CaMK II protein levels were monitored from

the 5-min stimulated samples. Including times
for pulse-chase, these samples were collected
16.5 min after NMDAR stimulation. Overall
levels of CaMK II protein were significantly
No 16.5 min increased by the treatment (n = 3, p < 0.05).
stimulation after Inset shows the electrophoretic mobility of
Interval between NMDA stimulation and stimulation
35S-methionine pulse radiolabeled proteins (Auto) compared to
CaMK II protein detected with a different

c d antibody via western blot (Blot). The slower
migrating band from the autoradiograph corre-
sponds to the CaMK II protein, the faster
migrating band corresponds to the band
marked in (d) with an arrowhead.
incorporation into
(c) Immunoprecipitation using an antibody

GluR2/3 precipitated protein
directed against GluR2/3 revealed no NMDA

(percent AP-5 control)
stimulation-induced increase in 35S-methionine

incorporation within five min of stimulation.
Representative autoradiographs from either
unstimulated (above, left) or NMDAR stimu-
lated (right) samples are shown. (d) CaMK II
35S-methionine
immunoprecipitates several additional 35S-

methionine labeled proteins (CaMK II, white
arrow). These gels are representative of three
No 5 min independent determinations of samples labeled
stimulation after five min after NMDAR stimulation ended.
stimulation Molecular mass standards are as in Fig. 1.
the rapid increase in 35S-methionine-labeled CaMK II (Fig. 2a, tent qualitative difference in the composition of these complex-
three minutes) and the increase in total CaMK II protein mea- es was observed between P8 and P13 synaptoneurosomes
sured 16.5 minutes after NMDAR stimulation (Fig. 2b) are con- (Fig. 2d). Newly synthesized proteins from P8 sSC synaptoneu-
sistent with the interpretation that synaptic activation of rosomes that co-immunoprecipitated with CaMK II included
NMDARs increases CaMK II levels via local protein synthesis. two proteins with masses of approximately 25 and 30 kDa
To assess the selectivity of NMDAR-induced protein transla- (arrows on left). Immuncomplexes derived from P13 synap-
tion, we immunoprecipitated several other synaptic proteins from toneurosomes contained a different set of labeled proteins
35S-methionine-labeled synaptoneurosomes, including GluR2/3 (arrows on right), including proteins of 10, 15, 35 and 42 kDa.
(Fig. 2c), NR1, calcineurin, spinophilin and eEF2 kinase (data NMDAR activation in the developing tadpole retinotectal pro-
not shown). None of these proteins showed any NMDAR-depen- jection induces an increase in synaptic phosphorylation of eEF2
dent increase in 35S incorporation. that is localized to the immediate postsynaptic cytoplasm 6.
At P11, the retinocollicular map within the developing rat Because eEF2 phosphorylation is both mediated by a Ca2+-depen-
sSC completes its refinement19,20 and NMDA receptor currents dent kinase and is associated with decreased protein synthesis via
are downregulated en masse 21 . To determine if changes in inhibition of peptide elongation1013, it is possible that NMDAR-
NMDAR-stimulated CaMK II synthesis correlated with in vivo induced eEF2 phosphorylation in sSC synaptoneurosomes could
alterations in synaptic organization and NMDAR function, we account for the observed rapid depression of overall synaptic pro-
analyzed sSC preparations from both P8 and P13 rats. NMDAR- tein synthesis. NMDAR activation of sSC synaptoneurosomes
mediated changes in CaMK II synthesis showed the same tem- resulted in a fivefold increase in phospho-eEF2 levels within one
poral pattern at both ages. However, studies suggest that minute of stimulation. Phospho-eEF2 levels remained elevated
Ca associates with other key synaptic proteins such as at least threefold for 5 minutes and then decreased, reaching base-
actin22 and the 2B subunit of the NMDAR23,24. Non-denaturing line levels after 15 minutes (Fig. 3a). Phospho-eEF2 levels were
immunoprecipitation five minutes after NMDAR stimulation not increased by NMDAR stimulation in samples treated before
revealed that CaMK II complexes in the developing sSC con- and during NMDA stimulation with either AP-5 or EGTA (data
tain multiple newly synthesized proteins. Furthermore, a consis- not shown). Thus, the rapid NMDAR-induced depression of

articles
Fig. 3. Inhibition of synaptic translation elongation. (a) Quantitation of

a phospho-eEF2 levels from NMDA-stimulated synaptoneurosomes.
(fold increase over AP-5 control) Immediately after 30 s of NMDAR activation, phospho-eEF2 levels were
significantly increased (t-test, p < 0.05). Phospho-eEF2 levels returned to
baseline 15 min after the addition of AP-5 (n = 5, mean s.d.). The
eEF2 phosphorylation
hatched bar indicates the interval during which we observed opposing

effects of NMDA stimulation on total protein and CaMK II synthesis
(compare Figs. 1b and 2a). (b) The effect of cycloheximide on
CaMK II synthesis depended on both time and concentration. The
potentiation of CaMK II synthesis was maximal after 15 min of cyclo-
heximide treatment at a concentration of 0.5 g per ml. (c) After 15 min
of treatment with 0.5 g per ml of cycloheximide, CaMK II synthesis
was significantly increased, whereas overall protein synthesis was
decreased to 60% of untreated levels. Higher concentrations of cyclo-
heximide (5 g per ml and 50 g per ml) decreased overall protein syn-
Time after NMDA thesis to 20% and 5% of untreated levels, respectively. Treatment with
stimulus (min)
5 g per ml of cycloheximide resulted in modest increases in CaMK II
synthesis (40% over baseline), whereas 50 g per ml severely decreased
CaMK II synthesis (50% below baseline). Data are presented as means
incorporation into CaMKII
and standard deviations from three independent determinations.

b 0.5 g/ml cycloheximide

5.0 g/ml cycloheximide
protein (percent untreated)
50 g/ml cycloheximide
overall protein synthesis via eEF2 phosphorylation could account
for the observed increase in CaMK II synthesis if CaMK II trans-
lation shared this paradoxical response to mild elongation inhibi-
tion. We therefore treated synaptoneurosomes with low doses of
cycloheximide, an agent that reduces elongation rate via a mecha-
35S-methionine
nism that does not require eEF2 phosphorylation.

After 15 minutes in the presence of cycloheximide (0.5 g
Time after
per ml), overall protein synthesis decreased by 60%, whereas
cyclohexamide application (min) CaMK II synthesis increased by 150% over baseline (Fig. 3c). In
other cell-based systems, this concentration of cycloheximide
reduces elongation to rates roughly equivalent to those observed in
after cycloheximide application (percent untreated)
incorporation into protein 15 min
the presence of phospho-eEF212 (A.C.N., unpublished observa-

c tions). The effect of cycloheximide on CaMK II synthesis was
time-dependent, occurring at a latency that was longer than the
NMDA-induced effect, probably because of the time necessary for
the drug to diffuse across the synaptoneurosome membranes
(Fig. 3b). Higher cycloheximide concentrations (5 and 50 g per
ml) reduced overall protein synthesis by 80% and 95%, respective-
ly (Fig. 3c). Synthesis of CaMK II was slightly increased by addi-
tion of cycloheximide at 5 g per ml, but showed a 50% decrease
with cycloheximide at 50 g per ml (Fig. 3c).
Our interpretation of these data is that CaMK II transcripts
35S-methionine
exhibit an increase rather than a decrease in translation efficiency in

response to mild inhibition of elongation. The relative resistance of
0.5 g/ml 5.0 g/ml 50 g/ml CaMK II synthesis (50% reduction) to levels of cycloheximide that
reduced total protein synthesis in the synaptoneurosomes to 5% is
n
n
n
I
I
ei
ei
ei
KI
KI
KI
also consistent with previous studies of proteins that respond para-

ot
ot
ot
M
M
pr
pr
pr
Ca
Ca
Ca
doxically to blockade of translation elongation. In fibroblasts, doses of

al
al
al
t
t
To
To
To
cycloheximide that reduce overall protein synthesis by 50% leave all

of the remaining synthetic activity accounted for by the synthesis of
synaptic protein synthesis at developing sSC synapses correlates just 7 proteins29. Here, CaMK II synthesis may account for much
with calcium-dependent phosphorylation of eEF2. of the residual 5% of control synthetic activity we observed follow-
We next examined if a generalized reduction of polypeptide ing 50 g per ml cycloheximide. Clearly, we cannot formally rule out
elongation via eEF2 phosphorylation could also account for the alternative explanations for our 35S- methionine data. Our results,
rapid and relatively selective increase in CaMK II synthesis however, indicate significant similarities between the effects of
induced by NMDAR stimulation. Studies of mRNA competition NMDAR stimulation and low doses of cycloheximide on translation
indicate that the efficiency of translation of certain transcripts is in synaptoneurosome fractions. Both treatments decrease total pro-
increased by slowing elongation, whereas translation of the major- tein synthesis but increase CaMK II synthesis, and for both of these
ity of transcripts is decreased2528. For example, doses of cyclohex- treatments, the opposing effects on translation are coupled in time.
imide that reduce overall protein synthesis by 50% potentiate the Thus, these data suggest that eEF2 phosphorylation and the conse-
translation of several transcripts by as much as 60% in normal quent decrease in elongation rate can account for the rapid increase
fibroblasts29. Thus, it seemed possible that transiently reducing in synaptic CaMK II synthesis induced by NMDAR stimulation.

articles
DISCUSSION The implications of our observations for brain maturation

We have demonstrated that brief NMDAR activation produced remain to be explored. Both NMDAR as well as eEF2 kinase show
rapid and dynamic changes in the synthesis of numerous pro- pronounced developmental regulation, and changes in the func-
teins in synaptoneurosomes from the developing sSC. Shortly tion of either of these proteins would significantly affect the
after NMDAR activation, both the synthesis and the absolute lev- mechanism we propose. In addition, the targeting of specific
els of CaMK II protein were increased, whereas synthesis of mRNAs to dendrites may change during development. This idea
most other proteins was severely depressed. Correlated with the is supported by the developmental change in the array of newly
increase in CaMK II synthesis, we observed an increase in the synthesized proteins that interact with CaMK II. Thus, rapid
Ca2+-dependent phosphorylation of eEF2, a response known to and selective local regulation of dendritic protein synthesis may
reduce elongation rate. Additionally, low doses of cycloheximide, constitute an important and previously unrecognized aspect of
which reduce elongation rate via an eEF2-independent mecha- neuronal differentiation.
nism, produce a similar, temporally coupled decrease in total
protein and an increase in CaMK II synthesis. METHODS
Taken together, these observations suggest the following Preparation of synaptoneurosomes. Synaptoneurosomes were prepared
sequence of events. NMDAR-mediated Ca2+ influx into dendrites using a described method14 with modifications. Unless otherwise noted,
activates Ca2+-dependent eEF2 kinase, which then phosphory- postnatal day 13 rats were used. Rats were anesthetized with carbon diox-
lates eEF2. This phosphorylation might slow the local rate of pro- ide and decapitated, and the superficial, retinorecipient layers of the SC
tein translation, and elongation, rather than initiation, would were dissected as previously described8. Samples were then homogenized
consequently become the rate-limiting step in protein synthesis. in ice-cold oxygenated buffer (118 mM NaCl, 4.7 mM KCl, 1.2 mM
MgSO4, 2.5 mM CaCl2, 1.53 mM KH2PO4 212.7 mM glucose) supple-
Such a shift should favor upregulation of translation of abun-
mented with 0.002 l per ml of Complete protease inhibitor cocktail
dant but poorly initiated transcripts such as CaMK II in den- (Boehringer-Mannheim, Indianapolis, Indiana), 0.04 units per ml of
drites. Other proteins are probably selectively translated, but their human placental RNase inhibitor (Ambion, Austin, Texas) and 200 g
identities as well as the downstream implications of such signal- per ml of chloramphenicol (Sigma). All subsequent steps were carried
ing remain unexplored. We propose that this pathway constitutes out at 4C. Samples were passed through a series of nylon filters of
a mechanistic link between the activation of a neurotransmitter descending pore size. The final pass was through a MLCWP 047 Milli-
receptor and the rapid and local control of protein production. pore filter with a 10-m pore size. Samples were then centrifuged for
A number of other studies suggest that dendritic protein syn- 15 min at 1,000 g. The supernatant was discarded and the pellet resus-
thesis may be regulated locally by activation of receptors on den- pended to a final protein concentration of 0.5 mg per ml.
drites. In the hippocampus, dendritic protein synthesis is increased A protocol with characterized specificity6,16 for NMDAR stimulation
consisted of a 30-s exposure to a cocktail of 10 M glutamate and 50 M
by simultaneous stimulation of afferents and activation of acetyl-
NMDA added from a concentrated stock solution. Synaptoneurosomes
choline receptors30; isolated processes of Aplysia sensory neurons (500 l, containing 0.5 mg per ml protein) were maintained at 37C for
stimulated with serotonin respond by increasing protein synthe- at least 10 min before NMDAR stimulation. Stimulation was terminated
sis levels1 and stimulation of metabotropic glutamate receptors by addition of AP-5 from a concentrated stock to a final concentration of
in isolated synaptic contacts from neonatal rat brain increases 120 M. Five min before NMDAR stimulation, negative control samples
protein synthesis31 as well as the synthesis of proteins from exoge- received AP-5.
nous mRNA in isolated neurites32. In addition, local application of
35S-Methionine pulsechase labeling. Synaptoneurosomes were first
neurotrophins to isolated hippocampal dendrites produces an
enhancement of synaptic transmission that is blocked by inhibitors warmed to 37C for 10 min. For each time point, two samples were pre-
pared, one sample was treated with AP-5 for five min before NMDAR
of protein synthesis2. Earlier data also show that regulation of
stimulation, and another sample received NMDAR stimulation followed
translation can be exerted on specific proteins. For example, syn- by AP-5 treatment. At designated times after NMDAR stimulation,
thesis of the fragile X protein is increased by activation of 50 Ci of 35S-methionine was added to each sample. After 1 min, non-
metabotropic glutamate receptors in synaptoneurosomes33, and radioactive methionine was added to a final concentration of 200 M.
several unidentified proteins are synthesized in response to depo- Samples were incubated for an additional 10 min to allow completion
larization of isolated synaptic fractions34. Rapid increases in den- of synthesis of proteins labeled during the pulse period.
dritic CaMK II levels have also been demonstrated after synaptic
activation, both in vitro35,36 and in vivo37, suggesting that local syn- Assessment of overall protein synthesis. To measure overall protein syn-
thesis of CaMK II in dendrites could have an appreciable effect thesis, synaptoneurosome samples were treated with an equal volume of
on the synaptic concentration of CaMK II. ice-cold 10% trichloroacetic acid (TCA) for 1 h. Insoluble material was
then pelleted by centrifugation and washed 3 times in ice-cold 5% TCA
Despite these compelling observations, the mechanism(s)
(for 1 h total). Pellets were washed 3 times with ice-cold methanol and
involved in the regulation of dendritic protein synthesis remain allowed to dry at room temperature. Samples were solubilized in either
unknown. Here we provide a plausible explanation for the speed 0.1N NaOH for scintillation counting or a buffer (8.0 M urea, 0.5% SDS
and selectivity of responses in dendritic translation, particular- and 0.4% 2-mercaptoethanol) for two-dimensional isoelectric focus-
ly in relationship to NMDARs, Ca2+ and CaMK II. However, ing24. Gels were treated with Amplify (Amersham Life Sciences, Piscat-
there are probably additional mechanisms of synaptic transla- away, New Jersey) before drying and exposure. Radioactive proteins were
tional control in neurons. For example, cytoplasmic polyadeny- detected either using X-ray film (Kodak, Rochester, New York) or phos-
lation of CaMK II mRNA in rat visual cortex is increased by 30 phorimager plates (Fuji, Chicago, Illinois).
minutes of light exposure after dark rearing38, and this polyadeny-
CaMK II immunoprecipitation. CaMK II was immunoprecipitated under
lation is correlated with an increase in CaMK II protein levels at
non-denaturing conditions as previously described40. Synaptoneurosome
synapses, suggesting that, as in oocytes39, polyadenylation increas- samples were labeled as described above before immunoprecipitation. Sam-
es translational efficiency. Interestingly, the time course observed ples were pelleted and solubilized in cold immunoprecipitation buffer
with a dark rearing/light exposure protocol correlates well with (20 mM Tris-HCl, pH 7.4, 5.0 mM EDTA, 1.0% Triton X-100, 153 mM NaCl,
the long-latency increase in CaMK II we observed in isolated 20 mg per ml BSA, 0.03% sodium azide and Complete protease inhibitors;
neonatal sSC synapses. Boehringer Mannheim). Samples were mixed with either 1 l of the beta sub-

articles
unit-specific antibody CB1 (Gibco, Rockville, Maryland) or IgG and incu- 13. Ryazanov, A. G., Shestakova, E. A. & Natapov, P. G. Phosphorylation of
bated overnight at 4oC. Immuncomplexes were collected by absorption with elongation factor 2 by EF-2 kinase affects rate of translation. Nature 334, 170173
50 l of a 1:1 slurry of protein A sepharose beads (Pharmacia, Piscataway, (1988).
14. Hollingsworth, E. B. et al. Biochemical characterization of a filtered
New Jersey) in immunoprecipitation buffer. Pellets were washed 3 times in synaptoneurosome preparation from guinea pig cerebral cortex: cyclic adenosine
wash buffer (PBS, 0.25% NP-40, 0.05% sodium azide) before addition of 4 3:5-monophosphate-generating systems, receptors and enzymes. J. Neurosci. 5,
sample buffer. Samples were boiled and centrifuged. Samples (5 l) were 22402253 (1985).
removed for scintillation counting, and the remainder was loaded on 8%18% 15. Lund, R. & Lund, J. Development of synaptic patterns in the superior colliculus of
the rat. Brain Res. 42, 120 (1972).
gradient acrylamide gels. Gels were exposed as described above. Scintillation 16. Scheetz, A. J. & Constantine-Paton, M. NMDA receptor activation-responsive
counting or densitometric scanning of phosphorimager plates generated com- phosphoproteins in the developing tectum. J. Neurosci. 15, 14601469 (1996).
parable results. Western blots using a monoclonal antibody (6G9, Boehringer 17. Burgin, K. E. et al. In situ hybridization histochemistry of Ca++/calmodulin-
Mannheim) were used to detect CaMK II protein. dependent protein kinase II in developing brain. J. Neurosci 10, 17881798
(1990).
18. Brocke, L., Srinivasan, M. & Schulman, H. Developmental and regional
Measurement of phospho-eEF2 levels. At designated times after NMDAR expression of multifunctional Ca2+/calmodulin-dependent protein kinase in rat
activation, samples were centrifuged and the supernatant removed. Pel- brain. J. Neurosci 15, 67976808 (1995).
lets were suspended in 4 sample buffer and applied directly to 8%18% 19. Simon, D. K. & OLeary, D. D. M. Limited topographic specificity in the targeting
gradient acrylamide gels. Separated protein was transferred to nitrocel- and branching of mammalian retinal axons. Dev. Biol. 137, 125134 (1990).
lulose and phospho-eEF2 detected with the polyclonal antibody cc81 20. Simon, D. K. & OLeary, D. D. M. Development of topographic order in the
mammalian retinocollicular projection. J. Neurosci 12, 12121232 (1992).
(refs. 6, 41) visualized using SuperSignal chemiluminescence detection 21. Shi, J., Aamodt, S. M. & Constantine-Paton, M. Temporal correlations between
(Pierce, Rockford, Illinois). Total levels of eEF2 were estimated by strip- functional and molecular changes in NMDA receptors and GABA
ping the blot and reprobing it using an affinity-purified anti-eEF2 anti- neurotransmission in the superior colliculus. J. Neurosci. 17, 62646276 (1997).
body that does not discriminate between phospho- and dephospho-eEF2. 22. Shen, K. & Meyer, T. Dynamic control of CaMKII translocation and localization
in hippocampal neurons by NMDA receptor stimulation. Science 284, 162166
Signal was quantitated as previously described8 using NIH Image 1.60. (1999).
23. Strack, S. & Colbran, R. J. Autophosphorylation-dependent targeting of calcium/
Cycloheximide treatment. A concentrated stock of cycloheximide (50 mg calmodulin- dependent protein kinase II by the NR2B subunit of the N-methyl-
per ml) was prepared in 0.2% DMSO. Synaptoneurosomes prepared from the D- aspartate receptor. J. Biol. Chem 273, 2068920692 (1998).
sSC of P13 rat pups were treated with cycloheximide at final concentrations 24. Leonard, A. S., Lim, I. A., Hemsworth, D. E., Horne, M. C. & Hell, J. W.
Calcium/calmodulin-dependent protein kinase II is associated with the
of 0.5 g per ml, 5.0 g per ml and 50.0 g per ml. At designated times, sam- N- methyl-D-aspartate receptor Proc. Natl. Acad. Sci. USA 96, 32393244 (1999).
ples were pulse-labeled as above for one min and chased with non-radioac- 25. Brendler, T., Godefroy-Colburn, T., Carhill, R. D. & Thach, R. E. The role of
tive methionine for ten min. Samples were then subjected to either TCA mRNA competition in regulating translation II. Development of a quantitative in
precipitation or CaMK II immunoprecipitation. Data were analyzed as above. vitro assay. J. Biol. Chem. 256, 1174711754 (1981).
26. Brendler, T., Godefroy-Colburn, T., Yu, S. & Thach, R. E. The role of mRNA
competition in regulating translation III. Comparison of in vitro and in vivo
results. J. Biol. Chem. 256, 1175511761 (1981).
ACKNOWLEDGEMENTS 27. Godefroy-Colburn, T. & Thach, R. E. The role of mRNA competition in
This work was supported by U.S. Public Health Service Grants EY 06039 to regulating translation IV. Kinetic model. J. Biol. Chem. 256, 1176211773
M.C.P and GM 50402 to A.C.N. (1981).
28. Walden, W. E., Godefroy-Colburn, T. & Thach, R. E. The role of mRNA
competition in regulating translation I. Demonstration of competition in vivo.
J. Biol. Chem. 256, 1173911746 (1981).
RECEIVED 11 JANUARY; ACCEPTED 24 JANUARY 2000 29. Walden, W. E. & Thach, R. E. Translational control of gene expression in a normal
fibroblast. Characterization of a subclass of mRNAs with unusual kinetic
properties. Biochemistry 25, 20332041 (1986).
1. Martin, K. C. et al. Synapse-specific, long-term facilitation of Aplysia sensory to 30. Feig, S. & Lipton, P. Pairing the cholinergic agonist carbacol with patterned
motor synapses: a function for local protein synthesis in memory storage. Cell 91, schaffer collateral stimulation initiates protein synthesis in hippocampal
927938 (1997). pyramidal cell dendrites via a muscarinic, NMDA-dependent mechanism.
2. Kang, H. & Schuman, E. M. A requirement for local protein synthesis in J. Neurosci. 13, 10101021 (1993).
neurotrophin-induced hippocampal plasticity. Science 273, 14021406 (1996). 31. Weiler, I. J. & Greenough, W. T. Metabotropic glutamate receptors trigger
3. Constantine-Paton, M., Cline, H. T. & Debski, E. A. Patterned activity, synaptic postsynaptic protein synthesis. Proc. Natl. Acad. Sci. USA 90, 71687171 (1993).
convergence and the NMDA receptor in developing visual pathways. Annu. Rev. 32. Crino, P. B. & Eberwine, J. Molecular characterization of the dendritic growth
Neurosci. 13, 129154 (1990). cone: regulated mRNA transport and local protein synthesis. Neuron 17,
4. Steward, O. & Falk, P. M. Protein-synthetic machinery at postsynaptic sites 11731187 (1996).
during synaptogenesis: a quantitative study of the association between 33. Weiler, I. J. et al. Fragile X mental retardation protein is translated near synapses
polyribosomes and developing synapses. J. Neurosci. 6, 412423 (1986). in response to neurotransmitter activation. Proc. Natl. Acad. Sci. USA 94,
5. Steward, O. & Falk, P. M. Selective localization of polyribosomes beneath 53955400 (1997).
developing synapses: a quantitative analysis of the relationships between 34. Leski, M. L. & Steward, O. Protein synthesis within dendrites: ionic and
polyribosomes and developing synapses in the hippocampus and dentate gyrus. neurotransmitter modulation of synthesis of particular polypeptides
J. Comp. Neurol. 314, 545557 (1991). characterized by gel electrophoresis. Neurochem. Res. 21, 681690 (1996).
6. Scheetz, A. J., Nairn, A. C. & Constantine-Paton, M. N-methyl-D-aspartate 35. Ouyang, Y., Rosenstein, A., Kreiman, G., Schuman, E. M. & Kennedy, M. B.
receptor activation and visual activity induce elongation factor-2 Tetanic stimulation leads to increased accumulation of Ca2+/calmodulin-
phosphorylation in amphibia tecta: a role for N-methyl-D-aspartate receptors in dependent protein kinase II via dendritic protein synthesis in hippocampal
controlling protein synthesis. Proc. Natl. Acad. Sci. USA 94, 1477014775 (1997). neurons. J. Neurosci. 19, 78237833 (1999).
7. Simon, D. K., Prusky, G. T., OLeary, D. D. M. & Constantine-Paton, M. NMDA 36. Ouyang, Y., Kantor, D., Harris, K. M., Schuman, E. M. & Kennedy, M. B.
receptor antagonists disrupt the formation of a mammalian neural map Proc. Visualization of the distribution of autophosphorylated calcium/calmodulin-
Natl. Acad. Sci. USA 89, 1059310597 (1992). dependent protein kinase II after tetanic stimulation in the CA1 area of the
8. Scheetz, A. J., Prusky, G. T. & Constantine-Paton, M. Chronic NMDA receptor hippocampus. J. Neurosci. 17, 54165427 (1997).
blockade during retinotopic map formation decreases the CaM kinase II 37. Steward, O. & Halpain, S. Lamina-specific synaptic activation causes domain-
differentiation in rat superior colliculus. Eur. J. Neurosci. 8, 13221328 (1996). specific alterations in dendritic immunostaining for MAP2 and CAM kinase II.
9. Kelly, P. T. Calmodulin-dependent protein kinase II. Multifunctional roles in J. Neurosci. 19, 78347845 (1999).
neuronal differentiation and synaptic plasticity. Mol. Neurobiol. 5, 153177 38. Wu, L. et al. CEPB-mediated cytoplasmic polyadenylation and the regulation of
(1991). experience-dependent translation of -CaMKII mRNA at synapses. Neuron 21,
10. Nairn, A. C. & Palfrey, H. C. Identification of the major Mr 100,000 substrate for 11291139 (1998).
calmodulin-dependent protein kinase III in mammalian cells as elongation 39. McGrew, L. L. & Richter, J. D. Translational control by cytoplasmic
factor-2. J. Biol. Chem. 262, 1729917303 (1987). polyadenylation during Xenopus oocyte maturation: characterization of cis and
11. Nairn, A. C. & Palfrey, H. C. in Translational Control (eds. Hershey, J. W. B., trans elements and regulation by cyclin/MPF. EMBO J. 9, 37433751 (1990).
Mathews, M. B. & Sonenberg, N.) 295318 (Cold Spring Harbor Press, 40. Baitinger, C., Alderton, J., Poenie, M., Schulman, H. & Steinhardt, R. A.
Plainview, New York, 1996). Multifunctional Ca2+/calmodulin-dependent protein kinase is necessary for
12. Redpath, N. T. & Proud, C. G. The tumour promoter okadaic acid inhibits nuclear envelope breakdown. J. Cell Biol. 111, 17631773 (1990).
reticulocyte-lysate protein synthesis by increasing the net phosphorylation of 41. Marin, P. et al. Glutamate-dependent phosphorylation of elongation factor-2 and
elongation factor 2. Biochem. J. 262, 6975 (1989). inhibition of protein synthesis in neurons. J. Neurosci. 17, 34453454 (1997).

articles
Rho GTPases regulate distinct

aspects of dendritic arbor growth
in Xenopus central neurons in vivo
Zheng Li1,2, Linda Van Aelst1 and Hollis T. Cline1,2
1 Cold Spring Harbor Laboratory, Beckman Bldg., 1 Bungtown Rd., Cold Spring Harbor, New York 11724, USA
2 Department of Neurobiology and Behavior, SUNY Stony Brook, New York 11794, USA
Correspondence should be addressed to H.T.C. (cline@cshl.org)
The development and structural plasticity of dendritic arbors are governed by several factors, includ-
ing synaptic activity, neurotrophins and other growth-regulating molecules. The signal transduction
pathways leading to dendritic structural changes are unknown, but likely include cytoskeleton regu-
latory components. To test whether GTPases regulate dendritic arbor development, we collected
time-lapse images of single optic tectal neurons in albino Xenopus tadpoles expressing dominant
negative or constitutively active forms of Rac, Cdc42 or RhoA. Analysis of images collected at two-
hour intervals over eight hours indicated that enhanced Rac activity selectively increased branch
additions and retractions, as did Cdc42 to a lesser extent. Activation of endogenous RhoA decreased
branch extension without affecting branch additions and retractions, whereas dominant-negative
RhoA increased branch extension. Finally, we provide data suggesting that RhoA mediates the
promotion of normal dendritic arbor development by NMDA receptor activation.
The structure of the dendritic arbor critically determines what regulating the cytoskeleton. Investigations of their function have
synaptic inputs a neuron receives and how they are integrated1. been aided by mutations that result in a constitutively active,
For instance, a neuron within the visual system whose dendritic GTP-bound form or a dominant-negative, GDP-bound form.
arbor has a large tangential spread can receive inputs from more Studies of neurite outgrowth from cultured neurons have given
visual afferents, leading to a larger receptive field over which us the first insights into the role of the Rho family of GTPases in
information is processed. Similarly, neurons that extend their regulating neuronal process growth911. Because neurites in cell
dendritic arbors into superficial laminae can receive inputs and culture often fail to take on the characteristics of dendrites and
process information from afferents in those laminae2. Conse- axons, few such studies have been able to distinguish effects of
quently, factors that regulate development and plasticity of the Rho GTPases on axonal and dendritic outgrowth. Rac, Cdc42
dendritic arbor control both the neurons structure and function and RhoA influence dendrite number in dissociated cortical neu-
and may ultimately affect circuit properties. rons12. Given that neuronal arbor elaboration in vivo is governed
The molecular mechanisms that underlie development of the by activity-dependent and activity-independent factors13, which
dendritic arbor are not clear yet. In vivo time-lapse imaging per- may operate through GTPases14, we wanted to determine whether
mits direct observation of dendritic arbor development. Time- Rho GTPases regulate dendritic arbor formation in the live ani-
lapse images of optic tectal neurons collected at intervals ranging mal with the normal pattern of synaptic inputs and local
from minutes to days in living Xenopus tadpoles show that den- environment. Our previous studies showed that NMDA recep-
dritic branches are very dynamic during arbor formation36. The tor activity is required for the initial phase of dendritic arbor
dynamic processes include addition of new branches, retraction growth in tectal neurons5,6. A potential link between synaptic
of branches, and selective extension or shortening of existing activity and the Rho GTPases remains to be defined.
branches. These events can be observed and quantified by col- As opposed to cultured cells, studies in intact animals pro-
lecting repeated images over several hours36. The data suggest vide the opportunity to investigate dendritic and axonal devel-
that the net growth of dendritic arbor occurs as a result of sev- opment in their normal complex environment. In vivo
eral distinct events: emergence of a new branch, selective main- experiments in transgenic flies, worms, Xenopus and mice all
tenance of the new branch, and extension of branch length. Each point to a crucial role of Rac and Cdc42 in axonal growth and
of these events may be individually regulated. Furthermore, it is target recognition1519. The roles of the Rho GTPases in regulat-
very likely that machinery controlling the actin cytoskeleton is ing dendritic arbor structural plasticity in vivo are relatively unex-
involved, because cytoskeleton reorganization accompanies struc- plored20. Although the dendrites of Purkinje cells in transgenic
tural changes in cells. mice expressing constitutively active Rac in the cerebellum branch
Members of the Rho family of small GTPases, Rac, Cdc42 and normally, dendritic spines are reduced in size and increased in
RhoA, are required components of signal transduction pathways number16. In Xenopus, retinal ganglion cell dendritic arbor elab-
through which extracellular signals cause morphological changes oration is inhibited by expression of constitutively active RhoA
in various cell types79. Rho GTPases mediate these changes by and constitutively active Cdc42, but promoted by constitutively

articles
Fig. 1. Expression of Rho GTPases and their effect on

the actin cytoskeleton. (a) Myc immunostaining shows a b
the distribution of infected, myc-tagged, GTPase-
expressing neurons in a horizontal section through the
midbrain of a stage 47 tadpole. Most cells near the tectal
ventricle (those imaged in later experiments) are
infected. (b) Propidium iodide staining of a section neigh-
boring the one shown in (a). Optic tectal cells (TC) are
stained and appear gray. The neuropil (NP) region is
unstained. (c) Phalloidin staining (left), GTPase expres-
sion (middle) in infected neurons, detected by myc
immunoreactivity for constitutively active Rac (RacV12),
dominant-negative Rac (RacN17) and constitutively
active Cdc42 (Cdc42V12) or EGFP for dominant-nega- c
tive Cdc42 (Cdc42N17), and the overlay of the two
images (right) in sections from animals infected at low
titer. Constitutively active forms of Rac and Cdc42
specifically increase actin polymerization in GTPase-
expressing neurons. Scale bars, 50 m.
active Rac18. In transgenic Drosophila, constitu-

tively active and dominant-negative Rac both block
axonal growth of sensory neurons, without affect-
ing dendrites, whereas constitutively active Cdc42
inhibits both axonal and dendritic growth15. As in
the cell culture studies, some of the phenotypic
outcomes resulting from expression of Rac and
Cdc42 vary between the different studies, likely
because of differences in neuron type15,17. Never-
theless, these studies provide growing evidence that
Rac and Cdc42 can regulate dendritic arbor mor-
phology. These studies do not indicate the poten-
tial function of Rho GTPases in dendritic arbor
development, nor do they reveal how activity of
different GTPases might cooperate during dendrite
elaboration.
To address these open issues, we investigated
whether Rho GTPases regulate branch dynamics
and dendritic arbor growth in optic tectal neurons in live Xeno- myc-tagged GTPases showed that after injection of high-titer virus,
pus tadpoles, by collecting in vivo time-lapse images of single DiI- most tectal cells near the brain ventricle were infected (Fig. 1a).
labeled tectal neurons. We used vaccinia virus-mediated gene Neurons in this region were DiI labeled for imaging experiments
transfer to express constitutively active and dominant-negative described below. Expression of foreign protein was detected start-
forms of RhoA, Rac and Cdc42 in tectal cells. We found that the ing six hours after injection and was maintained throughout the
three Rho GTPases have distinct effects in dendritic arbor devel- experiment. Retinal ganglion cells are not infected when virus is
opment: Rac and Cdc42 regulate dynamic branch additions and injected into the ventricle, so retinal axons do not express foreign
retractions, whereas RhoA regulates elongation of existing protein (Fig. 1), consistent with our previous data21.
branches. These results support the idea that dendritic arbor To test whether virally expressed Rho GTPases regulate the
growth occurs through a multi-step process in which Rac and actin cytoskeleton in infected tectal neurons, we studied the dis-
Cdc42 regulate the addition of short branches. RhoA regulates tribution of polymerized actin using double labeling with phal-
the selective extension of a subset of the added branches. loidin to visualize polymerized actin and either anti-myc antibody
or EGFP to visualize Rho GTPase-expressing cells. The animals
RESULTS were infected with low-titer (106 pfu) virus to infect only a few
Effects of Rho family GTPases on the actin cytoskeleton tectal cells per animal, so that we could detect the actin filaments
Tectal neurons were infected by ventricular injection of recombi- in individual GTPase-expressing neurons. In both uninfected ani-
nant vaccinia viruses expressing constitutively active and domi- mals and those infected with vaccinia virus expressing -gal, fila-
nant-negative forms of Rho family GTPases. The constitutively mentous actin (F-actin) was enriched in the neuropil region, with
active mutants we used were RacV12 and Cdc42V12, and the dom- relatively little F-actin found in the cell body region of the tectum
inant-negative mutants were RacN17, Cdc42N17 and RhoN19. (Fig. 1b and c). Actin polymerization was increased dramatically
The GTPases were also tagged with myc or enhanced green fluo- in neurons expressing constitutively active Rac and constitutively
rescent protein (EGFP; Methods). The myc-tagged GTPases were active Cdc42, as indicated by strong phalloidin staining in infect-
expressed as IRES (internal ribosome entry site)-EGFP constructs ed cells (Fig. 1c). In neurons expressing dominant-negative Rac,
that allowed us to identify infected regions of tectum for DiI label- dominant-negative Cdc42 or dominant-negative RhoA, the
ing. Cryostat sections of infected animals immunostained for the F-actin staining pattern was similar to that observed in control

articles
Fig. 2. RhoA activity blocks dendritic arbor elaboration. a

(a) Drawings of 2 representative tectal cells from control and
LPA-treated animals imaged over 12 h in vivo. (b) Drawings of 2
representative tectal cells from control and dominant-negative
RhoA vaccinia virus-infected animals imaged over 20 h.
(c) Change in total dendritic branch length (TDBL) over 20 h
for neurons imaged from animals infected with various viruses.
(d) Change in TDBL over 12 h for neurons from controls, LPA-
treated animals and LPA-treated animals expressing dominant-
negative RhoA. For each condition, 1847 cells were analyzed.
Scale bar, 50 m, applies to (a) and (b). *p < 0.05; **p < 0.002.
b
cells. F-actin staining was also indistinguishable in

lysophosphatidic acid (LPA)-treated animals and con-
trol animals, consistent with reports that neuronal cells
do not form stress fibers after LPA exposure22. These
experiments confirm viral expression of the GTPases.
They further show that constitutively active Rac and con-

stitutively active Cdc42 caused changes in actin cytoskele-
ton comparable to those reported previously7,9. One
possible scenario is that these actin cytoskeleton alter-
ations affect dendritic arbor growth. c d
In previous experiments, we collected time-lapse
TDBL growth (m per 12 h)

TDBL growth (m per 20 h)
images of tectal cell dendrites at three-minute intervals

over about an hour, at thirty-minute intervals over two
hours, at two-hour intervals over eight hours and at daily
intervals over five days36. These experiments demon-
strated that short branches are continually added and
retracted from the arbor and that arbor growth results
from selective stabilization of a small fraction of the
newly added branches and their subsequent elongation.
To assess the effects of Rho GTPases on branch addi-
tions and retractions (which we call branch dynamics) as
well as branch elongation, we chose an imaging protocol
that permits quantitation of parameters relating to
branch dynamics and net dendritic arbor growth. One day after cinia virus, 286.0 33.2 m per 20 h). We used LPA (10 M in
injecting virus, we labeled single tectal cells with DiI and col- rearing solution) to activate endogenous RhoA because we were
lected an initial image of the labeled tectal neurons two hours not able to generate a virus expressing constitutively active RhoA,
after DiI labeling. The following day, 12 hours after collecting the most likely because the protein is toxic. LPA has been shown to
first image, we found the same single cell and imaged it at 2-hour activate RhoA specifically in neuronal cell lines11. Activation of
intervals over the 8 hours from the 12-hour to 20-hour time RhoA by LPA treatment significantly reduced dendritic arbor
points. To determine the effect of GTPase activity on overall den- growth rate from the control value of 160.8 19.0 m per 12 h to
dritic arbor growth, we compared total dendritic branch length at 67.4 20.9 m per 12 h (p < 0.002; Fig. 2). To assure that the
the 0-hour and 20-hour timepoints. An increase in growth rate effect of LPA on growth rate is due to activating RhoA, we showed
indicates an increase in branch elongation. To determine the effect that the growth-inhibiting effect of LPA could be counteracted
of GTPase activity on dendritic arbor dynamics, we compared by expression of dominant-negative RhoA. Cells from animals
branch additions and retractions in sequential two-hour time infected with dominant-negative RhoA vaccinia virus and treat-
points over eight hours. This imaging interval permits us to fol- ed with LPA had a growth rate of 151.6 22.1 m per 12 h, com-
low the fate of every branch we image within the arbor over parable to control neurons (p = 0.85; Fig. 2). These data indicate
time35. This protocol underestimates the rates of branch addi- that LPA is acting through RhoA to control branch elongation
tions and retractions, because branches are continuously added in tectal neurons. In contrast to RhoA, Rac and Cdc42 did not
and retracted within the two-hour intervals. Nevertheless, it does alter dendritic growth rates (Fig. 2). Although RhoA is essential
provide a relative measure of arbor dynamics between control for dendritic branch extension, data shown below indicate that
and experimental conditions in the same neurons where we can neither LPA nor dominant-negative RhoA changed branch num-
also quantify arbor growth due to branch elongation (see Meth- ber and branch dynamics (Fig. 3). Together, these data suggest
ods for details). that RhoA regulates the growth of the dendritic arbor by affect-
ing the elongation of pre-existing branches.
Activated RhoA inhibits dendritic branch extension
Expression of dominant-negative RhoA significantly increased Rac regulates arbor dynamics
arbor growth rate compared to control cells (p < 0.05, Fig. 2; To study the effects of Rho family GTPases on dendritic arbor
growth rate in control cells, 213.7 18.8 m per 20 h; in neu- dynamics, we imaged the same neuron every two hours for a total
rons from animals infected with dominant-negative RhoA vac- of eight hours. During dendritic arbor formation, the structure

articles
a (60.9 7.8, constitutively active Rac; 38.5 3.0, control,

p < 0.002; Fig. 3). Constitutively active Rac cells also retracted
significantly more dendrites than controls (49.9 7.1, constitu-
tively active Rac; 29.4 2.5, control; p < 0.001). Because consti-
tutively active Rac enhanced both additions and retractions, the
final number of branch tips was not significantly different from
control cells. In contrast, dominant-negative Rac did not signif-
icantly alter rates of branch additions. Expression of dominant-
negative Rac, however, did cause a significant increase in branch
retractions compared to controls, but to a lesser extent than con-
stitutively active Rac (39.8 3.7, p < 0.05). Experiments were
also done using constitutively active Cdc42, dominant-negative
Cdc42, dominant-negative RhoA and LPA. Comparable analy-
sis indicated that neither Cdc42 nor RhoA affected branch
dynamics (Fig. 3).
This analysis demonstrated that Rac activity affects rates of
branch additions and retractions and suggested that the branch-
es may be more transient. To test directly whether Rac, Cdc42 or
RhoA alters the fate of dendritic branches, we designed an analy-

sis to identify changes in the proportion of transient branches in
an arbor. We divided all dendritic branches into four categories:
stable branches, lost branches, new branches and transient
b c branches (Fig. 4a). Stable branches are branches that are present
throughout the imaging period. Lost branches are those that are
Total branch retractions
Total branch additions
present at the first image, but retracted over the eight-hour peri-
od. Any branches that were added after the first image and were
still there at the last time point were categorized as new branch-
es. The branches that were added after the first image and then
retracted before the last image were categorized as transient
branches. This analysis showed that among all the branches ever
present during imaging, 34.8 1.2% of them were transient in
control cells. This number was significantly increased by the
expression of constitutively active Rac, and to a lesser extent by
constitutively active Cdc42 (Table 1; Fig. 4). Notably, dominant-
Fig. 3. Rac activity promotes dendritic branch dynamics. (a) Drawings negative Rac triggered an increase in transient branches (Fig. 4).
of 2 representative tectal cells from control, constitutively active Rac In addition, dominant-negative Rac significantly decreased the
virus-infected and dominant-negative Rac virus-infected animals imaged relative number of new branches, that is, those that were added to
over 20 h. Note the rapid changes in fine branch tips in cells expressing the arbor and maintained to the end of the observation period.
constitutively active Rac. (b, c) Total branch additions (b) and retrac- When we evaluated the fraction of stable branches, we noticed
tions (c) for neurons imaged every two hours over eight hours. For each that only constitutively active Rac significantly affected this cat-
condition, 1847 cells were analyzed. Scale bar, 50 m. *p < 0.05; egory, causing a 50% reduction of stable branches. No effects on
**p < 0.002; ***p < 0.001.
arbor dynamics were observed when we expressed dominant-
negative Cdc42 or dominant-negative RhoA or added LPA. Taken
together, the analysis of arbor dynamics indicated that Rac and
to a lesser extent Cdc42, but not RhoA are crucial in regulating
of the arbor is very dynamic, characterized by continuous branch arbor dynamics.
additions and retractions (Fig. 3). In animals infected with con- To test whether constitutively active Rac increases arbor dynam-
stitutively active Rac or dominant-negative Rac vaccinia virus, ics in the time frame of minutes, we imaged single labeled neurons
the dendritic arbor was more dynamic than in control neurons. every 3 minutes over periods up to 30 minutes. Six cells were imaged
As a result of these dynamics, the arbor structure was quite dif- from animals infected with either EGFP vaccinia virus (control) or
ferent from one time point to the next. Images of neurons infect- constitutively active Rac vaccinia virus. The rapid branch dynamics
ed with vaccinia virus expressing constitutively active Cdc42, are most easily recognized in a time-lapse movie of the cells (see
dominant-negative Cdc42 and dominant-negative RhoA or neu- http://neurosci.nature.com/web_specials/). The movie also demon-
rons exposed to LPA did not show obvious changes in arbor strates that dendritic arbors of neurons from animals infected with
dynamics during the eight-hour imaging period (see Fig. 3). constitutively active Rac vaccinia virus appear more dynamic.
To quantify whether an increase or decrease in RhoA, Rac or Branches within an arbor can have a variety of lifetimes, ranging
Cdc42 activity had significant effects on arbor dynamics, we from several minutes to several days3,4,23,24. The lifetimes of branch-
counted the numbers of newly added branches and retracted es can be shifted to longer or shorter times during development4,
branches that were imaged during the eight-hour period. All by increased CaMKII activity3 or by blocking NMDA receptor activ-
groups had a comparable number of branch tips and total den- ity6. Quantitation of the lifetimes of each branch in the arbors show
dritic branch length at the first imaging observation. Neurons that 67% of branches in constitutively active Rac neurons have life-
from animals infected with constitutively active Rac vaccinia virus times less than 9 minutes, whereas only 31% of branches in con-
added significantly more dendritic branches than control cells trol neurons have similarly short lifetimes. The shift toward shorter

articles
branch lifetimes observed with expression of constitutively active a

Rac is consistent with the increase in the fraction of transient
branches observed with the two-hour imaging protocol (Fig. 4). Stable branch
These data are consistent with the hypothesis that Rac affects Lost branch
cytoskeletal stability in neurons. Furthermore, they suggest that Rac New branch
affects arbor growth by affecting the rates of branch additions and Transient branch
retractions.
GTPases affect arbor complexity

The constitutively active Rac neurons shown in Fig. 3 appear to
b
have more densely branched dendritic arbors than control neu-
rons. We used Sholl analysis, in which the number of branches
crossing concentric rings around the cell body are counted (see
Methods) to analyze dendritic arbor complexity. Expression of
Stable branch
constitutively active Rac significantly altered dendritic arbor com-
plexity compared to controls (Fig. 5). The distal dendritic arbors of New branch
neurons from animals infected with constitutively active Rac vac- Lost branch
cinia virus were more complex than in controls, whereas proximal
Transient branch
dendritic arbors were significantly less complex than in controls.
Fig. 4. Rac and Cdc42 increase transient branches. (a) Schematic drawing
Sholl analysis also confirmed the impression from the drawings in
of the types of branch categories analyzed (see text for details). (b) Plots
Fig. 3 that decreased RhoA activity enhanced arbor complexity, of the fraction of stable, transient, new and lost branches in arbors from
whereas LPA treatment caused simpler dendritic arbors. each of the designated treatments. *p < 0.05; **p < 0.01; ***p < 0.001.
RhoA is involved in NMDAR-mediated arbor growth

We have previously shown that NMDA receptor activity is
required for normal dendritic arbor elaboration in tectal neu- (p = 0.55). These data support a model in which NMDA recep-
rons5,6. Normal arbor elaboration is prevented by exposing tectal tor-mediated control of dendritic arbor elaboration operates
neurons to the NMDA receptor antagonist 3-amino-phospho- through a pathway that decreases RhoA activity.
novaleric acid (APV) early during dendritic arbor development5,
when the glutamatergic synaptic transmission from the retina is DISCUSSION
mediated predominantly by the NMDA receptor25. By contrast, The structure of the neuronal dendritic arbor determines the
the stability of dendritic arbors is less affected by APV in more inputs received by the neuron as well as its integrative properties1,2.
mature neurons, when their retinotectal glutamatergic trans- Children with mental retardation have severely reduced dendritic
mission is mediated predominantly by the AMPA-type glutamate arbors, demonstrating a fundamental connection between neu-
receptor25. Here we report that increasing RhoA GTPase activity ronal structure and cognitive ability26. Consequently, mechanisms
inhibits dendritic arbor development,
whereas decreasing RhoA activity Table 1. Relative distribution of branches in each of the categories shown in Fig. 4.
enhances dendritic growth. These data Stable branch New branch Lost branch Transient branch
suggest that signals that decrease (Percent total) (Percent total) (Percent total) (Percent total)
endogenous RhoA activity promote Control 7.3 0.8 37.6 1.7 20.3 1.4 34.8 1.3
dendritic branch elongation. An
(n = 47)
intriguing model is that NMDA recep-
tor activity may promote dendritic
arbor development by decreasing RacV12 3.7 1.1 34.1 1.7 19.5 1.8 42.7 2.3
endogenous RhoA activity. To test this (n = 20) p < 0.013 p < 0.002
possibility, we determined whether
expression of dominant-negative RhoA RacN17 5.7 1.0 22.8 7.8 21.8 2.7 49.7 5.5
could prevent the reduced dendritic (n = 18) p < 0.008 p < 0.001
growth rate observed as a result of
blocking the NMDA receptor. As Cdc42V12 7.4 1.4 28.2 6.5 22.0 1.9 42.4 4.2
reported previously5, exposing animals
(n = 22) p < 0.03
to 100 M APV in their rearing solu-
tion inhibits dendritic arbor growth
compared to control neurons (APV, Cdc42N17 8.4 1.3 40.9 2.3 16.5 1.9 34.2 2.1
160 15 m per 20 h; control, (n = 21)
214 17 m per 20 h; p < 0.05; Fig. 6).
Cells from animals infected with dom- LPA 9.3 0.9 40.6 2.2 18.5 1.3 31.6 2.2
inant-negative RhoA and treated with (n = 27)
APV grow at 256 32.5 m per 20 h,
which is significantly faster than in RhoN19 10.4 1.4 40.1 2.4 16.4 1.4 33.2 2.2
APV-treated neurons (p < 0.005), and (n = 20)
comparable to neurons infected with
dominant-negative RhoA vaccinia virus Values for numbers of cells analyzed also apply to data in Figs. 2, 3 and 5.

articles
a b suggests that endogenous RhoA activity

in developing tectal neurons is high and
that dendritic arbor growth is actively
promoted under conditions that inhibit
RhoA activity. Because neither LPA nor
dominant-negative RhoA affects rates of
branch additions or retractions, these
experiments indicate that RhoA selective-
ly affects extension or retraction of exist-
ing branches.
We previously showed that blocking
the NMDA type of glutamate receptor
c d decreases the rate of branch additions and
decreases the extension of existing
branches during early stages of dendritic
arbor formation5,6. Later, as the neurons
mature, their dendritic arbor structure
becomes more stable. These mature neu-
rons express calcium/calmodulin-depen-

dent protein kinase (CaMKII), and
enhanced CaMKII activity decreases rates
of dendritic branch additions and retrac-
tions3. One intriguing scenario is that glu-
tamatergic synaptic activity may promote
dendritic arbor elaboration by decreas-
Fig. 5. Rac and RhoA affect dendritic arbor complexity. Sholl analysis of constitutively active Rac (a), ing endogenous RhoA activity in den-
dominant-negative Rac (b), dominant-negative Rho (c) and LPA-treated (d) neurons compared to drites of immature tectal cells. Indeed, we
controls. Expression of constitutively active Rac significantly decreases the number of dendritic find that the decreased arbor growth rate
branches close to the cell body, but significantly increases arbor complexity distal to the cell body. observed when NMDA receptors are
Decreasing RhoA activity increases arbor complexity, whereas increasing RhoA activity decreases blocked is counteracted by expression of
complexity. *p < 0.05; **p < 0.001; ***p < 0.0001. dominant-negative RhoA. These data
suggest a mechanism in which RhoA
activity and dendritic branch extension
may be controlled locally by synaptic
that influence the development of neuronal structure are likely to inputs. Furthermore, as neuronal structure matures and becomes
significantly affect brain function. Mutations in a Rho-GTPase more stable, glutamatergic synaptic inputs and CaMKII activity
activating protein are found in patients with X-linked mental retar- may operate through Rac and Cdc42 to control structural plas-
dation27, suggesting that GTPase signaling is required for the devel- ticity. The demonstration of a Ras GTPase-activating protein
opment of normal brain function. that is regulated by NMDA receptor and CaMKII activity28,29
Here we have shown that different members of the Rho fam- combined with evidence of crosstalk between Ras and Rac in
ily of GTPases regulate different aspects of dendritic arbor elab- controlling cell morphology30 further suggest that such a regu-
oration in the intact animal. By taking sequential images of latory pathway may exist.
individual dendritic arbors with the confocal microscope, we
found that Rac, and to a lesser extent Cdc42, regulate branch Rac regulates dendritic arbor dynamics
dynamics. RhoA specifically controls branch elongation, with- Data on the function of Rac in regulating dendritic arbor struc-
out directly affecting branch dynamics. These data suggest that ture in the intact animal is limited16,18,31. In Drosophila and mice,
dendritic arbor elaboration occurs as a result of Rac-mediated Rac does not affect the overall development of dendritic mor-
branch dynamics followed by RhoA-mediated branch extension. phology15,16, whereas in Xenopus retinal cells, Rac does affect den-
Together the activities of the GTPases contribute to the net dritic arbor outgrowth18. Our data show that Rac is involved in
growth of dendritic arbors in the intact animal. Because many the formation and turnover of dendrites, but not dendritic branch
tectal neurons are infected and express ectopic GTPases in these extension. At a single time point, the total number and length of
experiments, an interesting possibility, which we cannot exclude, dendrites in animals expressing constitutively active or domi-
is that the observed effects of the GTPases are in part due to nant-negative Rac are comparable to the control values, but time-
GTPase expression in cells other than the ones imaged. However, lapse imaging clearly demonstrates that constitutively active Rac
Rho affects dendritic arbor development in a cell-autonomous increases branch additions and retractions. The final branch
fashion in Drosophila20. numbers and total branch length are the same as control neu-
rons because enhanced rates of branch additions are balanced by
RhoA regulates dendritic branch extension increased branch retractions. One conceivable interpretation for
Exposure to LPA, an activator of RhoA, severely impairs den- the constitutively active Rac phenotype is that increased Rac activ-
dritic arbor elaboration, whereas expression of dominant-nega- ity triggers the initiation of dendritic branches, but Rac activity
tive Rho enhances dendritic arbor growth, consistent with the alone is not sufficient to maintain the newly formed branches.
findings in neuronal cell lines, where LPA causes neurite retrac- Consequently, they are retracted. Rac activity is also required to
tion and inhibition of RhoA induces neurite outgrowth10,11. This maintain dendritic branches, because more dendrites are retract-

articles
ed when Rac activity is reduced by expression of dom- a c

inant-negative Rac. In neurons from animals infected
with dominant-negative Rac vaccinia virus, the
enhanced rates of branch retractions contribute to the
increased fraction of transient branches in these arbors.
These experiments indicate the increased Rac activity
promotes branch additions. They further suggest that
Rac activity is necessary but not sufficient for branch
maintenance.
In addition to regulating the rate of branch addi-
tions, Rac is also involved in shaping the architecture
of the dendritic arbor. Sholl analysis indicates that
enhanced Rac activity increases the overall complexi-
ty of the dendritic arbor. Constitutively active Rac
increased the complexity of the distal dendritic arbor,
whereas it decreased the complexity of the proximal
b d
dendritic tree. The elaboration of proximal and distal
dendrites are differentially regulated2, possibly as a
result of differences in lamina-specific input activity

and the distributions of environmental signals such as
(m per 20 h)
TDBL growth
adhesion molecules32 and neurotrophins33. Indeed,
GTPase activity has been suggested to differentially reg-
ulate the elaboration of basal and apical dendrites in
dissociated cortical neurons12. Because the Rho fami-
ly GTPases are regulated by extracellular stimuli9, and
also seem to mediate neuronal responses to substrate-
bound guidance cues, including myelin and inte-
grins3436, normal dendritic arbor elaboration may
Fig. 6. NMDA receptor-mediated arbor growth operates through RhoA. Drawings of
reflect the complex composition of environmental cues two representative neurons from controls (a), APV-treated animals (b) and APV-
in intact animals. Consequently, in our experiments, treated animals expressing dominant-negative Rho (c). (d) Change in total dendritic
introduction of constitutively active Rac may have had branch length (TDBL) for neurons from the designated groups. For each condition,
different outcomes in distal and proximal dendrites 1224 neurons were analyzed. Scale bar, 50 m. *p < 0.05.
because of local variations in endogenous GTPase activ-
ity within the arbor.
Increased Cdc42 activity in tectal neurons increased
the proportion of transient dendrites. Active Cdc42 induces mechanisms, with respect to the cytoskeleton. The cytoskeleton
filopodia in growth cones37, which serve a sensory role in axon of transient dendritic branches may be entirely actin-based38
pathfinding. Cdc42 may serve a similar function to promote and their addition likely due to actin polymerization39. The
filopodia in the dendritic arbor, but does not seem to regulate maintenance of a newly added branch may be due to the inva-
either the stabilization or extension of dendritic branches. sion of the new branch by microtubules 39, as described for
growth cones40. Finally branch extension may be due to assem-
Interaction of GTPases with the cytoskeleton bly of microtubules. Although all three GTPases are reported to
Our data suggest that the GTPases modify specific aspects of regulate the actin cytoskeleton, RhoA may also regulate tubu-
dendritic arbor morphology. Rac seems to govern the rates of lin assembly41, supporting the idea that RhoAs principal effect
additions and retractions of new branches without affecting sub- on dendritic arbor elaboration is through the extension or
sequent regulatory events that determine whether the branch retraction of existing dendritic branches, which are enriched in
will extend. The increased actin filaments we observed by phal- microtubules38. In addition, GTPases can regulate cellcell con-
loidin staining in constitutively active Rac and constitutively tacts through presentation and clustering of cadherins and inte-
active Cdc42 cells suggest that these GTPases likely mediate their grins on the cell surface7. Evidence that cadherins can function
effects on the dendritic arbor dynamics by affecting the actin in synapse stabilization42 suggests an intriguing connection
cytoskeleton. Although constitutively active Cdc42 expression between GTPases and synaptogenesis.
increased filamentous actin according to the phalloidin stain- The GTPases are positioned in the midst of several signal
ing, it had only a modest effect on the parameters of dendritic transduction pathways, which govern cell shape and polarity7,9.
dynamics assayed here. It is possible that Cdc42 affects differ- The activity of each GTPase may be altered downstream of cell
ent aspects of cytoskeletal structure in neurons that were not surface receptors, including growth factors, cytokines and neu-
assessed in this study. Neither LPA nor dominant-negative RhoA rotransmitters7,14,37. We provide evidence that the NMDA-type
had a significant effect on branch dynamics, although they clear- glutamate receptor may be upstream of the RhoA GTPase in reg-
ly regulated branch shortening and extension, respectively. This ulating dendritic arbor elaboration. Indeed, it seems that NMDA
suggests that RhoA activity influences the extension or retrac- receptor activity decreases RhoA activity and thereby increases
tion of existing branches but does not govern the emergence of branch elongation. This observation is surprisingly comple-
new branches. Addition and retraction of short branches, main- mentary to a report that activation of the p75 neurotrophin
tenance of branches and their extension are distinct cell biolog- receptor can increase retinal axon elongation by decreasing
ical events that are likely controlled by different regulatory endogenous RhoA activity14. In addition, each of the GTPases

articles
affects the cytoskeleton through downstream effectors. It would Each optical section was an average of 816 frames. Animals were anes-
be of interest to identify the interacting partners through which thetized with 0.02% MS222 during DiI labeling, screening and imag-
the GTPases regulate dendritic arbor growth. Potential down- ing. Animals recovered from anesthesia between imaging sessions,
stream candidates for RhoA are the serine/threonine kinase, except for the three-minute imaging experiment, where animals were
ROCK, and mDia, which regulate the formation of different types anesthetized throughout the imaging protocol. In experiments using
LPA to activate RhoA or APV to block NMDA receptors, the drug was
of actin fibers43. Rac may act through LIM kinase and cofilin44,45.
added to the rearing solution immediately after the first image was
Filopodia formation induced by Cdc42 seems to depend on collected.
N-WASP, a ubiquitously expressed Cdc42 binding protein46.
Finally, Rac and Cdc42 share common effectors, including the Image analysis. Dendritic arbors were reconstructed by tracing the por-
serine/threonine kinase PAK, which is involved in the morpho- tion of the neuron in each optical section onto an acetate sheet until the
logical changes mediated by Rac and Cdc42 (refs. 9, 47). entire neuron was drawn. This method provides a more detailed repre-
sentation of the morphology than the three-dimensional reconstruction
generated by computer, because fine processes visible in the optical sec-
METHODS tions are lost in the computer-generated reconstruction. Total dendrit-
Construction of recombinant vaccinia virus. Human RacV12, RacN17
ic branch length was measured from scanned drawings of cells with NIH
and Cdc42V12 were myc tagged at the amino (N) terminal. Myc-RacV12,
Image 1.61. The number of branch tips was counted manually. To analyze
myc-RacN17 and myc-Cdc42V12 were cloned into the Sal I/Spe I site
the arbor dynamics, drawings of cells from sequential time points were
upstream of IRES-EGFP in the pBluescript vector. The myc-GTPase-
superimposed to identify added and retracted branches. The magnitude
IRES-EGFP constructs were cut out from pBluescript vector and cloned
of arbor dynamics determined in this and previous studies3,4, in terms
into the vaccinia virus vector pSC65 downstream of a strong synthetic

of numbers of branches added or retracted over a two-hour period, may
early/late vaccinia virus promoter. The cDNA encoding EGFP-Cdc42N17
be underestimated by five- to tenfold24,49, because unobserved branch-
and EGFP-RhoN19 N-terminal fusion proteins were cloned into the
es are both added and retracted during the two-hour intervals. The short-
Sal I /Sma I sites of pSC65 downstream of the strong synthetic early/late
est interval over which we can collect images is three minutes because
vaccinia virus promoter. The EGFP fusion proteins are active in vitro in
this is about the time it takes to collect and save a single z series through
a variety of cell lines assayed for cell morphology and cell adhesion
the optic tectum. Observations collected at such frequent time points
(L. Van Aelst, Cold Spring Harbor Laboratory, Cold Spring Harbor, New
over our 20-hour imaging period would provide a more accurate value of
York, personal communication). A virus expressing only EGFP driven
branch additions and retractions; however, neither the DiI-labeled neu-
by the early/late promotor was used as a control for the effects of viral
rons nor the animals can survive such prolonged imaging session or sus-
infection. All viruses also express -galactosidase (-gal) behind a weak-
tained anesthesia49. Branch additions, branch retractions and the change
er p7.5 viral promotor, which is used for plaque selection of recombi-
in total dendritic branch length in uninfected control animals and animals
nant viruses48. Constructs were confirmed by sequencing. Recombinant
infected with EGFP vaccinia virus were comparable, consistent with pre-
vaccinia viruses, obtained by homologous recombination of pSC65 and
vious observations that vaccinia virus does not affect the development
wild-type vaccinia virus as reported3, were purified and titered before
of tectal cell morphology3. We therefore pooled these two groups of con-
use48. High-titer virus (over 107 plaque-forming units, pfu) was used to
trol cells together.
infect animals in imaging experiments.
Sholl analysis was done with Object-Image software (NIH Image). A
scanned drawing of the neuron was overlaid on a series of concentric cir-
Viral infection. Albino Xenopus laevis tadpoles were obtained by mating
cles, spaced every 5 m with the cell body in the center. The number of
induced by human chorionic gonadotropin injections. Purified virus
dendritic branches crossing each concentric circle was marked on the
(100150 nl mixed with 0.1% fast green) was injected into the tectal ven-
composite image and counted.
tricle of stage 46 tadpoles anesthetized with 0.02% 3-aminobenzoic acid
Statistical analysis was done with two-tailed t-test.
ethyl ester (MS222).
Note: Time-lapse movies can be found on the Nature Neuroscience web site
Immunostaining. Animals were fixed in 4% paraformaldehyde in 0.1 M
(http://neurosci.nature.com/web_specials/).
phosphate buffer (PB, pH7.4) overnight at 4oC, rinsed in PB and cry-
oprotected in 30% sucrose before brains were cut into 20-m cryostat
sections. For myc immunostaining, sections were preincubated with block- ACKNOWLEDGEMENTS
ing solution containing 5% goat serum and 0.3% Triton X-100 in PB for We thank Wun Chey Sin for providing the EGFP-Cdc42N17 and EGFP-RhoN19
1 h, followed by an overnight incubation at 4oC in anti-myc antibody viruses, Kim Bronson for technical assistance and Neil Mahapatra for help with
(Calbiochem), diluted 1:100 in blocking solution. After rinsing, sections the Sholl analysis. Support for the work was provided by the NIH (H.T.C.,
were incubated in cy5-goat anti-mouse secondary antibody (diluted 1:100 L.V.A.) and the Eppley Foundation (H.T.C.).
in blocking solution; Jackson Immunoresearch Laboratories, West Grove)
for 30 minutes. We did not observe differing levels of myc immunoreac- RECEIVED 14 DECEMBER 1999; ACCEPTED 27 JANUARY 2000
tivity in neurons, suggesting that expression levels of the GTPases did not
vary over a detectable range. Rhodamine-phalloidin (Molecular Probes
Eugene, Oregon; final dilution 1:800) was added to the incubation solution 1. Stuart, G., Spruston, N. & Hausser, M. Dendrites (Oxford Univ. Press,
together with secondary antibody for double labeling. Oxford, 1999).
2. Koester, S. E. & OLeary, D. D. Development of projection neurons of the
mammalian cerebral cortex. Prog. Brain Res. 102, 207215 (1994).
Image acquisition. Single tectal neurons were labeled by iontophore- 3. Wu, G.-Y. & Cline, H. T. Stabilization of dendritic arbor structure in vivo by
sis of DiI (0.02% 1,1-dioctadecyl-3,3,33-tetramethylindocarbocya- CaMKII. Science 279, 222226 (1998).
nine perchlorate, Molecular Probes) in ethanol. Positive current 4. Wu, G.-Y., Zou, D. J., Rajan, I. & Cline, H. T. Dendritic dynamics in vivo
(110 nA) was used in 310 pulses of 200-ms duration. In animals change during neuronal maturation. J. Neurosci. 19, 44724483 (1999).
infected with EGFP fusion proteins or IRES-EGFP constructs, DiI- 5. Rajan, I. & Cline, H. T. Glutamate receptor activity is required for normal
development of tectal cell dendrites in vivo. J. Neurosci. 18, 78367846
labeled neurons were located within an infected region of the tectum; (1998).
however, it was not always possible to verify that each DiI-labeled neu- 6. Rajan, I., Witte, S. & Cline, H. T. NMDA receptor activity stabilizes
ron was infected. We selected animals with single brightly labeled neu- presynaptic retinotectal axons and postsynaptic optic tectal cell dendrites in
rons two hours after DiI labeling. Images were collected at 2-m steps vivo. J. Neurobiol. 38, 20892096 (1999).
7. Hall, A. Rho GTPases and the actin cytoskeleton. Science 279, 509514
through the entire z dimension of labeled neurons with a Noran Instru- (1998).
ments XL laser scanning confocal attachment mounted on an upright 8. Narumiya, S. The small GTPase Rho: cellular functions and signal
Nikon Optiphot through a 40 Nikon oil immersion lens (1.30 NA). transduction. J. Biochem. (Tokyo) 120, 215228 (1996).

articles
9. Van Aelst, L. & DSouza-Schorey, C. Rho GTPases and signaling networks. 29. Kim, J. H., Liao, D., Lau, L.-F. & Huganir, R. L. SynGAP: a synaptic RasGAP that
Genes Dev. 11, 22952322 (1997). associates with the PSD-95/SAP90 protein family. Neuron 20, 683691 (1998).
10. Van Leeuwen, F. N. et al. The guanine nucleotide exchange factor Tiam1 30. Ridley, A. J. et al. The small GTP-binding protein rac regulates growth factor-
affects neuronal morphology; opposing roles for the small GTPases Rac and induced membrane ruffling. Cell 70, 401410 (1992).
Rho. J. Cell Biol. 139, 797807 (1997). 31. Luo, L., Liao, Y. J., Jan, L. Y. & Jan, Y. N. Distinct morphogenetic functions of
11. Gebbink, M. F. et al. Identification of a novel, putative Rho-specific similar small GTPases: Drosophila Drac1 is involved in axonal outgrowth
GDP/GTP exchange factor and a RhoA-binding protein: control of neuronal and myoblast fusion. Genes Dev. 8, 17871802 (1994).
morphology. J. Cell Biol. 137, 16031613 (1997). 32. Miskevich, F., Zhu, Y., Ranscht, B. & Sanes, J. R. Expression of multiple
12. Threadgill, R., Bobb, K. & Ghosh, A. Regulation of dendritic growth and cadherins and catenins in the chick optic tectum. Mol. Cell Neurosci. 4,
remodeling by Rho, Rac, and Cdc42. Neuron 19, 625634 (1997). 240255 (1998).
13. Katz, L. C. & Shatz, C. J. Synaptic activity and the construction of cortical 33. McAllister, A. K., Lo, D. C. & Katz, L. C. Neurotrophins regulate dendritic
circuits. Science 274, 11321138 (1996). growth in developing visual cortex. Neuron 15, 791803 (1995).
14. Yamashita, T., Tucker, K. L. & Barde, Y.-A. Neurotrophin binding to the p75 34. Kuhn, T. B. et al. Myelin and collapsin-1 induce motor neuron growth cone
receptor modulates Rho activity and axonal outgrowth. Neuron 24, 585605 collapse through different pathways: inhibition of collapse by opposing
(1999). mutants of rac1. J. Neurosci. 19, 19651975 (1999).
15. Luo, L., Liao, Y. J., Jan, L. Y. & Jan, Y. N. Distinct morphogenetic functions of 35. Jin, Z. & Strittmatter, S. M. Rac1 mediates collapsin-1-induced growth cone
similar small GTPases: Drosophila Drac1 is involved in axonal outgrowth and collapse. J. Neurosci. 17, 62566263 (1997).
myoblast fusion. Genes Dev. 8, 17871802 (1994). 36. Kuhn, T. B., Brown, M. D. & Bamburg, J. R. Rac1-dependent actin filament
16. Luo, L. et al. Differential effects of the Rac GTPase on Purkinje cell axons and organization in growth cones is necessary for beta1-integrin-mediated advance
dendritic trunks and spines. Nature 379, 837840 (1996). but not for growth on poly-D-lysine. J. Neurobiol. 37, 524540 (1998).
17. Kaufmann, N., Wills, Z. P. & Van Vactor, D. Drosophila Rac1 controls motor 37. Kozma, R., Sarner, S., Ahmed, S. & Lim, L. Rho family GTPases and neuronal
axon guidance. Development 125, 453461 (1998). growth cone remodeling: relationship between increased complexity induced
18. Ruchhoeft, M. L. et al. The neuronal archicture of Xenopus retinal ganglion by Cdc42Hs, Rac1, and acetylcholine and collapse induced by RhoA and
cells is sculpted by Rho-family GTPases in vivo. J. Neurosci. 19, 84548463 lysophosphatidic acid. Mol. Cell Biol. 17, 12011211 (1997).
(1999). 38. Fiala, J. C., Feinberg, M., Popov, V. & Harris, K. M. Synaptogenesis via
19. Zipkin, I. D., Kindt, R. M. & Kenyon, C. J. Role of a new Rho family dendritic filopodia in developing hippocampal area CA1. J. Neurosci. 18,
member in cell migration and axon guidance in C. elegans. Cell 90, 89008911 (1998).
883894 (1997). 39. Ziv, N. E. & Smith, S. J. Evidence for a role of dendritic filopodia in
20. Lee, T. et al. Essential roles of Drosophila RhoA in the regulation of synaptogenesis and spine formation. Neuron 17, 91102 (1996).
neuroblast proliferation and dendritic but not axonal morphogenesis. 40. Bentley, D. & OConnor, T. P. Cytoskeletal events in growth cone steering.
Neuron (in press). Curr. Opin. Neurobiol. 4, 4348 (1994).
21. Wu, G.-Y., Zou, D.-J., Koothan, T. & Cline, H. T. Infection of frog neurons 41. Hirose, M. et al. Molecular dissection of the Rho-associated protein kinase
with vaccinia virus permits in vivo expression of foreign proteins. Neuron 14, (p160ROCK)-regulated neurite remodeling in neuroblastoma N1E-115 cells.
681684 (1995). J. Cell Biol. 141, 16251636 (1998).
22. Jalink, K. et al. Inhibition of lysophosphatidate- and thrombin-induced 42. Fannon, A. M. & Colman, D. R. A model for central synaptic junctional
neurite retraction and neuronal cell rounding by ADP ribosylation of the complex formation based on the differential adhesive specificities of the
small GTP-binding protein Rho. J. Cell Biol. 126, 801810 (1994). cadherins. Neuron 17, 423434 (1996).
23. ORourke, N. A., Cline, H. T. & Fraser, S. E. Rapid remodeling of retinal 43. Maekawa, M. et al. Signaling from Rho to the actin cytoskeleton through
arbors in the tectum with and without blockade of synaptic transmission. protein kinases ROCK and LIM-kinase. Science 285, 895898 (1999).
Neuron 12, 921934 (1994). 44. Arber, S. et al. Regulation of actin dynamics through phosphorylation of
24. Witte, S., Stier, H. & Cline, H. T. In vivo observations of timecourse and cofilin by LIM-kinase. Nature 393, 805809 (1998).
distribution of morphological dynamics in Xenopus retinotectal axon arbors. 45. Yang, N. et al. Cofilin phosphorylation by LIM-kinase 1 and its role in Rac-
J. Neurobiol. 31, 219234 (1996). mediated actin reorganization. Nature 393, 809812 (1998).
25. Wu, G.-Y., Malinow, R. & Cline, H. T. Maturation of a central glutamatergic 46. Miki, H., Sasaki, T., Takai, Y. & Takenawa, T. Induction of filopodium
synapse. Science 274, 972976 (1996). formation by a WASP-related actin-depolymerizing protein N-WASP. Nature
26. Purpura, D. P. Dendritic differentiation in human cerebral cortex: normal 391, 9396 (1998).
and aberrant developmental patterns. Adv. Neurol. 12, 91134 (1975). 47. Obermeier, A. et al. PAK promotes morphological changes by acting
27. Billuart, P. et al. Oligophrenin-1 encodes a rhoGAP protein involved in X- upstream of Rac. EMBO J. 17, 43284339 (1998).
linked mental retardation. Nature 392, 923926 (1998). 48. Mackett, M., Smith, G. L. & Moss, B. DNA Cloning: A Practical Approach (IRL
28. Chen, H. J., Rojas, S. M., Oguni, A. & Kennedy, M. B. A synaptic Ras-GTPase Press, Oxford, 1985).
activating protein (p135 SynGAP) inhibited by CaM kinase II. Neuron 20, 49. Cline, H. T. et al. in Imaging: A Laboratory Manual (eds. Yuste, R., Lanni, F.
895904 (1998). & Konnerth, A.) 13.113.12 (Cold Spring Harbor Laboratory Press, 1999).

articles
Anatomical and physiological evidence

for D1 and D2 dopamine receptor
colocalization in neostriatal neurons
Oleg Aizman1, Hjalmar Brismar1, Per Uhln1, Eivor Zettergren1, Allan I. Levey2, Hans Forssberg1,
Paul Greengard3 and Anita Aperia1
1 Department of Woman and Child Health, Karolinska Institutet, Astrid Lindgren Childrens Hospital, Q2:09, 171 76 Stockholm, Sweden
2 Department of Neurology, Emory University School of Medicine, Atlanta, Georgia, Georgia 30322, USA
3 Laboratory of Molecular and Cellular Neuroscience, The Rockefeller University, New York, New York 10021-6399, USA
Correspondence should be addressed to A.A. (aniap@child.ks.se)
O.A. and H.B. contributed equally to this work.
Despite the importance of dopamine signaling, it remains unknown if the two major subclasses of
dopamine receptors exist on the same or distinct populations of neurons. Here we used confocal
microscopy to demonstrate that virtually all striatal neurons, both in vitro and in vivo, contained
dopamine receptors of both classes. We also provide functional evidence for such colocalization: in
essentially all neurons examined, fenoldopam, an agonist of the D1 subclass of receptors, inhibited
both the Na+/K+ pump and tetrodotoxin (TTX)-sensitive sodium channels, and quinpirole, an agonist
of the D2 subclass of receptors, activated TTX-sensitive sodium channels. Thus D1 and D2 classes of lig-
ands may functionally interact in virtually all dopamine-responsive neurons within the basal ganglia.
Of the vast number of biogenic amines and peptide neurotrans- protein (Fig. 1c). DIC imaging verified the absence of unlabeled
mitters acting in the mammalian brain, dopamine receives by far (glial) cells. Cultured neurons were labeled using antibodies
the most attention1. This is partly explained because abnormal- against either D1-subclass or D2-subclass receptors, each raised in
ities of dopaminergic neurotransmission are implicated in the two species (Fig. 1dl). The cells were examined by confocal
etiology of several major neurological and psychiatric disorders, microscopy. The improved signal-to-noise ratio for the deter-
including Parkinsons disease2,3, schizophrenia4,5, attention deficit mination of dopamine receptors, made possible by the smaller
hyperactivity disorder6,7 and drug abuse8,9. There is considerable depth of focus and the high lateral resolution, permits detection
evidence from immunocytochemical as well as in situ hybridiza- of substances present in low abundance. In all cases, we detected
tion studies that the two major subclasses of dopamine receptors immunofluorescence from virtually all cells. With both anti-
are enriched in different projection neurons in the neostria- bodies against the D1 receptor subclass, the signal was random-
tum1015. Thus, the D1 subclass of dopamine receptors is enriched ly distributed in the cytoplasm and in the vicinity of the plasma
in striatal neurons containing substance P and dynorphin that membrane; with both antibodies against the D2 receptor sub-
project to the substantia nigra, pars reticulata and entopedun- class, the signal was more distinct in the region of the plasma
cular nucleus16,17. Conversely, the D2 subclass of dopamine recep- membrane and in the perinuclear region. Both D1 and D2 sig-
tors is enriched in enkaphalin-containing striatal neurons nals were present in dendrites. Double labeling confirmed the
projecting to the external segment of the globus pallidus16,17. In co-existence of D1- and D2- like receptors in virtually all cells.
contrast to these anatomical data, electrophysiological measure- DARPP-32 (dopamine- and cyclic AMP-regulated phospho-
ments as well as measurements of mRNA in single cells suggest protein, 32 kDa) is present in all medium spiny neurons and only
that many striatal neurons contain both classes of dopamine in medium spiny neurons in the striatum and nucleus accum-
receptors18. The advent of confocal microscopy led us to re-exam- bens19,20. We used this selective distribution to identify cells con-
ine the possible colocalization of D1 and D2 subclasses of recep- taining both D 1 and D 2 subclasses of receptors. DARPP-32
tors. Moreover, on-line recording of intracellular sodium immunofluorescence was observed in virtually all cultured neu-
concentration in these cells permitted functional evaluation of rons. Double labeling of cultured neurons for D1-subclass recep-
the co-expression of the two subclasses of dopamine receptors. tors and DARPP-32 (Fig. 2ac) or D2-subclass receptors and
DARPP-32 (Fig. 2df) indicated colocalization of both subclass-
RESULTS es of receptors with DARPP-32 in virtually all cells; no neurons
After two to three weeks in culture, virtually all cells from embry- were singly labeled for D1- or D2-subclass receptors. These results
onic rat striatum seemed to be mature neurons, as they resem- identify the dopamine receptor-labeled cultured cells as medi-
bled typical medium spiny neurons and stained positively for um spiny neurons.
three neuronal markers: -tubulin III (Fig. 1a), the alpha 3 iso- To evaluate the possibility that the high degree of colocalization
form of Na+/K+-ATPase (Fig. 1b) and neuron-specific nuclear of D1 and D2 subclasses of dopamine receptors resulted from an

articles
Fig. 1. Primary culture of medium spiny neurons from rat embryonic

striatum stained for neuronal markers and for D1 and D2 receptors.
(ac) Combined DIC and confocal immunofluorescence images.
(a) Alexa 488 labeling of -tubulin III. (b) Alpha 3 isoform of Na+/K+-
ATPase labeled with Alexa 488. (c) Neuron-specific nuclear protein
labeled with Alexa 488. (d) D1 receptors labeled with rabbit anti-human
D1 receptor primary antibody and Alexa 488-labeled goat anti-rabbit
secondary antibody. (e) D2 receptors labeled with goat anti-human D2
receptor primary antibody and Cy3-labeled donkey anti-goat secondary
antibody. (f) D1 and D2 receptors, double labeled using same antibodies
as for (d, e). (gi) Immunocytochemical labeling with antibodies used in
(df) at higher magnification. (j) D1 receptors labeled with rat mono-
clonal anti-human D1 receptor primary antibody and Alexa 488-labeled
goat anti-mouse secondary antibody. (k) D2 receptors labeled with rab-
bit anti-human D2 receptor primary antibody and Alexa 546-labeled
goat anti-rabbit secondary antibody. (l) D1 and D2 receptors double
labeled using same antibodies as for (j, k). Green signal, D1; red signal,
D2. Yellow indicates overlap.
alteration in the properties of neurons under culture conditions,

we examined the distribution of dopamine receptors in ten slices
freshly prepared from neostriatum as well as from accumbens shell
of each of three adult rats. In these slices, D1 and D2 subclasses of
receptors were colocalized as in the cultured neurons (Fig. 3). Con-
focal recordings of the slices were inspected by an observer blind
to the experimental procedure. In 10 randomly selected fields of
view (250 m 250 m), a total of 519 cells were identified. We
found that 517 of 519 cells (> 99% of the neurons) were positive
for both D1 and D2 receptor subclasses. We also examined the colo-
calization of D1 or D2 subclasses with DARPP-32 in these slices
(Fig. 2gl). Receptors were detected only in DARPP-32-labeled cells, We obtained evidence of a functional role for the ubiquitously
demonstrating that dopamine receptors were present only on neu- expressed D1- and D2-subclass dopamine receptors in studies of
rons, not on glial cells. Na+ influx and efflux in individual cells. This required measure-
ment of initial rates of TTX-sensitive Na+ influx and ouabain-
sensitive Na+ outflux21 (Fig. 4a; see Methods). Cells were loaded
with the sodium-sensitive dye, SBFI-AM. After recording the
basal level, extracellular Na+ was replaced by choline chloride.
When the intracellular Na+ concentration approached zero, extra-
cellular Na+ was restored, and K+ was removed from the medi-
um to prevent pump-mediated Na+ efflux. This resulted in a rapid
influx of Na+. In all protocols, TTX decreased the initial rate of
Na+ influx by approximately 6070%. When the concentration
of intracellular Na+ started to level off, the pump was reactivated
by adding 4 mM K+ to the medium and replacing Na+ by choline
chloride. To prevent influx of Na+ via channels, both TTX and
the NMDA- and AMPA-receptor inhibitors MK801 and CNQX
were added to the medium. The rapid efflux of Na+ from the cells
was almost completely ouabain dependent.
The specific D1 subclass agonist fenoldopam caused a pro-
nounced and highly significant decrease in TTX-sensitive Na+
Fig. 2. D1 and D2 subclasses of receptors colocalize with DARPP-32 in

primary cultures and slices of rat neostriatal neurons. (af) Primary cul-
tures stained for D1 receptor (a), DARPP-32 (b), D1 receptor and
DARPP-32 (c), DARPP-32 (d), D2 receptor (e) or DARPP-32 and D2
receptor (f). (gl) Slices stained for DARPP-32 (g), D1 receptor (h),
DARPP-32 and D1 receptor (i), DARPP-32 (j), D2 receptor (k) or
DARPP-32 and D2 receptor (l). D1 and D2 receptor antibodies were as
described in Fig. 1d and e in all frames except (h); here Alexa 546-
labeled goat anti-rabbit was used. DARPP-32 was stained with mouse
monoclonal anti-bovine DARPP-32 primary antibody and TexasRedX-
(b, c), Oregon Green- (d, f) and and Alexa 488- (g, i, j, l) labeled goat
anti-mouse secondary antibodies.

articles
virtually all neurons provides functional evidence for colocal-

ization of the two classes of dopamine receptors.
DISCUSSION
Our immunocytochemical studies indicated that virtually all stri-
atal projection neurons contained both D1 and D2 subclasses of
dopamine receptors. Moreover, electrophysiological measure-
ments of Ca2+ and Na+ currents18,22, together with our
a b 0
60
0
50
0
40
[Na+]i (mM)
Ratio/min
0
30
0
20 0
10 0
0 0
Fig. 3. D1-subclass receptors colocalize with D2-subclass receptors in neurons of
0.0 5.0 10.0 15.0 20.0
rat neostriatal slices. (ac) Neostriatum, high magnification. (df) Neostriatum,
140 mM Na+ min
low magnification. (gi) Accumbens shell, low magnification. (a, d, g) D1 receptor
labeling. (b, e, h). D2 receptor labeling. (c, f, i) D1 and D2 receptor labeling. 4 mM K+
Antibodies were as described in Fig. 1d and e.
b 0.140
influx (Fig. 4b, upper panel). The effect was reversed in the pres-
0.120 Influx
ence of the D1 subclass antagonist, SCH 23390. Fenoldopam also
significantly inhibited Na+/K+-ATPase-dependent Na+ efflux, and Efflux
0.100
this effect of fenoldopam was also abolished by SCH23390. The
Ratio/min
D2 subclass agonist quinpirole caused a pronounced and highly 0.080

significant stimulation of TTX-sensitive Na+ influx, an effect that
was abolished by the D2 subclass antagonist sulpiride (Fig. 4b, 0.060
lower panel). Quinpirole had no significant effect on Na+/K+-
ATPase-dependent Na+ efflux. 0.040 *
*
In resting cells, free intracellular Na+ was 16.0 mM. When
0.020
quinpirole was added to the medium, intracellular Na+ rapidly
increased (Fig. 5b); this occurred in virtually all cells and was 0.000
attributable to the activation of TTX-sensitive Na+ channels
Fenoldopam
Fenoldopam
Fenoldopam
Fenoldopam
+SCH23390
+SCH23390
Control
Control
0.0
(Fig. 4b, lower panel). Virtually all cells also showed a slow min
increase in intracellular Na+ in response to fenoldopam (Fig. 5a),
indicating that the effect of this agonist of D1-subclass receptors
on pump inhibition masked its effect on channel inhibition 0.200
(Fig. 4b, upper panel). The finding that sodium levels were altered *
both by stimulation of D1 and by stimulation of D2 receptors in 0.180
0.160
0.140 Influx
Fig. 4. Effects of D1 and D2 agonists and antagonists on initial rates of Efflux
Ratio/min
0.120
Na+ influx and efflux in primary cultures of striatal cells.
0.100
(a) Experimental design to study the initial rates of Na+ influx and efflux
under Vmax conditions21. (b) Fenoldopam (105 M; a D1 agonist), 0.080
SCH23390 (105 M; a D1 antagonist), quinpirole (105 M; a D2 agonist) 0.060
and sulpiride (105 M; a D2 antagonist) were present in the incubation
0.040
mixture as indicated. Images were taken every 20 s. From each cover-
slip, a field of view containing 515 cells was examined. Values (means 0.020
s.e.) for each coverslip were calculated as rate of change in sodium 0.000
Quinpirole
Quinpirole
+ sulpiride
Quinpirole
Quinpirole
+ sulpiride
(340:380 ratio) per min. Each group included 612 experiments. For
Control
Control
both influx and efflux studies, 50120 cells were used. The control and
treatment experiments were carried out on the same day on paired cov-
erslips from the same batch of cells. *Significant differences between
control and treated cells, p < 0.01, Students t-test.

articles
Na+/K+-ATPase indicate that the 1 isoform of the sodium pump

has a higher affinity for Na+ than the 3 isoform31,32.
The effects of D1 and D2 subclasses of receptors on sodium
fluxes and intracellular sodium concentration have important
physiological implications. Regulation of the TTX-sensitive sodi-
um channels directly affects the onset of the action potential. The
hyperpolarization between bursts of action potentials results from
sodium extrusion via the electrogenic sodium pump; hence inhi-
bition of the sodium pump affects the firing rate33. In addition,
sodium may serve as a signaling ion in postsynaptic neurons:
intracellular sodium regulates NMDA receptor activity34. Thus
opposing effects of D1 and D2 receptors on sodium fluxes and
their net effect on intracellular sodium are probably important
in the physiological and pathophysiological roles of dopamine
receptors in various complex behaviors.
Our results provide compelling evidence for the presence of
both classes of dopamine receptors on virtually all dopaminocep-
tive neurons in the striatum and indicate physiological activity
of these ubiquitously distributed D1 and D2 subclasses. This sug-

gests intracellular rather than intercellular mechanisms for many
of the synergistic and antagonistic actions exerted by activation of
D1 and D2 subclasses of dopamine receptors on numerous sig-
nal-transduction pathways.
METHODS
Cell culture. Cultures of striatal neurons were prepared from 1819 day-
old rat embryos as described35, with modifications. The cells were plated
Fig. 5. Effects of D1 and D2 agonists on [Na+]i in primary cultures of rat on poly-D-lysine-coated glass coverslips and cultured in Eagles minimal
striatal cells. At time 0, SBFI-loaded cells were incubated with essential medium (MEM) plus F-12 (1:1). The cultures were grown in
5 105 M fenoldopam (a) or 5 105 M quinpirole (b), and the ratio the presence of 5% FBS. To suppress the growth of glial cells, FBS was
images were recorded every 30 s. Values represent ratio changes from exchanged with N2 (1%) two days after plating, and cytosine arabinoside
the basal level. The results were corrected for SBFI bleaching and leak- (5 M) was added to the culture medium from day 5 to day 7. Cells were
age. Curves were fit by using the method of least squares. Digital images maintained in culture for two to three weeks before experiments.
of SBFI fluorescence under control conditions and eight min after
fenoldopam (a) or quinpirole (b) treatment are shown in insets. Intracellular Na+ measurement with SBFI. Cells were loaded with the
Changes in color from blue to red on the pseudo-color scale used in sodium-sensitive fluorescent probe, SBFI-AM (sodium-binding benzo-
these images indicate increases in [Na+]i. furan isophthalate; 10 M). Coverslips were then transferred to an exper-
imental chamber (FCS2, Bioptechs, Butler, Pennsylvania) with laminar
flow that allowed instantaneous exposure to the indicated drugs. The
cells were excited at wavelength 340/10 nm and 380/10 nm every
2030 s with a 400-ms exposure time. Emission fluorescence was select-
results on Na+ influx and efflux, provide physiological evidence ed and collected with a 510/30 nm band-pass filter. Recordings were made
for a high degree of colocalization of D1- and D2-subclass recep- from cell soma. Collected data were then processed by an image acquisi-
tors. Contradictory evidence suggesting a low degree of overlap of tion program from Inovision Corporation, Raleigh, North Carolina. Cal-
D1 and D2 subclasses of receptors1013 can be reconciled with our ibration of intracellular [Na+] was carried out in living cells.
conclusion by assuming that the striatonigral projection neurons
contain high levels of D1-subclass receptors and low levels of D2- Immunostaining. Cells were fixed for 20 min in ice-cold
subclass receptors, and that the converse is true for the stri- 2% paraformaldehyde, permeabilized with 0.1% saponin, blocked with
atopallidal pathway. Thus, previous failure to substantially 7% normal goat (NGS) or donkey (NDS) serum and then incubated
colocalize the two subclasses using immunocytochemistry23 and overnight at 4C with primary antibody in PBS containing 2.8% serum
and 0.1% saponin. The cells were washed, incubated with fluorescent
in situ hybridization24 is probably attributable to inadequate sen- secondary antibody in PBS with 2.8% serum and 0.1% saponin and
sitivity of these analytical procedures. Conversely, the high degree washed again. For double labeling, the cells were incubated again
of overlap of D1 and D2 subclasses of receptor mRNA seen using overnight at 4C with the second primary antibody, followed by incuba-
PCR25 is attributable to the extremely high sensitivity of this tech- tion with the second fluorescent secondary antibody. The slides were
nique. Furthermore, our data suggest that almost undetectable subsequently mounted in Prolong antifade (Molecular Probes, Eugene,
levels of D1-subclass receptors in striatopallidal neurons and of Oregon). Similar protocols were used for the immunostaining of adult
D2-subclass receptors in striatonigral neurons were adequate to rat striatal slices with minor modifications. D1 subclass dopamine recep-
produce physiological responses. tors were probed with an affinity-purified rabbit polyclonal antibody
Intracellular Na+ in cultured striatal neurons was 16 mM, con- against human D1 dopamine receptor (1:100; ref. 36) or with a rat mon-
oclonal antibody against the human D1 dopamine receptor (1:100;
siderably higher than in non-neuronal cells, in which [Na+]i is
ref. 23). D2-subclass dopamine receptors were probed with polyclonal
generally 37 mM2628. We attribute the high [Na+]i in neurons goat affinity-purified anti-human D2 dopamine-receptor antibody (1:200;
to the presence of the neuron-specific catalytic subunit of the Santa Cruz Biotechnology, Santa Cruz, California) or with an affinity-
sodium pump, 3. Most non-neuronal cells express only the purified rabbit polyclonal antibody against human D2 (1:100; ref. 23).
more widespread 1 isoform29,30. Studies on purified Na+/K+- The two D1 antibodies used are both reported to be specific for the D1
ATPase and on cells transfected with different isoforms of receptor in the D1 subclass of dopamine receptors and not to crossreact

articles
with the D5 receptor23,36. Likewise, the two D2 antibodies are both report- 13. Surmeier, D. J., Reiner, A., Levine, M. S. & Ariano, M. A. Are neostriatal
ed to be specific for the D2 receptor in the D2 subclass of dopamine recep- dopamine receptors co-localized? Trends Neurosci. 16, 299305 (1993).
tors and not to crossreact with the D3 and D4 receptor23 (statement from 14. Gerfen, C. R. The neostriatal mosaic: multiple levels of compartmental
organization. Trends Neurosci. 15, 133139 (1992).
Santa Cruz Biotechnology). DARPP-32 was detected with a monoclonal 15. Le Moine, C. & Bloch, B. D1 and D2 dopamine receptor gene expression in
mouse antibody against bovine DARPP-32 (1:1000; ref. 37). The -tubu- the rat striatum: sensitive cRNA probes demonstrate prominent segregation
lin III was detected with a monoclonal mouse antibody against human - of D1 and D2 mRNAs in distinct neuronal populations of the dorsal and
tubulin III (1:400; Sigma), the neuron-specific nuclear protein was ventral striatum. J. Comp. Neurol. 355, 418426 (1995).
16. Gerfen, C. R. et al. D1 and D2 dopamine receptor-regulated gene
detected with a mouse monoclonal antibody against neuronal nuclei expression of striatonigral and striatopallidal neurons. Science 250,
(1:50; Chemicon, Temecula, California). The 3 isoform of Na+/K+- 14291432 (1990).
ATPase was probed with polyclonal rabbit affinity-purified anti-rat anti- 17. Yung, K. K. et al. Immunocytochemical localization of D1 and D2 dopamine
body against the 3 isoform of Na+/K+-ATPase (gift from M. Caplan, receptors in the basal ganglia of the rat: light and electron microscopy.
Yale University, New Haven, Connecticut). Donkey-anti-goat Cy3 (Jack- Neuroscience 65, 709730 (1995).
18. Surmeier, D. J. et al. Dopamine receptor subtypes colocalize in rat
son ImmunoResearch, West Grove, Pennsylvania), goat-anti-rabbit Ore- striatonigral neurons. Proc. Natl. Acad. Sci. USA 89, 1017810182 (1992).
gon green, goat-anti-rabbit TexasRedX, goat-anti-rabbit Alexa 488, 19. Ouimet, C. C. & Greengard, P. Distribution of DARPP-32 in the basal
goat-anti-rabbit Alexa 546, goat-anti-mouse Oregon green, goat-anti- ganglia: an electron microscopic study. J. Neurocytol. 19, 3952 (1990).
mouse TexasRedX and goat-anti-mouse Alexa 488 (all from Molecular 20. Anderson, K. D. & Reiner, A. Immunohistochemical localization of DARPP-
32 in striatal projection neurons and striatal interneurons: implications for
Probes) were used at 1:200 as secondary antibodies. the localization of D1-like dopamine receptors on different types of striatal
neurons. Brain Res. 568, 235243 (1991).
Confocal microscopy. The immunolabeled cells were recorded with a 21. Borin, M. L. Dual inhibitory effects of dopamine on Na+ homeostasis in rat
Zeiss LSM410 or a Leica TCS SP inverted confocal scanning laser micro- aorta smooth muscle cells. Am. J. Physiol. 272, C428438 (1997).
scope using 63/1.4 N.A. and 20/0.75 N.A. objectives. Green fluores- 22. Cepeda, C., Buchwald, N. A. & Levine, M. S. Neuromodulatory actions of
dopamine in the neostriatum are dependent upon the excitatory amino acid
cence was excited at 488 nm and detected with a 515540 nm band-pass receptor subtypes activated. Proc. Natl. Acad. Sci. USA 90, 95769580 (1993).
filter. Red fluorescence were excited at 543 nm and detected with a 570 23. Levey, A. I. et al. Localization of D1 and D2 dopamine receptors in brain with
nm long-pass filter. Preparations in which the primary antibody was subtype-specific antibodies. Proc. Natl. Acad. Sci. USA 90, 88618865 (1993).
omitted from the staining protocol were used as negative controls. 24. Le Moine, C., Normand, E. & Bloch, B. Phenotypical characterization of the
rat striatal neurons expressing the D1 dopamine receptor gene. Proc. Natl.
Acad. Sci. USA 88, 42054209 (1991).
ACKNOWLEDGEMENTS 25. Surmeier, D. J., Song, W. J. & Yan, Z. Coordinated expression of dopamine
receptors in neostriatal medium spiny neurons. J. Neurosci. 16, 65796591
We thank Ann-Christine Eklf for experimental assistance. This work was supported (1996).
by grants from the Swedish Medical Research Council (H.B. and A.A.), Mrta and 26. Belusa, R. et al. Mutation of the protein kinase C phosphorylation site on rat
Gunnar V. Philipsons Foundation (H.B.), NIH grants MH40899 and DA 10044 alpha1 Na+,K+-ATPase alters regulation of intracellular Na+ and pH and
influences cell shape and adhesiveness. J. Biol. Chem. 272, 2017920184
(P.G.) and a grant from Stiftelsen Frimurare Barnhuset I Stockholm (O.A.). (1997).
27. Borin, M. L., Goldman, W. F. & Blaustein, M. P. Intracellular free Na+ in
resting and activated A7r5 vascular smooth muscle cells. Am. J. Physiol. 264,
RECEIVED 29 NOVEMBER 1999; ACCEPTED 3 JANUARY 2000 C15131524 (1993).
28. Satoh, H. et al. Regulation of [Na+]i and [Ca2+]i in guinea pig myocytes: dual
loading of fluorescent indicators SBFI and fluo 3. Am. J. Physiol. 266,
1. Missale, C., Nash, S. R., Robinson, S. W., Jaber, M. & Caron, M. G. Dopamine H568576 (1994).
receptors: from structure to function. Physiol. Rev. 78, 189225 (1998). 29. Peng, L., Martin-Vasallo, P. & Sweadner, K. J. Isoforms of Na,K-ATPase alpha
2. Chase, T. N., Oh, J. D. & Blanchet, P. J. Neostriatal mechanisms in Parkinsons and beta subunits in the rat cerebellum and in granule cell cultures.
disease. Neurology 51, S3035 (1998). J. Neurosci. 17, 34883502 (1997).
3. Grace, A. A., Gerfen, C. R. & Aston-Jones G. Catecholamines in central 30. Pietrini, G., Matteoli, M., Banker, G. & Caplan M. J. Isoforms of the Na,K-
neurvous system. Overview. Adv. Pharmacol. 42, 655670 (1998). ATPase are present in both axons and dendrites of hippocampal neurons in
4. Knable, M. B. & Weinberger, D. R. Dopamine, the prefrontal cortex and culture. Proc. Natl. Acad. Sci. USA 89, 84148418 (1992).
schizophrenia. J. Psychopharmacol. 11, 123131 (1997). 31. Jewell, E. A. & Lingrel, J. Comparison of the substrate dependence properties
5. Egan, M. F. & Weinberger, D. R. Neurobiology of schizophrenia. Curr. Opin. of the rat Na,K-ATPase alpha1, alpha 2 and alpha 3 isoforms expressed in
Neurobiol. 7, 701707 (1997). HeLa cells. J. Biol. Chem. 266, 1692516930 (1991).
6. Castellanos, F. X. Toward a pathophysiology of attention-deficit/hyperactivity 32. Zahler, R., Zhang, Z.-T., Manor, M. & Boron, W. F. Sodium kinetics of Na,K-
disorder. Clin. Pediatr. (Phila.) 36, 381393 (1997). ATPase isoforms in intact transfected HeLa cells. J. Gen. Physiol. 110,
7. Swanson, J., Castellanos, F. X., Murias, M., LaHoste, G. & Kennedy J. 201213 (1997).
Cognitive neuroscience of attention deficit hyperactivity disorder and 33. Johnson, S. W., Seutin, V. & North, R. A. Burst firing in dopamine neurons
hyperkinetic disorder. Curr. Opin. Neurobiol. 8, 263271 (1998). induced by N-methyl-D-aspartate: role of electrogenic sodium pump. Science
8. Self, D. W., Barnhart, W. J., Lehman, D. A. & Nestler, E. J. Opposite 258, 665667 (1992).
modulation of cocaine-seeking behavior by D1- and D2-like dopamine 34. Yu, X. M. & Salter, M. W. Gain control of NMDA-receptor currents by
receptor agonists. Science 27, 15861589 (1996). intracellular sodium. Nature 396, 469474 (1998).
9. Koob, G. F. & Le Moal, M. Drug abuse: hedonic homeostatic dysregulation. 35. Bockaert, J. et al. Primary culture of striatal neurons: a model of choice for
Science 278, 5258 (1997). pharmacological and biochemical studies of neurotransmitter receptors.
10. Gerfen, C. R. & Keefe, K. A. Neostriatal dopamine receptors. Trends Neurosci. J. Physiol. (Paris) 81, 219227 (1986).
17, 23 (1994). 36. OConnell, D. P. et al. Localization of dopamine D1A receptor protein in rat
11. Bloch, B. & Le Moine, C. Neostriatal dopamine receptors. Trends Neurosci. kidneys. Am. J. Physiol. 268, F11851197 (1995).
17, 34 (1994). 37. Hemmings, H. C. Jr. & Greengard, P. DARPP-32, a dopamine- and adenosine
12. Surmeier, D. J., Reiner, A., Levine, M. S. & Ariano, M. A. Neostriatal 3:5-monophosphate-regulated phosphoprotein: regional, tissue, and
dopamine receptors. Trends Neurosci. 17, 45 (1994). phylogenetic distribution. J. Neurosci. 6, 14691481 (1986).

articles
Growth cone and dendrite dynamics

in zebrafish embryos: early events in
synaptogenesis imaged in vivo
James D. Jontes, JoAnn Buchanan and Stephen J. Smith
Department of Molecular and Cellular Physiology, Stanford University School of Medicine, Stanford, California 94305-5435, USA
Correspondence should be addressed to S.J.S. (sjsmith@leland.stanford.edu)
We used time-lapse fluorescence microscopy to observe the growth of Mauthner cell axons and their
postsynaptic targets, the primary motor neurons, in spinal cords of developing zebrafish embryos.
Upon reaching successive motor neurons, the Mauthner growth cone paused briefly before continu-
ing along its path. Varicosities formed at regular intervals and were preferentially associated with
the target regions of the primary motor neurons. In addition, the postsynaptic motor neurons
showed highly dynamic filopodia, which transiently interacted with both the growth cone and the
axon. Both Mauthner cell and motor neurons were highly active, each showing motility sufficient to
initiate synaptogenesis.
During development of the central nervous system, a mass of Analysis of brain development in vivo involving mainly electron
undifferentiated cells develops into a precisely organized system microscopy (EM) reveals many synapses on dendritic filopo-
that acquires information, processes that information and uses it dia2527; however, such static studies do not reveal cellular dynam-
to generate appropriate behavior. Formation of central chemical ics or detailed interactions between cells.
synapses, an important part of this self-organization, is poorly We used time-lapse confocal laser-scanning microscopy and
understood. Here we take advantage of the small size, optical two-photon laser-scanning microscopy of neurons labeled in liv-
transparency and rapid development of the zebrafish (Danio rerio) ing zebrafish embryos to characterize the cell motility events
to study central synaptogenesis in an intact vertebrate embryo. involved in making initial contacts. We found that the Mauthn-
The Mauthner cells (M-cells) are large, paired, identified hind- er growth cone moved rapidly and relatively smoothly down the
brain neurons that are well characterized in the teleost fishes17 and spinal cord and seemed to interact transiently with successive
can be identified on the basis of size, position and morphology4,5 target cells. Primary motor neurons interacted with the nearby
(Fig. 1a and b). The M-cells mediate the escape response, a rapid axon via dynamic filopodia. These findings are consistent with
movement away from auditory, vibrational or tactile stimuli1,2,7. an active involvement of both pre- and postsynaptic elements in
Each M-cell axon crosses the ventral midline to extend down the establishing synaptic contacts. In addition, regularly spaced vari-
contralateral spinal cord to synapse on primary motor neurons cosities formed along the Mauthner axon in close proximity to
(Pmns) in each spinal hemisegment2,7 (Fig. 1a and c). Serially motor neurons; these varicosities may represent nascent presy-
repeating, bilaterally paired sets of three primary motor neurons, naptic boutons.
the RoP (rostral primary), MiP (middle primary) and CaP (cau-
dal primary)811, are activated during the escape reflex and fast RESULTS
swimming12. In goldfish and tench, the M-cellPmn synapse forms Growth of the Mauthner cell axon
between a dorsal collateral of the M-cell axon and a spine-like struc- To observe migration of the M-cell growth cone down the spinal
ture on the ventral dendrite of the Pmn2,13,14 (Fig. 1d). cord, lipophilic dye, DiD (confocal imaging) or DiO (two-pho-
Time-lapse imaging in vivo is used to study growth cone ton imaging)28, was iontophoretically applied to the cell bod-
dynamics and axonal pathfinding of the retinotectal projections ies of the M-cell 1820 hours after fertilization (Fig. 1b).
in Xenopus laevis1519 and zebrafish20, innervation of axial muscle Consistent with estimations from fixed embryos5,9, the Mauth-
by the Pmns11 and embryonic olfactory neuron pathfinding in ner growth cone moved at a mean rate of 99.6 22.7 m per
zebrafish21, as well as the effect of collapsin 1 on sensory neuron hour (n = 12, 28C; Fig. 2). In general, the Mauthner growth
growth cones22; however, these studies do not address the dynam- cone moved smoothly and steadily, but time-lapse sequences
ics of pre- or postsynaptic cells as they contact one another. variably showed slowing or brief pauses every 1015 m; these
During synapse formation in vitro, dendritic filopodia of cen- were more pronounced in some cases (Fig. 2b, top and mid-
tral neurons show rapid and extensive dynamics23,24, suggesting an dle) than in others (Fig. 2b, bottom). In all sequences exam-
active role in synaptogenesis. The number of filopodia gradual- ined, varicosities formed with variable timing along the nascent
ly decreases as the number of dendritic spines increases, sug- axon (Fig. 2c), often just behind the growth cone but not always
gesting that filopodia may represent precursors to the more stable in sequence. Thus, as it moved, the growth cone seemed to leave
spines23,24. However, even in a tissue slice, the similarity of events a trail of axonal thickenings spaced at 13.5 4.3 m (67 vari-
in culture to those in an intact, developing organism is unknown. cosities from 6 axons).

articles
Fig. 1. Illustration of the Mauthner cellprimary motor

a b neuron system. (a) Schematic diagram of the embryonic
zebrafish hindbrain and spinal cord showing the overall
relationship between the Mauthner cells and the serially
repeating primary motor neurons. OV, otic vesicle; Pms,
primary motor neurons; My, myotome; M, Mauthner cell;
SC, spinal cord. (b) Confocal image of bilaterally symmet-
ric Mauthner neurons in rhombomere 4, which have been
labeled with the lipophilic dye, DiD. The image is a super-
position of the labeled cells and a DIC image of a more
dorsal plane, chosen to give a clearer view of the anatom-
ical relationships. An additional, unidentified neuron with
an ascending growth cone has also been labeled on the
c d upper side of the hindbrain. (c) Confocal image of three
DiD-labeled primary motor neurons, RoP, MiP and CaP.
(d) Schematic representation of the mature Mauther-pri-
mary motor neuron synapse, based on ultrastructural
analyses in adult goldfish and tench. The basic geometry
holds true in the zebrafish: the axon of the motor neuron
passes medially and ventrally to the Mauthner axon
before leaving the spinal cord through the ventral root.

DS, dendritic spine; MA, Mauthner axon; MC, Mauthner
collateral; Mn, motor neuron axon; MnVD, motor neuron
ventral dendrite.
Because the varicosities we observed may have been nascent od investigated (data not shown). However, we cannot rule out
presynaptic termini, we investigated their structure using EM. short-term changes corresponding to the growth cone pauses we
Micrographs of cross-sectioned embryonic spinal cord revealed observed.
long, narrow (0.6 m) segments of Mauthner axon largely
devoid of synaptic vesicles (Fig. 2d). In addition, short areas of Double-label time-lapse analysis
the axon were of larger diameter (1.2 m) and contained small To examine initial contact and possible responses between cells
(40 nm), clear synaptic vesicles (Fig. 2e). Accumulation of during synaptogenesis, we labeled both Mauthner axons and
synaptic vesicles in Mauthner axon swellings suggests that these motor neurons and collected time-lapse sequences. Both the
varicosities were sites of early synaptic contact. Mauthner growth cone and the motor neurons behaved essen-
tially the same as in single-label experiments; motor neurons
Labeling of primary motor neurons showed dynamic filopodia, and growth cones moved rapidly and
We injected the ventrolateral myotomes of segments 1215 smoothly.
with DiD or DiO to label motor neurons. An injection was gen- As the Mauthner growth cone first approached and became
erally confined to a single hemisegment, labeling only the neu- apposed to a primary motor neuron, its path crossed close to
rons innervating it. Because of the injection placement, the the motor neurons ventral dendrites. In some cases, the growth
RoP and CaP motor neurons were preferentially labeled, cones seemed to transiently interact with the target regions of
although occasionally the MiP motor neuron could be identi- the motor neurons (Fig. 4a). As in single-label experiments,
fied (Fig. 1c). Before each time-lapse sequence, we used migra- analysis of motility of the M-cell growth cone revealed brief and
tion of the lateral line primordium to assess embryonic stage periodic pauses. Such behavior could either represent the spe-
to prim-stages 518 (ref. 29), roughly corresponding to 2433 cific recognition of targets, or it could simply represent the sto-
hours post-fertilization (hpf; at 28.5C). chastic nature of growth cone progress. To investigate this, we
Labeled motor neurons consistently showed variable numbers analyzed growth cone dynamics relative to labeled motor neu-
of characteristic filopodia extending from the cell bodies and ven- rons (Fig. 4bd). Averaging the growth cone motilities from nine
tral dendrites (Figs. 1c and 3). Time-lapse analysis revealed suitable time-lapse data sets smoothed away much of the tem-
extremely dynamic and variable protrusive/retractive behavior of poral variability, leaving only a marked discontinuity when the
filopodia, occurring on a time scale of seconds to minutes (Fig. 3). growth cone approached a motor neuron (Fig. 4d). This sup-
Filopodia were analyzed in 19 primary motor neurons for sev- ports the conclusion that the Mauthner growth cone interacts
eral basic properties, length, lifetime, density and rates of exten- transiently with each successive motor neuron, and that these
sion and of retraction. The lifetimes of individual filopodia interactions have a subtle but detectable effect on growth cone
seemed exponentially distributed with a time constant, , of dynamics. However, we observed no overt change in growth cone
approximately three minutes (Fig. 3b). The filopodium length structure, such as collapse or loss of filopodia.
(mean, 2.7 0.7 m; n = 571; Fig. 3c) typically varied from 1.3 to The axonal varicosities observed in time-lapse analysis of
4.0 m, but could reach lengths of 13 m. Density of filopodia labeled M-cell axons were also observed in double-label time-
on the ventral dendrites was highly variable, both over time in lapse experiments. Importantly, they formed preferentially near
the same cell and from cell to cell, and varied from 0.1 to 0.7 the ventral dendrites of motor neurons, over time associating
filopodia per m of dendrite. Rates of extension essentially with them very closely (Figs. 5 and 6). In 5 of 13 observations,
matched those of retraction (1.2 0.3 m per min). These para- an axonal varicosity formed within 5 minutes of the passing of
meters did not change significantly over the developmental peri- the growth cone, suggesting that the varicosity had arisen from

articles
Fig. 2. Time-lapse imaging of the Mauthner cell

growth cone. (a) DiO-labeled growth cone imaged a b
with two-photon microscopy. Image stacks (dorsal
view) were acquired every minute for one hour.
Distance (m)
This sequence shows the emergence of an axonal
varicosity that occurred well after the growth cone
had moved on. Each displayed image is a projection
from a stack of 16 optical sections. (b) Quantitative
analysis of the growth cones reveals transient
pauses not readily apparent from direct observa-
tion of time-lapse sequences. The occurrence and
periodicity of these pauses is variable, as shown in
this analysis of three different growth cones.
Changes in velocity of two growth cones are indi-
Distance (m)
cated by an apparent staircase pattern (b, top and
middle), whereas another growth cone seemed to
move more smoothly (b, bottom). (c) A dorsal
view of a DiD labeled M-cell axon imaged with a
confocal microscope showing periodicity of the
varicosities. (d) An electron micrograph of a
Mauthner axon from a cross-sectioned zebrafish

spinal cord at 28 hpf. The axon (arrow) is narrow
and contains only a small number of vesicles. (e) An
electron micrograph of another section from the
same embryo shown in (d), revealing an axonal
Distance (m)
varicosity in the same Mauthner axon. The axon
has a larger diameter and contains many synaptic
vesicles (arrow), as well as a dense core vesicle
(arrowhead) and a mitochondrion (M). Dorsal is to
the right, and scale bar is 0.25 m in (d, e).
c
the growth cone. However, in the remaining Time (min)

eight time-lapse sequences, either no varicos-
ity formed or one appeared long after the
d e
growth cone had passed. An axonal varicosi-
ty formed near the ventral dendrite of a CaP
motor neuron (Fig. 6).
Dendritic filopodia were quite active near
the M-cell axon and interacted with it fre-
quently, touching both the passing growth
cone and the axonal shaft (Fig. 7). These
interactions were transient, lasting only a few
minutes before retraction of the filopodia.
The relatively short distance ( 3 m)
between the M-cell axon and motor neurons,
placed the axon well within reach of the filopodia (mean length, Events in synaptogenesis
2.7 m; maximum, 13 m; Fig. 7). A conventional expectation is that an active axonal growth cone
initiates synaptogenesis with a passive postsynaptic target. This
DISCUSSION mechanism is most vividly illustrated by the neuromuscular
Mauthner growth cone junction, at which the motor growth cone grows up to and ter-
The large Mauthner axon growth cone moved rapidly (100 m minates on a muscle fiber before developing into a presynaptic
per hour at 28C; Fig. 2a). Considering that the Mauthner bouton. In contrast with this configuration en terminale, many
growth cone is believed to pioneer the medial longitudinal fas- synapses in the CNS, including the M-cellPmn synapse, are
ciculus, surprisingly little exploratory behavior was observed; formed en passant. Like the en-terminale synapse, the en-pas-
growth cones largely proceeded along a direct path with few, if sant synapse might develop from a motile axonal growth cone.
any, deviations. This behavior is characteristic of a growth cone Alternatively, filopodia from the dendrite could be responsible
moving down a permissive corridor rather than exploring a com- for establishing contact between two neurons.
plex environment. Brief, periodic pauses presented an intrigu- If the Mauthner growth cone were responsible for establishing
ing feature of growth cone movement (Fig. 2c). These pauses each synapse as it descended the spinal cord, one might expect
were spaced at 1015 m, roughly the distance between successive the growth cone to show some overt evidence of such an inter-
Pmns in the spinal cord, suggesting interaction of the growth action. In the M-cellPmn system, we found some evidence for
cone with targets. Analysis of double-labeling experiments reveals growth conemotor neuron interactions. The growth cone
correlation of these brief reductions in growth cone velocity with migrated along a path that approached the motor neuron ven-
proximity to a motor neuron (Fig. 4). tral dendrite, but our analysis revealed only a brief pause when it

articles
Fig. 3. Time-lapse imaging of primary motor

neurons and filopodia dynamics. (a) Brief time- a
lapse sequences of two RoP motor neurons. Each
image is a projection of 2 (left column) or 14
(right column) sections collected by confocal
microscopy of a DiD-labeled cell and two-photon
microscopy of a DiO-labeled cell, respectively.
(b) Distribution of lifetimes of 571 filopodia from b
19 cells imaged by confocal microscopy (28C).
Percent filopodia
Fit to a single exponential reveals a filopodia life-
time of 3 min. (c) Distribution of filopodium
lengths for the same data set as in (b). The mean
length is 2.7 0.7 m.
came into contact with successive motor

neurons (Fig. 4). In addition, growth cone
Time (min)
structure or dynamics seemed to sustain lit-
tle or no perturbation, in contrast with

growth cone behavior affected by guidance
cues such as the collapsins/semaphorins,
which induce growth cone collapse and tem-
porary paralysis of motility30,31. If the M-cell
growth cone is responsible for establishing c
synaptic contact, the early events must occur
Percent filopodia
very rapidly and without significant disrup-
tion of growth cone structure or dynamics.
In time-lapse images, we consistently
observed formation of varicosities on Mau-
thner axons in the wake of growth cone
migration (Figs. 2, 5 and 6). Electron micro-
graphs showing synaptic vesicles within vari-
cosities along the Mauthner axon (Fig. 2d)
support the idea that varicosities represent Length (m)
nascent synapses. Similar swellings are
observed on developing corticorubral neu-
rons; EM reveals these varicosities as sites of
synaptic contact25. The formation of varicosities along the Mau- ily from motility of the growth cone or motility of the dendrite.
thner axon may, therefore, represent the early organization of Synaptogenesis, even in what seems a hardwired system like the
presynaptic boutons. Mauthnermotor neuron synapse, is probably a stochastic
Perhaps our most provocative and striking observation was process, depending on neither the growth cone nor dendritic
the presence of highly dynamic filopodia extending from the ven- filopodia exclusively. Moreover, possible redundancy in capaci-
tral dendrites of primary motor neurons, which interact with the ties of both the pre- and postsynaptic cell to initiate synaptoge-
M-cell axon (Fig. 3). Present throughout the period of synapto- nesis may be important for ensuring the formation of such a
genesis, these filopodia made contact with Mauthner axons and stereotyped synapse. Thus, further investigations may reveal
were enriched in regions of the motor neurons that ultimately examples of synaptogenesis resulting from action of both growth
receive large numbers of synaptic inputs. The varicosity shown in cones and dendritic filopodia.
Fig. 6 formed after filopodiaaxon interactions and was frequently
contacted by dendritic filopodia during the time-lapse sequence. Timing of synaptogenesis
However, it is unclear whether varicosity formation was actually Currently, our only marker for the presence of a nascent synapse
induced by contact with filopodia. Our observations are consis- is the formation of an axonal varicosity. By this criterion, there
tent with proposals that dendritic filopodia may be involved in can be a considerable and variable lag between the interaction of
initiating synaptogenesis and might represent developmental pre- the growth cone with the motor neuron and synapse formation.
cursors of spines23,24. Electron microscopy demonstrates dendrit- Zebrafish embryos respond to touch by 21 hpf; hindbrain
ic filopodia as frequent sites of early synaptic contact25,26, although lesions destroy this touch sensitivity32. Based on the timing of
spine synapses may develop primarily from shaft synapses26. axon extension by the reticulospinal neurons, the Mauthner cell
If synaptogenesis does not result directly from growth cone was suggested to participate in this early behavior32. The Mau-
motility, then dendritic filopodia may actively help to bridge the thner growth cone reaches the rostral spinal cord by 20 hpf,
gap between the two cells and to initiate contact. However, it may indicating a lag of 1 h between growth cone passage and initia-
be unrealistically simplistic to assume that synaptogenesis results tion of synaptic function. The time frame for synapse formation
exclusively from the actions of either the growth cone or of the established by the behavioral studies is consistent with our obser-
dendritic filopodia. Contact between the cells may be the criti- vations of formation of axonal varicosities with a variable lag fol-
cal event, and it may be irrelevant if this contact results primar- lowing passage of Mauthner growth cones.

articles
Fig. 4. Simultaneous two-photon imaging of the a b

Mauthner growth cone and primary motor neurons. (a)
Maximum-intensity projections from a representative
time-lapse sequence during which an M-cell growth cone
migrated past a CaP motor neuron. The growth cone
Distance (m)
was highly active and transiently interacted with the CaP
cell, one of its synaptic partners. The growth cone
moved past the CaP cell without collapsing, stalling or
significantly altering its morphology. (bd) Analysis of
growth cone migration with respect to primary motor
neurons. To assess whether growth cone deceleration
was related to growth conetarget interactions, we mea-
sured rate of growth cone motility relative to the posi-
tions of labeled motor neurons. In single-label
experiments, growth cone motility varied in overall
c
speed as well as the degree of saltation. (b) This growth
cone distinctly paused as it passed the labeled motor
Distance (m)
neuron. (c) The growth cone represented in this graph
(shown in a) moved without clear periodic changes in
rate and did not pause near the motor neuron. (d) To fil-
ter out stochastic fluctuations and thus reveal whether

the rates of growth cone motility indicated any target
cell interactions, data from nine time-lapse sequences
were averaged after alignment with respect to the
labeled motor neurons. The rate of growth cone
progress distinctly fluctuates near the motor neurons.
d
Distance (m)
METHODS
Labeling of neurons. Zebrafish embryos were collected from
a laboratory colony33 and were allowed to develop at 28.5C
in a beaker of fresh tank water. At times ranging from 2435
hpf (motor neuron labeling) or 1821 hpf (Mauthner label-
ing), embryos were de-chorionated and mounted on glass
depression slides in 12% agarose with 0.05% tricaine (3-
amino benzoic acid ethyl ester; Sigma). DiD (1,1-dioctade-
cyl-3,3,3,3-tetramethylindocarbocyanine; Molecular Probes, Time (min)
Eugene, Oregon; 0.51% in dimethylformamide) or DiO
Imaging. Labeled embryos were remounted on cover slips in 12%
(3,3-dioctadecyloxacarbocyanine perchorate; Molecular Probes; 0.51%
agarose with 0.05% tricaine and imaged by confocal or two-photon
in dimethylformamide) was pressure injected into the ventral myotome
microscopy. The embryos were imaged either on a confocal laser-scan-
of anesthetized embryos or iontophoretically applied (0.51% in 50%
ning microscope using a Zeiss 20/NA 0.75 dry objective, or on a two-
dimethylformamide/50% ethanol) to the cell bodies of Mauthner cells.
photon laser-scanning microscope using a Zeiss 63/NA 0.9
Embryos were then allowed to develop for several hours, allowing time for
water-immersion objective. (Both microscopes were designed and built
growth cones to invade the spinal cord. Well-labeled growth cones were
by S.J.S., software by N.E. Ziv.) The 633-nm line of a helium-neon laser
further selected either for single-label time-lapse analysis, or for motor
(0.51 mW at the specimen) was used to excite the DiD, and fluores-
neuron labeling and double-label time-lapse analysis.
cence passed through a 650-nm long-pass filter before detection. Z-stacks
Electron microscopy. Embryos (2430 hpf) were fixed for 1 h at room
temperature in 3% glutaraldehyde, 2% formaldehyde, 1% acrolein and Fig. 5. Projections of the
1% DMSO in PBS and then further processed for electron microscopy Mauthner axon and sets of
in a Pelco 3450 Laboratory Microwave Processor (Ted Pella, Redding, primary motor neurons. In
California). Embryos were washed with 0.1 M cacodylate buffer for 5 both panels, an image stack
min at room temperature and post-fixed with 1% osmium tetroxide in was collected a few seg-
0.1 M cacodylate buffer containing 0.8% potassium ferricyanide on the ments rostral to the caudally
bench for 5 min on ice, then given 2 short (10 s) pulses in the microwave. directed growth cone. In
The embryos were washed in distilled water for 5 min, stained en bloc each case, distinct varicosi-
with 2% aqueous uranyl acetate for 20 min and given another short pulse ties were associated with the
(10 s) in the microwave. They were then dehydrated in an ascending ventral dendrites of primary
ethanol series (50%, 70%, 95%, 100% 3) in the microwave oven for 10 motor neurons. Secondary
s each. The embryos were infiltrated in the microwave in 1:1, 2:1 (resin: motor neurons (asterisks;
ethanol) for 5 min each, and in 100% resin for 5 min using Embed 812 characterized by their small
resin (Electron Microscopy Sciences, Fort Washington, Pennsylvania). size, round cell bodies and
The embryos were flat embedded and polymerized in a 60C oven ventral position) are present
overnight. Thin sections were then cut using a Reichert-Jung Ultracut E in both panels. In the bottom
ultramicrotome (Leica, Deerfield, Illinois). After staining with 6% uranyl panel, one varicosity does
acetate and 1% lead citrate, the sections were imaged in a Philips CM12 not seem to be associated
transmission electron microscope (FEI/Philips, Hillsboro, Oregon) oper- with a primary motor neuron. However, it is in the correct position
ating at an accelerating voltage of 80 kV. to be associated with an unlabeled MiP motor neuron.

articles
Fig. 7. Interaction of dendritic filopodia with the Mauthner axon. This

two-photon time-lapse sequence illustrates the interaction of a
Mauthner growth cone with a RoP motor neuron as these two cells first
came into contact. The top panel shows the start of this time-lapse, as
the Mauthner growth cone (gc) first reached the RoP ventral dendrite.
The bottom panels are a segment of the time-lapse sequence (collected
at 1-min intervals) beginning 40 min after the initial cellcell contact.
Each image is a maximum-intensity projection of 14 sections collected at
1-m steps. Numerous dendritic filopodia extended and retracted from
the RoP ventral dendrite, many of which interacted transiently with the
Mauthner axon. During this experiment, no stable cellcell contacts
were observed. Because a synapse will ultimately form between the
Mauthner axon and the RoP ventral dendrite, it is possible that a
filopodium may be responsible for initiating synaptogenic contact.
(6 sections spaced at 1 m) were collected every min over periods of 30

min2 h. To image DiO with two-photon microscopy, the Ti:sapphire
laser (Mira 900, Coherent, Santa Clara, California) was tuned to 900910
nm. Every min, we collected 1018 sections spaced at 1 m.
Deconvolution. Data sets were deconvolved using the commercial software

AutoDeblur (AutoQuant Imaging, Watervliet, New York) using an algo-
rithm for maximum-likelihood estimation that did not require measure-
ment of the point spread function. Maximum-intensity projections resulting
Fig. 6. Formation of an axonal varicosity. This two-photon time-lapse from five iterations of deconvolution were used to make the figures. In gen-
sequence shows the formation of an axonal varicosity in close association eral, image processing produced neither new specimen features nor loss of
with a CaP motor neuron. The images are maximum-intensity projections observed image detail. However, background noise was strongly suppressed,
of image stacks displayed at five-min intervals. Initially, the axon swelled, and image features such as axons and filopodia were sharpened.
forming a large, elongated varicosity close to the CaP cell. Over time, the
varicosity shrank and became more spherical. The arrow in the 5-min Analysis of filopodia dynamics. The highest quality time-lapse sequences
panel marks the future site of the stable varicosity, emphasized by a second were selected for analysis. A mouse-driven program (written by N.E. Ziv)
arrow in the 45-min panel. Throughout the time lapse, there was extensive was used to measure the lengths of filopodia in each frame of each time-
filopodia activity on the CaP cell near the developing varicosity. lapse sequence.

articles
Filopodia lengths. Lengths were defined as the distance from the base of 6. Metcalfe, W. K., Mendelson, B. & Kimmel, C. B. Segmental homologies
the filopodium to its tip. Only filopodia 3 pixels (0.7 m) were scored, among reticulospinal neurons in the hindbrain of zebrafish larva. J. Comp.
Neurol. 251, 147159 (1986).
as shorter filopodia could not be identified reliably. Lengths were mea- 7. Fetcho, J. R. & Faber, D. S. Identification of motoneurons and interneurons
sured from maximum-intensity projections. Thus the values we present in the spinal network for escapes initiated by the Mauthner cell in goldfish.
slightly underestimate (by 13%) the actual three-dimensional lengths34. J. Neurosci. 8, 41924213 (1988).
Rates. The rate of extension/retraction for a given filopodium was defined 8. Myers, P. Z. Spinal motoneurons of the larval zebrafish. J. Comp. Neurol. 236,
as the change in unit length between successive frames. Filopodial exten- 555561 (1985).
9. Myers, P. Z., Westerfield, M. & Eisen, J. S. Development and axonal
sion/retraction rate for a given motor neuron was calculated as the mean outgrowth of identified motoneurons in the zebrafish. J. Neurosci. 6,
rate of extension or retraction for all filopodia observed on that neuron. 22782289 (1986).
10. Westerfield, M., McMurray, J. V. & Eisen, J. S. Identified motoneurons and
Lifetimes. The lifetime of an individual filopodium was defined as length their innervation of axial muscles in the zebrafish. J. Neurosci. 6, 22672277
(1986).
of time for which its length was measureable. A filopodium extending 11. Eisen, J. S., Myers, P. Z. & Westerfield, M. Pathway selection by growth cones
from a position at which one had previously disappeared was counted of identified motoneurones in live zebra fish embryos. Nature 320, 269271
as a second filopodium. Lifetime data were pooled and plotted in a his- (1986).
togram (bin size, 2 min). Population lifetime was calculated from the 12. Liu, D. W. & Westerfield, M. Function of identified motoneurones and co-
time constant, , of a single-exponential fit to the histogram. ordination of primary and secondary motor systems during zebra fish
swimming. J. Physiol. (Lond.) 403, 7389 (1988).
13. Celio, M. R., Gray, E. G. & Yasargil, G. M. Ultrastructure of the Mauthner
Densities. Filopodia density was calculated as the number of filopodia axon collateral and its synapses in the goldfish spinal cord. J. Neurocytol. 8,
per unit length for each motor neuron averaged over the duration of the 1929 (1979).
time-lapse sequence. 14. Yasargil, G. M. & Sandri, C. Topography and ultrastructure of commisural
interneurons that may establish reciprocal inhibitory connections of the

Mauthner axons in the spinal cord of the tench, Tinca tinca L. J. Neurocytol.
Analysis of growth cone motility. For each frame of a time-lapse 19, 111126 (1990).
sequence, we recorded the leading edge of the growth cones lamel- 15. Harris, W. A., Holt, C. E. & Bonhoeffer, F. Retinal axons with and without their
lipodium. In double-label experiments, growth cone velocities and somata, growing to and arborizing in the tectum of Xenopus embryos: a time-
movement analyses were generated from the position of the growth cone lapse video study of single fibres in vivo. Development 101, 123133 (1987).
16. ORourke, N. A. & Fraser, S. E. Dynamic changes in optic fiber terminal
relative to the position of the ventral dendrite of the labeled motor neu- arbors lead to retinotopic map formation: an in vivo confocal microscopic
ron. After aligning to labeled motor neurons, suitable data sets were aver- study. Neuron 5, 159171 (1990).
aged. As only those with 15 time points on either side of the motor 17. ORourke, N. A., Cline, H. T. & Fraser, S. E. Rapid remodeling of retinal
neuron were included, 9 of 13 data sets were used. Although two-pho- arbors in the tectum with and without blockade of synaptic transmission.
ton-imaged growth cones had reduced and more variable velocities (68.5 Neuron 12, 921934 (1994).
18. Witte, S., Stier, H. & Cline, H. T. In vivo observations of timecourse and
29.8 m per h) that may be attributed to the 6C-cooler ambient tem- distribution of morphological dynamics in Xenopus retinotectal axon arbors.
perature of the two-photon microscope stage (22C) compared with J. Neurobiol. 31, 219234 (1996).
that of the warmer confocal microscope (28C), this temperature dif- 19. Wu, G. Y. & Cline, H. T. Stabilization of dendritic arbor structure in vivo by
ference probably did not significantly affect our results. All embryos were CaMKII. Science 279, 222226 (1998).
20. Kaethner, R. J. & Stuermer, C. A. Dynamics of terminal arbor formation and
incubated at a constant 28C except during injections and during the target approach of retinotectal axons in living zebrafish embryos: a time-
12 h periods of the time-lapse experiments. Additionally, apart from lapse study of single axons. J. Neurosci. 12, 32573271 (1992).
the enhanced resolution and reduced growth cone rate observed with 21. Dynes, J. L. & Ngai, J. Pathfinding of olfactory neuron axons to stereotyped
the two-photon microscope, all the phenomena we report here were glomerular targets revealed by dynamic imaging in living zebrafish embryos.
observed on both microscopes at both temperatures. Neuron 20, 10811091 (1998).
22. Shoji, W., Yee, C. S. & Kuwada, J. Y. Zebrafish semaphorin Z1a collapses
specific growth cones and alters their pathway in vivo. Development 125,
Note: Movies of these data can be found on the Nature Neuroscience web site 12751283 (1998).
(http://neurosci.nature.com/web_specials/). 23. Dailey, M. E. & Smith, S. J. The dynamics of dendritic structure in developing
hippocampal slices. J. Neurosci. 16, 29832994 (1996).
24. Ziv, N. E. & Smith, S. J. Evidence for a role of dendritic filopodia in
ACKNOWLEDGEMENTS synaptogenesis and spine formation. Neuron 17, 91102 (1996).
25. Saito, Y. et al. Developing corticorubral axons of the cat form synapses on
Susan Pike conducted initial work on this project. We thank S. Pike and faculty filopodial dendritic protrusions. Neurosci. Lett. 147, 8184 (1992).
members of the MBL Neural Development and Genetics of the Zebrafish course 26. Fiala, J. C., Feinberg, M., Popov, V. & Harris, K. M. Synaptogenesis via
dendritic filopodia in developing hippocampal area CA1. J. Neurosci. 18,
(Woods Hole, Massachusetts) for help and advice. J.D.J. is a fellow of the Helen
89008911 (1998).
Hay Whitney Foundation. This work was supported by NIH grants to S.J.S. 27. Saito, Y., Song, W.-J. & Murakami, F. Preferential termination of
corticorubral axons on spine-like dendritic protrusions in developing cat.
J. Neurosci. 17, 87928803 (1997).
RECEIVED 29 NOVEMBER; ACCEPTED 23 DECEMBER 1999 28. Honig, M. G. & Hume, R. I. Fluorescent carbocyanine dyes allow living
neurons of identified origin to be studied in long-term cultures. J. Cell Biol.
103, 171187 (1986).
1. Yasargil, G. M. & Diamond, J. Startle-response in teleost fish: an elementary 29. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. & Schilling, T. F. Stages
circuit for neural discrimination. Nature 220, 241243 (1968). of embryonic development of the zebrafish. Dev. Dyn. 203, 253310 (1995).
2. Diamond, J. in Fish Physiology Vol. 5 (eds. Hoar, W. S. & Randall, D. J.) 30. Luo, Y., Raible, D. & Raper, J. A. Collapsin: a protein in brain that induces the
265346 (Academic, New York, 1971). collapse and paralysis of neuronal growth cones. Cell 75, 217227 (1993).
3. Kimmel, C. B., Powell, S. L. & Metcalfe, W. K. Brain neurons which project to 31. Kolodkin, A. L. et al. Fasciclin IV: sequence, expression, and function during
the spinal cord in young larvae of the zebrafish. J. Comp. Neurol. 205, growth cone guidance in the grasshopper embryo. Neuron 9, 831845 (1992).
112127 (1982). 32. Saint-Amant, L. & Drapeau, P. Time course of the development of motor
4. Mendelson, B. Development of reticulospinal neurons of the zebrafish. I. behaviors in the zebrafish embryo. J. Neurobiol. 37, 622632 (1998).
Time of origin. J. Comp. Neurol. 251, 160171 (1986). 33. Westerfield, M. The Zebrafish Book (Univ. of Oregon Press, 1995).
5. Mendelson, B. Development of reticulospinal neurons of the zebrafish. II. 34. Maletic-Savatic, M., Malinow, R. & Svoboda, K. Rapid dendritic
Early axonal outgrowth and cell body position. J. Comp. Neurol. 251, 172184 morphogenesis in CA1 hippocampal dendrites induced by synaptic activity.
(1986). Science 283, 19231927 (1999).

articles
Enrichment induces structural changes

and recovery from nonspatial memory
deficits in CA1 NMDAR1-knockout mice
Claire Rampon, Ya-Ping Tang, Joe Goodhouse, Eiji Shimizu, Maureen Kyin and Joe Z. Tsien
Department of Molecular Biology, Princeton University, Washington Road, Princeton, New Jersey 08540-1014, USA
The first two authors contributed equally to this work.
Correspondence should be addressed to J.Z.T. (jtsien@molbio.princeton.edu)
We produced CA1-specific NMDA receptor 1 subunit-knockout (CA1-KO) mice to determine the

NMDA receptor dependence of nonspatial memory formation and of experience-induced structural

plasticity in the CA1 region. CA1-KO mice were profoundly impaired in object recognition, olfactory
discrimination and contextual fear memories. Surprisingly, these deficits could be rescued by enrich-
ing experience. Using stereological electron microscopy, we found that enrichment induced an
increase of the synapse density in the CA1 region in knockouts as well as control littermates.
Therefore, our data indicate that CA1 NMDA receptor activity is critical in hippocampus-dependent
nonspatial memory, but is not essential for experience-induced synaptic structural changes.
Comparative anatomical studies of hippocampal cytoarchitecture ing in adult rats16. Based on findings of NMDA receptor function
reveal a selective expansion of the CA1 region in the hippocampus in the developing brain17,18, it is postulated that NMDA receptors
in primates with respect to rodents and in humans with respect might be required for behavioral experience-induced structural
to monkeys, suggesting that this subregion is important in human plasticity in adult brain. However, new dendritic spines can form
hippocampal function1. Damage to the hippocampus or the CA1 on mature hippocampal neurons in vitro in the absence of synap-
subregion in humans leads to deficits in memories of people, tic activity19, and LTP has no effect on synapse number in the CA1
objects, places and events (declarative memory)24. Similar spa- region20. These findings imply that the role of the NMDA recep-
tial and nonspatial memory deficits are also observed in a variety tor in structural changes in adult brain might differ from that in
of laboratory animals with hippocampal lesions4,5. Recordings of the developing brain. By using unbiased stereological electron
neuronal activity in the CA1 subregion of rodents during behav- microscopy, we examined the structural changes in our CA1-KO
ioral tests show the importance of this subregion in the encoding mice before and after enriched experience.
of nonspatial information6. However, the molecular mechanisms
underlying these processes remain unknown. RESULTS
N-methyl-D-aspartate (NMDA) receptors are widely distrib- NMDAR1 knockout in the CA1 region
uted in the brain7 and are essential for the induction of major We confirmed the complete deletion of the NMDAR1 gene in
forms of long-term potentiation (LTP) and long-term depres- CA1-KO mice using in-situ hybridization histochemistry. No
sion (LTD)8,9. Enhanced NMDA receptor function in the fore- NMDAR1 mRNA was detected in the CA1 subregion (Fig. 1),
brain improves learning and memory10, indicating its crucial role indicating that the genetic deletion of NMDAR1 in young adult
in these processes. Using the Cre/loxP-mediated recombination mice was heritable after three years of breeding and was complete.
system, we developed a region-specific gene-knockout technique
to generate conditional-knockout mice in which NMDAR1, a key Nonspatial learning and memory deficits
subunit of NMDA receptor, was selectively deleted in the CA1 CA1-KO and control mice used in all behavioral tests were 24
subregion11. These CA1-specific NMDAR1 knockout mice lacked month-old littermates. We used three hippocampus-dependent
NMDA-mediated currents and plasticity in the CA1 region and behavioral tasks to assess associative, nonspatial memory functions.
were profoundly impaired in spatial memory tasks12,13. The hippocampus is important in the formation of recognition
In this study, our first goal was to examine the role of CA1 memory in both human patients and animals 35. However, little is
NMDA receptor activity in the formation of nonspatial memory known about its anatomical basis and its molecular and cellular
and the effects of enriched experience on these memory func- mechanisms21. We evaluated recognition memory with a novel-
tions. Our second goal was to examine whether enriched experi- object recognition-memory task10. During the training session, the
ence during adulthood could lead to structural changes in the total amount of time spent exploring two objects was 28.62 3.94
CA1 region, and whether CA1 NMDA receptor-mediated respons- seconds in control mice (n = 17) and 34.71 3.56 seconds in CA1-
es are required for such structural modifications. Synaptic struc- KO mice (n = 12), and no significant exploratory preference was
tural changes are assumed to be the anatomical substrate of found between control and CA1-KO mice (data not shown). These
long-term storage of learned experience14,15. For example, numbers observations indicate that these two types of mice have the same
of dendritic spines in the hippocampus increase after spatial learn- levels of motivation, curiosity, and interest in exploring novel objects.

articles
a b between control and CA1-KO

mice (1.723 0.248 g versus
1.584 0.378 g, respectively).
Because these two types of scent-
ed food were counterbalanced
throughout the experiments,
these results indicated that the
deficits observed in CA1-KO
mice were due neither to abnor-
mal food intake nor to a scent
bias per se.
We then examined nonspatial
memory using a contextual fear-
Fig. 1. Cre/loxP-mediated deletion of NR1 gene in CA1-KO mice (2.5 months old) is complete. An anti- 10
sense 42-mer oligonucleotide recognizing the NR1 gene was used for in situ hybridization. CTX, cortex; ST, conditioning task . In this hip-
striatum; DG, dentate gyrus. pocampus-dependent task 25 ,
animals learn to fear an environ-
ment by associating it with an
aversive stimulus (foot shock).
For retention tests, at 30 minutes, 2 hours and 24 hours after Control and CA1-KO mice significantly differed in contextual
the training (Fig. 2), one object was replaced by a novel object. freezing, but not in freezing responses immediately after the
As expected, control mice showed a significant preference for shocks, in retention tests at 2 hours and 24 hours (p < 0.05,
exploring the novel object during each retention test. In contrast, p < 0.01, respectively; Fig. 4a), indicating that CA1-KO mice were
CA1-KO mice showed only marginal preference for the novel impaired in contextual fear memory. However, in a cued-condi-
object (Fig. 2). A self-ANOVA between training and retention tioning test, which is hippocampus independent 25, freezing
tests revealed a significant difference in control mice (F3,64 = 7.744, responses of CA1-KO mice to a tone were similar to those of con-
p < 0.001) but not in CA1-KO mice. Repeated measures of trol mice in retention tests at 2 hours (data not shown) and 24
ANOVA on retention tests showed a highly significant difference hours (Fig. 4b). In addition, no abnormal nociceptive responses
between control and CA1-KO mice (F2,27 = 19.435, p < 0.001). A were found in CA1-KO mice: current required to elicit flinch-
post-hoc analysis using Dunnetts test demonstrated significance ing/running, jumping or vocalization in the CA1-KO mice was
of this difference in retention at 30 minutes, 2 hours and 24 hours the same as in control mice (data not shown). These data clear-
(p < 0.05, p < 0.01 and p < 0.01, respectively; Fig. 2). These results ly showed that hippocampus-dependent, but not hippocampus-
indicate profound deficits in novel object recognition memory independent, fear memory was impaired in CA1-KO mice.
in CA1-KO.
We then examined olfactory discrimination memory using a
social transmission of food preference 22. Rodents develop a
Exploratory preference (%)
preference for foods they have recently smelled on the breath of Control
other individuals 23. Lesions restricted to the hippocampus CA1 - KO
impair this kind of memory 24 hours after social interaction
(training)24. However, the molecular mechanisms underlying
this kind of memory remain unknown22,24. The task used here
consisted of three stages. First, animals became accustomed to
eating from a food cup placed on the cage floor. Second, in
training sessions, observer mice were allowed to interact with
demonstrator mice fed with either cinnamon-scented (1% per
weight) or cocoa-scented (2% per weight) food. Third, observ-
er mice were then tested by presentation of both scented foods, Training session
and consumption of each food was recorded. Control mice
showed a significantly higher preference for food smelled dur-
ing the training session over unsmelled food than did CA1-KO b
mice (F1,26 = 5.291; Fig. 3), indicating that olfactory-discrimi-
nation memory was impaired in CA1-KO mice. To ensure that
Control
this observation was not due to a difference in feeding behav- CA1 - KO
iors, we measured the total amount of food taken during the ** **
retention session; this quantity did not significantly differ *
Fig. 2. Impaired novel-object recognition memory in CA1-KO mice.

Recognition memory is expressed in terms of exploratory preference in
the retention tests. The memory for control (n = 17) or CA1-KO
(n = 12) mice was measured at three different time intervals after train-
30 min 2h 24 h
ing. The data are expressed as mean s.e.; *p < 0.05, **p < 0.01, deter-
mined by post-hoc analysis. We observed similar results in separate Retention interval
groups of CA1-KO and control mice.

articles
Fig. 3. Social transmission of food preference. Olfaction-discrimination

memory is expressed in terms of food preference. (a) The memory of a
either control (n = 15) or CA1-KO (n = 13) mice was measured 24 h Control
Food preference (%)

after training. (b) Food consumption data are expressed as mean s.e.
*p < 0.05 determined by one-way ANOVA.
CA1 - KO *
Enrichment-induced recovery of memory deficits

These results indicated a profound deficit in three forms of non-
spatial memory in CA1-KO mice. Enriched experience can sig-
nificantly improve performance in spatial maze tests in normal
animals26 and attenuate memory deficits in animals with hip- 24-hour retention
pocampal lesions27. However, it is not known whether enriched
experience can enhance performance of the animals in the three
nonspatial memory tasks we used, nor if it is able to rescue the b
memory deficits observed in CA1-KO mice.
Total food consumption (grams)

To address these issues, we evaluated nonspatial memory Control
using the same tasks with additional groups of adult animals after CA1 - KO
they were trained daily in an enriched environment for two

months. In the novel-object recognition task, we observed an
increase in preference for exploring novel objects in both con-
trol and CA1-KO mice (n = 14; n = 12, respectively; Fig. 5a). A
self-ANOVA between training and retention tests revealed sig-
nificant differences for control (F3,52 = 24.771, p < 0.001) and
CA1-KO mice (F3,44 = 3.797, p < 0.05), indicating that training
in the enriched-environment not only enhanced the performance
of normal mice but also attenuated the deficits in CA1-KO mice.
An integrated statistical analysis of animals with the same geno- enriched control and enriched CA1-KO mice (F2,24 = 11.619,
type confirmed the effects of training on memory in control mice p < 0.01), indicating that enriching experience could only par-
(F 1,29 = 7.674, p < 0.05) and in CA1-KO mice (F 1,22 = 6.674, tially rescue the deficits observed in the CA1-KO mice.
p < 0.05). However, a repeated measures ANOVA on retention In the test of social transmission of food preference, we found
tests still revealed a highly significant group difference between that enriched control animals (n = 15) did not increase food pref-
erence more than naive animals. This possibly reflected either a
a ceiling effect or a delicate balance between choosing the preferred
food and eating any food after 24 hours of food deprivation. How-
Control
Contextual freezing (%)
ever, food preference of enriched CA1-KO mice (n = 12) was dra-

CA1 - KO
** matically increased over that of naive CA1-KO animals
(F1,22 = 4.746, p < 0.05, Fig. 5b), indicating that the enriched expe-
rience completely rescued the memory deficits in the CA1-KO mice.
We also observed that the enrichment training significant-
ly enhanced contextual freezing in control mice (F1,25 = 7.366,
p < 0.05) and completely rescued the deficits observed in the
CA1-KO mice (F1,23 = 14.559, p < 0.001). No significant dif-
ference was found between enriched control and enriched
CA1-KO mice (Fig. 5c). In addition, cued freezing also dra-
Immediate freezing 24-hour retention matically increased following enriched experience in both con-
trol (F1,25 = 6.213, p < 0.05) and CA1-KO mice (F1,23 = 4.372,
p < 0.001; Fig. 5d). These results clearly indicate that the
enriched experience could enhance these two kinds of fear
b memory in control animals and could rescue the deficits in
CA1-KO mice.
Control
CA1 - KO
Cued freezing (%)
Fig. 4. Fear-conditioning tasks. (a) Contextual fear conditioning.

Memory is expressed as percent of freezing responses. Immediate learn-
ing indicates the freezing response during the 30 s immediately after
shock in the training session. Contextual fear memory in controls
(n = 14) and CA1-KO (n = 12) was measured 24 h after training.
(b) Cued fear conditioning. Memory is expressed as percentage of
freezing responses. There was no significant difference in proportion of
freezing responses either at pre-CS or retention test between control
mice (n = 14) and the CA1-KO mice (n = 12). All data are expressed as
mean s.e.; **p < 0.01, determined by one-way ANOVA. Similar results
Pre-CS CS
were obtained from another set of experiments.

articles
Fig. 5. Effects of enriched experience

a c on nonspatial memory in both control
and CA1-KO mice. (a) Enriched expe-
Contextual freezing (%)

Naive Naive
Enriched ** Enriched rience increased exploratory prefer-
* ence in both control (n = 14) and
* *** CA1-KO (n = 12) mice in the novel-
*
object recognition task. (b) Enriched
experience rescued memory deficits
of CA1-KO (n = 12) mice in the social
transmission of food preference. For
control mice, n = 15. (c) Enriched
Control CA1-KO Control CA1-KO experience enhanced contextual
24-hour retention freezing response in both control (n =
24-hour retention
13) and CA1-KO (n = 11) mice in the
b d fear-conditioning task. There was no
significant difference in immediate
Naive Naive freezing between the two groups
Enriched Enriched
Food preference (%)
(data not shown). (d) Enriched

Cued freezing (%)
* *
*
experience increased cued freezing in
the fear-conditioning task. There was
no significant difference in pre-CS

freezing or cued freezing between
control mice (n = 13) and CA1-KO
(n = 11) mice. All data are expressed
Control CA1-KO Control CA1-KO as mean s.e. *p < 0.05, ** p < 0.01,
***p < 0.001 in one-way ANOVA.
24-hour retention 24-hour retention
Enrichment-induced anatomical changes CA1 pyramidal cells of enriched animals. We observed a signifi-
What are the molecular and structural mechanisms underlying cantly higher density of dendritic spines in enriched contol mice
the enrichment-induced effects? Enriched experience promotes compared with naive animals (10.8 1.1 versus 8.1 0.8;
various biochemical and morphological changes in several brain p < 0.05; Fig. 6c). This is consistent with a previous report of
regions26,2831. Synaptic structural changes are believed to be the increased spine density on CA1 neurons in rats after enrich-
anatomical substrate of long-term storage of learned experience, ment16. Surprisingly, enrichment similarly increased spine density
and it is assumed that NMDA receptors are required for experi- in CA1-KO mice (10.0 0.3 versus 8.6 0.6; p < 0.05). Statistical
ence-dependent anatomical changes in the adult brain. As the analysis revealed no significant difference between enriched con-
first step toward the genetic dissection of the molecular mecha- tol and enriched CA1-KO animals (Fig. 6c).
nisms of experience-induced structural plasticity in the adult hip-
pocampal CA1 region, we used light and electron microscopy to
examine the effects of enrichment on anatomical changes in the a b
CA1 region and the role of the NMDA receptor in this effect.
Nissl-stained tissue of CA1-KO mice shows no gross anatomical
abnormality12. We used the Golgi-impregnation technique32 to ini-
tially assess dendritic structures. Golgi-impregnated CA1 pyrami-
dal neurons of control and CA1-KO naive animals presented similar
dendritic morphology (Fig. 6a and b). Quantitative analysis of den-
dritic-spine density on CA1 pyramidal cells revealed no significant
difference between naive control (8.1 0.8 spines per 10 m of api-
cal dendrite, mean s.e.) and naive CA1-KO mice (8.6 0.6 spines;
Fig. 6c). This suggests that conditional knockout of the NMDAR1
gene in the CA1 region, which occurs in the postnatal third and
fourth weeks, did not result in abnormal spine density. c
To characterize the effects of enrichment on dendritic spine
density, we then counted the number of spines on dendrites of
Dendritic spines per 10 m
Fig. 6. Enrichment-induced increase in CA1 spine density in both nor-

mal control and CA1-KO mice. Photomicrographs of Golgi-impregnated
apical dendritic segments of CA1 pyramidal neurons of naive control (a)
and naive CA1-KO (b) mice. (c) Bars represent a scatter plot of individ-
ual average numbers of dendritic spines per 10 m length of CA1 pyra-
midal dendritic segments. Diamonds represent group means s.e.
(n = 3 for each group). The sexes (M or F) and ages in months of animals
in each column (from top to bottom) are naive normal control (M, 3.5;
M, 4; M, 3.5), enriched normal control (M, 3.5; M, 5; M, 4.5), naive CA1- Naive Enriched Naive Enriched
KO (M, 4.5; M, 5; M, 4) and enriched CA1-KO (M, 5; M, 5; M, 4.5). Control CA1-KO
*p < 0.05, Mann-Whitney U-test.

articles
Fig. 7. Enrichment-induced increase in CA1 synapse density in both con-

trol and CA1-KO mice. Representative electron photomicrographs illus- a b
trating synapses (indicated by arrows) in the stratum radiatum of the CA1
region of naive controls (a), naive CA1-KO mice (b), enriched controls (c)
and enriched CA1-KO mice (d). Scale bars, 1 m. (e) Diamonds represent
group means s.e. of estimated synaptic densities in the stratum radiatum
of the CA1 region before and after enrichment, calculated with the stereo-
logical disector. Bars represent a scatter plot of the CA1 synaptic density in
individual animals in each group. The sexes (M or F) and ages in months of
animals in each column are (from top to bottom) naive control (M, 4.5; M,
5; M, 4; M, 4; M, 3; M, 3.5; M, 4), enriched control (M, 4.5; M, 5; M, 5.5; M,
4; M, 5), naive CA1-KO (M, 4; M, 4.5; F, 3.5; M, 4) and enriched CA1-KO c d
(M, 3.5; F, 4; M, 4.5; M, 5). *p < 0.05, determined by Mann-Whitney U-test.
These data indicate that enriched experience resulted in

changes in dendritic-spine density and suggest that this increase
in spine density can occur in the absence of NMDA receptor
activity. These results are consistent with the finding that mature
CA1 pyramidal neurons in vitro can grow spines in absence of

NMDA-receptor activation19. However, Golgi staining should be e
interpreted cautiously. First, this method does not reveal spines
hidden beneath or sitting above the dendritic segment, thus, spine
density measured this way probably underestimates the actual Synapses per 100 m3
number of spines. Second, some branched spine heads receive
no innervation33; therefore, spine density should not be simply
extrapolated to synapse density.
To address these concerns, we used electron microscopy to quan-
titatively analyze synapse density in the stratum radiatum of the
CA1 hippocampal region (Fig. 7ad). We ensured unbiased sam-
pling by using the stereological disector technique34,35. No signifi-
cant difference in synaptic density was observed between naive
control mice (76.9 1.9 synapses per 100 m3, mean s.e.) and Naive Enriched Naive Enriched
(n = 7) (n = 5) (n = 4) (n = 4)
naive CA1-KO mice (70.0 4.4 synapses, Fig. 7e). However, after Control CA1-KO mice
the enrichment training, control animals showed a significant
increase of CA1 synaptic density (93.8 5.3) over the naive con- To investigate the possible differential effect of enrichment and
trol group (p < 0.05). Synaptic density was also strikingly higher in NMDA receptor activity on changes in the density of synaptic sub-
trained CA1-KO mice (91.2 2.9) compared to the naive CA1-KO types, we examined the distribution of CA1 synapses in three major
group (p < 0.01, Fig. 7e). Furthermore, no significant difference in subclasses (non-perforated, perforated and shaft synapses) catego-
synaptic density was found between enriched control and enriched rized according to the profile of the postsynaptic density (PSD) in
CA1-KO mice. These data indicate that enriched experience pro- pairs of serial sections. The PSD was classified as non-perforated if
motes synaptic structural changes in the CA1 hippocampal region. the thickening was continuous or as perforated if the PSD was inter-
Furthermore, these results suggest that the NMDA receptor is not rupted by electron-lucent regions (Fig. 8a). Synapses with asym-
required for the increase of CA1 synaptic density. metric PSDs on dendritic shafts were identified using Grays criteria36
Fig. 8. Major subclasses of CA1

synapses in the CA1 region.
a b
(a, b) Electron micrographs illustrating
the ultrastructure of different types of
synapses in the stratum radiatum of
CA1. Scale bars, 0.5 m. (a) Non-perfo-
rated axospinous synapse (np) and per-
forated axospinous synapse (p).
(b) Axodendritic synapses involving
dendritic shafts (s). (c, d) Estimated dis-
tribution of CA1 synapses in the cate-
gories described above for control c d
Synapses per 100 m3
(c) and CA1-KO (d) animals, either

Synapses per 100 m3
Perforated
naive or reared in an enriched environ- Non-perforated
** *
Shaft
ment. Synaptic densities were calculated
per surface area and then expressed as
the number of synapses per 100 m3 by
using the stereological coefficient
obtained with the disector method. The
data are expressed as mean s.e.;
Naive control Enriched control Naive CA1-KO Enriched CA1-KO
*p < 0.05, Mann-Whitney U-test.

articles
and were counted separately from those on dendritic spines (Fig. 8b). unit 2B of the NMDA receptor also leads to overall enhancement
We found a significantly higher density of non-perforated synapses in of learning and memory10. Thus, changes in the composition (for
control mice after enrichment (77.1 5.2, mean s.e.) than in naive instance, ratio of NR2A to NR2B subunits) of the NMDA recep-
control animals (62.6 2.3, p < 0.05; Fig. 8c). Remarkably, enrich- tor complex may be a possible mechanism for the enrichment-
ment also significantly increased the density of non-perforated induced effects. Taken together, these findings indicate that
synapses in CA1-KO mice (78.3 5.4 versus 60.7 5.1; p < 0.05; Fig. learning and memory might be enhanced in mammals by genet-
8d), whereas the densities of perforated and shaft synapses remained ic factors as well as behavioral experience.
unchanged after enrichment in both control and CA1-KO mice. For
better observation of the synaptic structural changes, a sophisticat-
ed approach using three-dimensional reconstruction of synapses METHODS
Animals. The CA1-KO and control mice were produced as described11.
viewed through serial electron micrographs would be valuable33.
Throughout the behavioral and histological experiments, observers were
Possible tissue shrinkage during the embedding process can blind to the genotype of an individual animal.
be ruled out as explanation for these findings. First, the effect is
synapse specific. If there were a generalized shrinkage in the In situ hybridization. Described procedures10 were used. Briefly, an anti-
enriched animals with no other change, then all types of synaps- sense 42-mer oligonucleotide probe (5-ACC ACT CTT TCT ATC
es should be affected uniformly. Second, the mitochondrial cross- CTG CAG GTT CTT CCT CCA CAC GTT-3), which recognizes NR1
sectional diameters were uniform across the naive and enriched exon 20, was end labeled with [35S]-dATP. After hybridizing with the
conditions, showing that no generalized shrinkage of individual oligonucleotide probe (5 105 cpm per slide) at 47C for 24 h, brain sec-
process occurred (data not shown). tions (20 m) were washed in 2 SSC at room temperature (RT) fol-
lowed by two washes in 0.2 SSC at 60C and one wash in 0.1 SSC at
RT. Interestingly, preliminary observations indicated that the deletion
DISCUSSION of the NR1 gene in six-month-old knockouts seemed to spread to other
Here we examined three kinds of nonspatial memory in mice forebrain regions such as the cortex, probably reflecting the recombina-
lacking NMDAR1 in the CA1 region. Our analysis provides evi- tion threshold effect as the CaMKII promoter-driven Cre accumulated.
dence supporting the important role of NMDA receptor activity
in the CA1 region for the formation of hippocampus-dependent Enriched environment training. Adult littermates (1.52 months old) were
nonspatial memory. Taking into account the deficit in spatial randomly distributed into two experimental groups. One group was kept
learning in the CA1-KO mice12, we conclude that NMDA recep- in standard cages (naive group) and the other group trained in an enriched
tor activity in CA1 is essential for the formation of both spatial environment for three hours daily for two months (enriched group). The
and nonspatial memory. enrichment training was carried out in specially designed boxes (1.5 m
0.8 m 0.8 m), in which various toys, running wheels and small houses were
More interestingly, we found that the memory deficits associ-
changed every other day to encourage exploration. Food and water were
ated with the disruption of the hippocampal NMDA receptor could also available in the boxes. Because electrophysiological, behavioral and
be rescued by daily training in an enriched environment. We show anatomical experiments showed similar results for wild-type and Cre trans-
that behavioral experience can enhance learning and memory in genic mice, these two genotypes of mice were pooled together as controls.
the mutant animals. Thus, the memory deficits observed in gen-
tic mutant animals are not necessarily irreversible, and enriched Novel-object recognition task and fear-conditioning tasks. The experi-
experience could promote recovery of some of the deficits. mental protocol was the same as described previously10, except that the
To investigate further the possible cellular mechanisms under- duration of the training period in the object-recognition task was 15 min.
lying these enrichment effects, we systematically evaluated the
Social transmission of food preference task. In this experiment, observ-
anatomical changes using light and electron microscopy. We
er mice were tested for memory and demonstrator wild-type mice were
showed that enriched experience promotes a significant increase used to interact with observers according to a described procedure22.
in synapse density in the CA1 hippocampal field. Because there is After interaction, the observers were deprived of food for 24 h and then
no neurogenesis in the CA1 region26, the increased synapse den- tested for memory. During testing, observers were offered both cinna-
sity is likely to reflect a higher number of synapses per neuron mon- and cocoa-scented food for 2 h, and consumption of each food
rather than an increased number of pyramidal cells in CA1. At the was subsequently measured. The preferred food had the same scent as
ultrastructural level, we showed that mice lacking the NMDA food eaten by the demonstrator with whom the observer interacted. Pref-
receptor in the CA1 region also increase synaptic density follow- erence was calculated as percentage ratio of the amount of preferred food
ing exposure to an enriched environment. This suggests that consumed over 50% of the total amount of food consumed.
NMDA-receptor activity is not required for structural plasticity in
Histology. For electron microscopy, contol groups were composed of males,
the CA1 region of adult animals induced by behavioral experience,
3.55.5 months old. The CA1-KO naive mice were 3 males (4.5, 4 and 4
and thus the molecular mechanism underlying adult activity- months) and 1 female (3.5 months), and the CA1-KO enriched mice were
dependent structural plasticity is very different from mechanisms 3 males (5, 4.5 and 3.5 months) and 1 female (4 months). Because CA1
active in the developing brain17,18. As synaptic plasticity can also dendritic spine density in rats fluctuates over the estrous cycle39, females
be induced in the CA1 region by an NMDA receptor-independent were killed during proestrus, the stage at which the spine numbers of
process37,38 such as LTP dependent on voltage-gated calcium chan- female rats approach those of male rats (although it is not known whether
nels, other mechanisms might serve as putative candidates for expe- the same phenomenon exists in mice). Animals were anesthetized and per-
rience-induced structural plasticity in adult brain. fused with 2% paraformaldehyde, 2% glutaraldehyde and 1.5% saturated
It remains to be determined whether the enrichment-induced picric acid in 0.1 M phosphate buffer (pH 7.4) and postfixed overnight in
the same solution. Following a modified version of the single-section Golgi
increase in CA1 synaptic density has any functional role in the
impregnation technique32, 100 m-thick coronal sections were cut with a
concomitant behavioral effects. It is likely that other brain regions vibratome into a bath of 3% potassium dichromate in distilled water and
may undergo similar biochemical and structural changes after incubated overnight in this solution. The slide assemblies were incubated
enrichment, and also might participate in the enrichment-induced in the dark in a solution of 1.5% silver nitrate for 2 days.
behavioral improvement. Enhanced synaptic coincidence detec- For electron microscopy, we used a vibratome to cut 250 m-thick coro-
tion in the forebrain of transgenic mice via upregulation of sub- nal sections into a bath of 0.1 M sodium cacodylate buffer. Hippocampal

articles
sections corresponding to figures 4750 of the mouse brain atlas40 were 6. Wood, E. R., Dudchenko, P. A. & Eichenbaum, H. The global record of
dissected, rinsed in distilled water and postfixed in 1% osmium tetroxide memory in hippocampal neuronal activity. Nature 397, 613616 (1999).
7. Moriyoshi, K. et al. Molecular cloning and characterization of the rat NMDA
followed with 1% ferrocyanide-reduced osmium tetroxide. Tissues were receptor. Nature 354, 3137 (1991).
rinsed, stained en bloc overnight in 1% aqueous uranyl acetate, rinsed and 8. Nicoll, R. A. & Malenka, R. C. Expression mechanisms underlying NMDA
dehydrated and subsequently infiltrated and flat embedded with Embed receptor-dependent long-term potentiation. Ann. NY Acad. Sci. 868, 515525
812 resin. Sections (0.5 m) were cut and stained with 1% toluidine blue. (1999).
Sections were examined to determine how the block should be trimmed 9. Bear, M. F. & Malenka, R. C. Synaptic plasticity: LTP and LTD. Curr. Opin.
Neurobiol. 4, 389399 (1994).
for ultrathin sectioning to include the apex of the CA1 pyramidal cell layer, 10. Tang, Y.-P. et al. Genetic enhancement of learning and memory in mice.
from the stratum radiatum to the stratum lacunosum moleculare. Ultra- Nature 401, 6369 (1999).
thin sections were cut with a diamond knife, placed on carbon Formvar- 11. Tsien, J. Z. et al. Subregion- and cell type-restricted gene knockout in mouse
coated slot grids and stained with 1% uranyl acetate and lead citrate. brain. Cell 87, 13171326 (1996).
12. Tsien, J. Z., Huerta, P. T. & Tonegawa, S. The essential role of hippocampal
Section thickness was measured by scanning electron microscope mea- CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell
surements to range from 0.087 to 0.105 m, depending on the series. 87, 13271338 (1996).
13. McHugh, T. J., Blum, K. I., Tsien, J. Z., Tonegawa, S. & Wilson, M. A. Impaired
Data analysis: spine density. Golgi-impregnated dendritic segments hippocampal representation of space in CA1-specific NMDAR1 knockout
selected for analysis were located 100250 m from the pyramidal cell mice. Cell 87, 13391349 (1996).
14. Woolf, N. J. A structural basis for memory storage in mammals. Prog.
bodies in the stratum radiatum, belonged to a thoroughly impregnated Neurobiol. 55, 5977 (1998).
pyramidal neuron, remained approximately in the plane of focus and 15. Bailey, C. H. & Kandel, E. R. Structural changes accompanying memory
were >10 m in length and 0.801 m in width. Calculations of spine storage. Annu. Rev. Physiol. 55, 397426 (1993).
density were obtained by measuring dendritic segments matching the 16. Moser, M. B., Trommald, M. & Andersen, P. An increase in dendritic spine
above criteria (control enriched n = 37, CA1-KO naive n = 38, CA1-KO density on hippocampal CA1 pyramidal cells following spatial learning in
adult rats suggests the formation of new synapses. Proc. Natl. Acad. Sci. USA
enriched n = 41 and control naive n = 43) from age- and sex-balanced 91, 1267312675 (1994).
sets of 3 animals of each group. These segments represented a total den- 17. Goodman, C. S. & Shatz, C. J. Developmental mechanisms that generate
dritic length of 10011284 m per group of mice. precise patterns of neuronal connectivity. Cell 72, Suppl. 7798 (1993).
18. Li, Y., Erzurumlu, R. S., Chen, C., Jhaveri, S. & Tonegawa, S. Whisker-related
neuronal patterns fail to develop in the trigeminal brainstem nuclei of
Data analysis: synapse density. Synapse density was estimated by the dis- NMDAR1 knockout mice. Cell 76, 427437 (1994).
ector method34,35. From each brain, 15 electron micrographs of separate 19. Kirov, S. A. & Harris, K. M. Dendrites are more spiny on mature hippocampal
regions within the stratum radiatum, 100250 m from the pyramidal neurons when synapses are inactivated. Nat. Neurosci. 10, 878883 (1999).
cell layer, were taken from different sections at 10,000 on a JEOL 100C 20. Sorra, K. E. & Harris, K. M. Stability in synapse number and size at 2 hr after
electron microscope to create 15 reference planes. The micrographs cov- long-term potentiation in hippocampal area CA1. J. Neurosci. 18, 658671
(1998).
ered an area of approximately 2100 m2 per animal. Micrographs of exact- 21. Mansuy, I. M., Mayford, M., Jacob, B., Kandel, E. R. & Bach, M. E. Restricted
ly the same regions were taken on an adjacent section to create look-up and regulated overexpression reveals calcineurin as a key component in the
planes. The number of synapses contained in a reference plane but absent transition from short-term to long-term memory. Cell 92, 3949 (1998).
in the corresponding look-up plane (Qi) was counted to determine the 22. Kogan, J. H. et al. Spaced training induces normal long-term memory in
CREB mutant mice. Curr. Biol. 7, 111 (1996).
number of synapses present within the volume defined by the reference 23. Strupp, B. J. & Levitsky, D. A. Social transmission of food preferences in adult
plane, the look-up plane and the distance between them (hi). Two sides hooded rats (Rattus norvegicus). J. Comp. Psychol. 98, 257266 (1984).
of the rectangular picture were assigned randomly as inclusion or exclu- 24. Bunsey, M. & Eichenbaum, H. Selective damage to the hippocampal region
sion edges to create a counting frame that minimized potential edge effects blocks long-term retention of a natural and nonspatial stimulus-stimulus
across samples41. To increase the sampling reliability, 6061043 synapses association. Hippocampus 5, 546556 (1995).
25. Phillips, R. G. & LeDoux, J. E. Differential contribution of amygdala and
were counted over a total area of 8,40014,700 m2 from 600 electron hippocampus to cued and contextual fear conditioning. Behav. Neurosci. 106,
micrographs for each experimental group of mice. Section thickness (hi) 274285 (1992).
was determined by scanning electron microscope measurements (see His- 26. Kempermann, G., Kuhn, G. H. & Gage, F. H. More hippocampal neurons in
tology). Estimated synapse density (est Nv) was calculated as adult mice living in an enriched environment. Nature 386, 493495 (1997).
27. Dalrymple-Alford, J. C. & Benton, D. Preoperative differential housing and
dorsal hippocampal lesions in rats. Behav. Neurosci. 98, 2334 (1984).
est Nv = Qi / (Ai hi) 28. Rosenzweig, M. R. Environmental complexity, cerebral change, and behavior.
Am. Psychol. 21, 321332 (1966).
Note: Additional methods details can be found on the Nature Neuroscience web 29. Diamond, M. C. Enriching Heredity: The Impact of the Environment on the
site (http://neurosci.nature.com/web_specials/). Anatomy of the Brain (Free Press, New York, 1988).
30. Fiala, B. A., Joyce, J. N. & Greenough, W. T. Environmental complexity
modulates growth of granule cell dendrites in developing but not adult
hippocampus of rats. Exp. Neurol. 59, 372383 (1978).
ACKNOWLEDGEMENTS 31. Greenough, W. T., Withers, G. S. & Wallace, C. S. in The Biology of Memory
We thank E. Gould for help with stereology and statistics and reading the manuscript, and (eds. Squire, L. R. & Lindenbaum, E.) 159185 (Schattauer, Stuttgart, 1990).
P. Tanapat for technical advice concerning the Golgi method. This work was supported in 32. Gabbott, P. L & Somogyi, J. The single Golgi impregnation procedure:
methodological description. J. Neurosci. Methods 11, 221230 (1984).
part by a postdoctoral fellowship from Fondation pour la Recherche Medicale to C.R. and 33. Sorra, K. E., Fiala, J. C. & Harris, K. M. Critical assessment of the involvement
by grants from Princeton University, Beckman Foundation and NIH to J.Z.T. of perforations, spinules, and spine branching in hippocampal synapse
formation. J. Comp. Neurol. 398, 225240 (1998).
34. Sterio, D. C. The unbiased estimation of number and sizes of particles using
RECEIVED 12 OCTOBER; ACCEPTED 28 DECEMBER 1999 the disector. J. Microsc. 134, 127136 (1984).
35. DeGroot, D. M. G. & Bierman, E. P. B. A critical evaluation of methods for
estimating the numerical density of synapses. J. Neurosci. Methods 18, 79101
1. Rosene, D. L. & Van Hoesen, G. W. in Cerebral Cortex Vol. 6: Further Aspects of (1986).
Cortical Function, Including Hippocampus (eds. Jones, E. G. & Petes, A.) 36. Gray, E. G. Axosomatic and axodendritic synapses of the cerebral cortex: An
345447 (Plenum, New York, 1987). electron microscopic study. J. Anat. 83, 420433 (1959).
2. Scoville, W. B. & Milner, B. Loss of recent memory after bilateral hippocampal 37. Nicoll, R. A. & Malenka, R. C. Contrasting properties of two forms of long-
lesion. J. Neurol. Neurosurg. Psychiatry 20, 1112 (1957). term potentiation in the hippocampus. Nature 377, 115118 (1995).
3. Zola-Morgan, S., Squire, L. R. & Amaral, D. G. Human amnesia and the medial 38. Stevens, C. F. & Sullivan, J. Synaptic plasticity. Curr. Biol. 8, R151153 (1998).
temporal region: enduring memory impairment following a bilateral lesion 39. Woolley, C. S., Gould, E., Frankfurt, M. & McEwen, B. S. Naturally occurring
limited to field CA1 of the hippocampus. J. Neurosci. 6, 29502967 (1986). fluctuation in dendritic spine density on adult hippocampal pyramidal
4. Squire L. R. Memory and the hippocampus: a synthesis from findings with neurons. J. Neurosci. 10, 40354039 (1990).
rats, monkeys, and humans. Psychol. Rev. 99, 195231 (1992). 40. Franklin, K. B. J. & Paxinos, G. The mouse brain in stereotaxic coordinates.
5. Mumby, D. G. et al. Ischemia-induced object-recognition deficits in rats are (Academic, San Diego, 1997).
attenuated by hippocampal ablation before or soon after ischemia. Behav. 41. Gundersen, H. J. G. Notes on the estimation of the numerical density of
Neurosci. 110, 266281 (1996). arbitrary profiles: the edge effect. J. Microsc. 111, 219223 (1977).

articles
Muscles express motor patterns of

non-innervating neural networks by
filtering broad-band input
Lee G. Morris1, Jeff B. Thuma2 and Scott L. Hooper2
1 Department of Physiology and Biophysics, Mt. Sinai Medical School, Box 1218, 1 Gustave L. Levy Place, New York, New York 10029, USA
2 Neuroscience Program, Department of Biological Sciences, Irvine Hall, Ohio University, Athens, Ohio 45701, USA
The first two authors contributed equally to this work.
Correspondence should be addressed to S.L.H. (Hooper@ohio.edu)
We describe three slow muscles that responded to low-frequency modulation of a high-frequency

neuronal input and, consequently, could express the motor patterns of neural networks whose
neurons did not directly innervate the muscles. Two of these muscles responded to different
frequency components present in the same input, and as a result each muscle expressed the motor
pattern of a different, non-innervating, neural network. In an analogous manner, the distinct
dynamics of the multiple intracellular processes that most cells possess may allow each process to
respond to, and hence differentiate among, specific frequency ranges present in broad-band input.
Neurons receive input across an extremely wide frequency range output is modulated by two other, much slower rhythmic neur-
(from high-frequency sound at tens of kHz to circadian rhythms al networks. The muscles responded to these modulations and
at 10 Hz). Muscles receive input across a narrower but still followed the slow network activity, even though no neuron of
wide range (10 Hz to several hundred Hz, the maximum firing the slow networks innervates the muscles. These results extend
rate of neurons). Neurons and muscles cannot accurately follow the known range of functional low-frequency response some
the higher portions of these ranges. For instance, the refractory two orders of magnitude, and provide an example of a signal
period limits neuron firing to a few hundred Hz, and relaxation carried partially or exclusively in a slow modulation imposed
kinetics prevent muscles from accurately following motor neu- on the targets driver from elsewhere in the nervous system.
ron bursts faster than 1 Hz1. Neuronal and muscle intracellu- Furthermore, two of these muscles are innervated by the same
lar processes (for example, Ca 2+ and second messenger motor neurons and, hence, receive identical neuronal input.
concentration, protein kinase activation, protein phosphoryla- However, because of differences in their contractile properties,
tion and gene expression) are also slow, with time courses rang- one muscle followed both slow networks, whereas the other fol-
ing from tens of milliseconds (10100 Hz) to minutes (0.01 Hz) lowed only one. These two muscles responded to different fre-
or longer215, and therefore would also be expected to be unable quency domains of one signal, and thus provide a clear example
to accurately follow high-frequency input. of both encoding of information by the nervous system and its
Restriction of responsiveness to a particular frequency range decoding by neuronal followers, on the basis of frequency. Pre-
might be thought to limit function. Indeed, the inhibitory and liminary accounts of this work appeared in abstract form
neuromodulatory muscle innervation in several invertebrates (L.G.M. et al., Soc. Neurosci. Abstr. 25, 1642, 1999; J.B.T. &
may be present to increase the frequency range over which the S.L.H., Soc. Neurosci. Abstr. 25, 1641, 1999).
muscles can accurately follow their inputs1619. However, selec-
tivity for low frequencies may also be advantageous because it RESULTS
allows receivers to demodulate, and hence respond to, low-fre- We studied the pyloric neuromuscular system of the lobster
quency signals carried in a modulated high-frequency (carrier (Panulirus interruptus), a part of the stomatogastric system that
wave) input. The wide range of kinetics present in the respons- is driven by the pyloric neural network. Every 0.5 to 2 seconds,
es of neurons and muscles to input suggests that these systems the motor neurons of this network produce 510 action poten-
might have the cellular mechanisms necessary to perform such tial bursts of 100500 millisecond duration25. Pyloric muscles
demodulation, and different patterns of temporal input can dif- are non-spiking muscles that contract as a graded function of
ferentially alter various intracellular processes12,2024. However, their neuronal input26. Based on the pyloric networks cycle
except for the exclusion of inputs with frequencies of 1 Hz or period and bursting nature, it was thought that each motor
greater in vertebrate skeletal muscle1, no examples in which the neuron burst induces a single muscle contraction that fully
signal of interest is carried in the low-frequency component of relaxes between bursts, and that the muscles thus temporally
a broad-band input have been described in biological systems mirror the bursting activity of their input 25,2730. Work on
on the cellular level. pyloric muscle contractions evoked by nerve stimulation in the
We describe here three slow muscles driven by a bursting crab, Cancer borealis, and the shrimp, Palaemon serratus, sup-
neural network with a cycle frequency of 1 Hz. This networks port this belief3133.

articles
Fig. 1. In some pyloric muscles, overall spike frequency codes steady-

state average contraction amplitude. (a) Top trace, intracellular PD neu- a
ron recording; bottom trace, isotonic contraction of an extremely slow PD
muscle innervated by the PD neurons (the dorsal PD muscle) induced neuron
by motor nerve stimulation mimicking the first burst of action potentials
(vertical lines below muscle trace) in the PD neuron trace. If the muscle 15 mV
were driven by the neuron, the next contraction (arrow) would occur Dorsal PD 48 m
before the first contraction finished, resulting in intercontraction sum- muscle 0.5 s
mation. Most hyperpolarized PD neuron membrane potential, 62 mV.
(b) Rhythmic stimulation (burst duration, 390 ms; spike frequency,
20 Hz; cycle period, 2 s) of the motor nerve innervating the dorsal PD b
muscle resulted in (after intercontraction summation had stabilized) a
sustained tonic contraction on which rhythmic contractions rode. (c)
amplitude (m)
Contraction
After intercontraction summation had stabilized, overall spike frequency
(spike number per burst divided by cycle period) coded average con-
traction amplitude (one half phasic plus tonic contraction) for this mus-
cle. All data are from the cpv1b muscle39.
Time (s)
However, in Panulirus some pyloric muscles contract and relax

much too slowly to follow their motor neuron bursts faithful- c
amplitude (mm)
ly34,35 (Fig. 1a). The top trace shows the characteristic rhythmic
Average
action potential bursts of a pyloric dilator (PD) neuron. The sec-
ond trace shows an isotonic contraction (muscle shortening
against constant load) of an extremely slow muscle innervated
by the PD neurons (the dorsal PD muscle). This contraction was
induced by stimulation of the motor nerve with parameters
matching the first burst in the neuron trace. If the neuron were Overall spike freq. (Hz)
driving this muscle (in all experiments the motor nerve was cut
to prevent spontaneous neuron input to the muscle), the next summation had stabilized, the muscles response therefore con-
contraction would have occurred before the contraction induced sisted of a sustained, tonic baseline contraction on which rode
by the previous burst had ended (arrow; angled to account for phasic contractions that matched the stimulation cycle period.
the long delay between the neuron burst and the contraction Here the average contraction amplitude (one half the phasic con-
beginning). traction amplitude plus the tonic contraction amplitude) of the
When the motor nerve was rhythmically stimulated with muscle is particularly important, and after muscle summation
bursts of shocks using physiologically relevant parameters, the has stabilized, this average amplitude is well predicted by the
muscle contractions temporally summated (Fig. 1b). After the overall spike frequency (spike number per burst divided by cycle
period) of the motor neuron (Fig. 1c; ref. 34 and L.G.M. & S.L.H.,
Soc. Neurosci. Abstr. 24, 1891, 1998). These data suggest that if
1 the overall spike frequencies of pyloric motor neurons were slow-
PY ly modulated by extrinsic influences, average contraction ampli-
neuron
tude of pyloric muscles might similarly vary.
The stomatogastric nervous system also contains the gastric
mill and cardiac sac networks, which have much slower cycle
10 mV
periods (510 seconds and 1 minute, respectively) than the
pyloric network25. Gastric mill and cardiac sac activity modu-
2 lates pyloric activity25,28,3638, and the overall spike frequency of
aln
several pyloric neurons varies with gastric mill and cardiac sac
activity (J.B.T. & S.L.H., Soc. Neurosci. Abstr. 25, 1641, 1999).
Using stimulation parameters determined from previously mea-
sured effects of gastric mill and cardiac sac network activity on
3 10
Overall 7.5
spike 5 Fig. 2. A muscle innervated by PY neurons expressed a gastric mill net-
freq. (Hz) work contraction pattern even though no gastric mill neurons innervate
2.5
the muscle. Trace 1, intracellular recording of the PY neuron whose
0 activity was used to stimulate the muscles motor nerve. Most hyperpo-
larized PY neuron membrane potential, 55 mV. Trace 2, extracellular
4 0.15 mm recording from the aln of the activity of the GM neurons of the gastric
mill network. Trace 3, PY neuron overall spike frequency. Trace 4, iso-
PY tonic contractions of a PY neuron-innervated muscle. Gray shading
muscle
shows the tonic component of the muscle contraction; the muscle
almost fully relaxed only at the end of each gastric mill cycle (dashed
lines mark one cycle period). Solid horizontal line, fully relaxed muscle
2s length. Recording is from p8 muscle39.

articles
Fig. 3. Although no cardiac sac neurons innervate it, the extremely slow
1 Cardiac dorsal PD muscle primarily contracted in a pattern matching the very
sac
slow cardiac sac rhythm. Trace 1, schematic representation of cardiac sac
network activity; rectangles correspond to cardiac sac bursts, dashed
lines mark one cardiac sac cycle period. Trace 2, intracellular recording of
2 PD
neuron
the PD neuron whose activity was used to stimulate the muscles motor
nerve. Trace 3, time expansion of PD neuron activity immediately before,
during and after a cardiac sac burst. Most hyperpolarized PD neuron
membrane potential, 65 mV. Trace 4, PD neuron overall spike fre-
quency. Trace 5, isotonic contractions of a dorsal PD muscle (inset is a
time expansion showing small contractions in pyloric time). Scale bars,
3 20 s (traces 1, 2, 4 and 5) or 2 s (trace 3) and 5 mV (traces 2,3) or 60 m
(trace 5). Recording is from the cpv1b muscle39.
5 mV
2s
20
4 priate for this new spike frequency. If overall spike frequency
Overall 15 remained at the new value, the muscle would stabilize at the aver-
spike 10 age contraction amplitude predicted by the equivalent of Fig. 1c
freq. (Hz)
5 for the PY muscle. However, overall spike frequency does not sta-
0 bilize; consequently, the muscle is driven toward different regions
of this curve each pyloric cycle, which generates the slow varia-
5 tions in average contraction amplitude (Fig. 2).
For the relatively fast PY muscle, the contractions in pyloric
Dorsal PD time were a large percentage of total contraction amplitude.
muscle
baseline
In contrast, the extremely slow dorsal PD muscle responded
(full relaxation) 60 m to PD neuron input modulated by the very slow, 0.01 Hz, car-
20 s
diac sac network primarily with a slow, rhythmic variation in
tonic contraction amplitude (Fig. 3). During cardiac sac net-
pyloric network output, we explored the effects of altering pyloric work bursts (trace 1; each rectangle represents one cardiac sac
neuron activity on pyloric muscles by stimulating motor nerves burst, dashed lines mark one cardiac sac cycle), PD neuron
innervating various pyloric muscles (Fig. 2). The first trace shows activity was markedly altered, but the neuron continued to
an intracellular recording from a pyloric (PY) motor neuron; the burst in pyloric time (trace 2; expanded over one cardiac sac
second trace is an extracellular recording from the anterior lat- burst in trace 3). PD neuron overall spike frequency (trace 4)
eral nerve (aln), which carries the axons of the gastric mill (GM) showed a triphasic increasedecreaseincrease pattern with
neurons of the gastric mill network39. PY neuron firing waxed each cardiac sac burst. When the motor nerve innervating the
and waned as a function of gastric mill network activity (the extremely slow dorsal PD muscle was stimulated with this
dashed lines mark one gastric mill network cycle), and the third activity pattern, the muscles contractions in pyloric time
trace shows that this change in PY neuron activity resulted in (shown on expanded time scale in inset) were only 8% of the
corresponding changes in PY neuron overall spike frequency. maximum contraction amplitude, whereas its contractions in
The fourth trace shows the contractions of a muscle inner- cardiac sac time were 50% of maximum contraction ampli-
vated by the PY neurons induced by motor nerve stimulation tude (trace 5). Thus, even though no motor neurons of the
exactly matching the PY neuron spiking pattern shown in the cardiac sac network innervate it39, the muscle primarily con-
first trace. As overall spike frequency increased at the beginning tracted in cardiac sac time.
of each GM neuron burst, the phasic amplitude and summation These data show that the PD and PY muscles can respond to
of the pyloric-timed PY muscle contractions increased and, con- slow modulations of PD and PY neuron activity. However,
sequently, the muscles tonic contraction component (gray shad- because the activity patterns of the PD and PY neurons are dif-
ing) increased. As overall spike frequency decreased during the ferent, the PD and PY muscles receive different input; these data,
GM neuron interburst interval, the muscles phasic contraction therefore, do not show how different muscles respond to the same
amplitude and summation and, thus, its tonic contraction, input. As well as innervating the extremely slow dorsal PD mus-
decreased, allowing it to return nearly to rest length before the cle, the PD neurons also innervate the faster ventral PD muscle,
next GM neuron burst. Thus, although no gastric mill neuron which allowed us to investigate whether the two muscles selec-
innervates this muscle, it nonetheless displayed both a pyloric tively respond to different frequency domains within an identical
(the rapid rhythmic contractions) and a gastric mill (the tonic neural input.
contraction component) motor pattern. In addition to showing large (15 Hz) variations in overall
The changes in the PY muscle tonic contraction can be qual- spike frequency in time with the slow cardiac sac network
itatively understood by considering Fig. 1b and c. Contractions of (0.01 Hz; Fig. 3), the PD neurons often also showed small
the extremely slow dorsal PD muscle took 1015 pyloric cycles (2.5 Hz) variations in overall spike frequency (Fig. 4, trace 1)
to stabilize (Fig. 1b). The PY muscle is faster than the dorsal PD in time with the faster gastric mill network (0.10.2 Hz; trace
muscle, but it nonetheless took 23 cycles to stabilize in response 2; dashed lines mark one gastric mill cycle). The extremely slow
to rhythmic stimulation and shows similar coding dependence dorsal PD muscles showed much smaller variations in gastric
(M. Rehn, L.G.M. & S.L.H., unpublished observations). When mill time than did the faster ventral PD muscles and, depend-
PY neuron overall spike frequency changes, the muscles con- ing on gastric mill cycle period, sometimes completely exclud-
traction amplitude changes toward an average contraction appro- ed the gastric mill signal. In these cases (Fig. 4, trace 3), the

articles
Fig. 4. Different muscles can respond to different frequency domains in

1 12.5 the same neural signal. The PD neurons innervate both the very slow
dorsal PD muscle and the faster ventral PD muscle. PD neuron overall
10 frequency (trace 1) varied as a function of gastric mill phase (trace 2, GM
Overall 7.5 neuron activity recorded from the aln; dashed lines mark a single cycle
spike period). Muscle contraction amplitude of the extremely slow dorsal PD
freq. (Hz) 5 muscle did not vary as a function of gastric mill phase (trace 3), whereas
2.5 that of the faster ventral PD muscle did (trace 4). Trace 5, intracellular
PD neuron recording used to stimulate the muscles motor nerves.
0 Most hyperpolarized potential of PD neuron membrane,65 mV. The
dorsal PD muscle is the cpv1b muscle; the ventral is the cpv2b muscle39.
2
aln
3 Dorsal
PD
input. These results have implications not only for interpreting
muscle studies of the pyloric system, but also, more generally, for
0.2 mm
baseline (full relaxation) understanding how and to what extent biological systems can
respond to input signals carried in different frequency domains.
4 Pyloric network function

In response to bath application of various neuromodulatory
Ventral
PD substances 29,4043 or stimulation of modulatory
muscle inputs25,28,29,37,38,4448, the pyloric network produces different
0.2 mm neural outputs. It is tempting to interpret these changes as
baseline (full relaxation)
inducing changes in a pyloric-timed motor pattern. The con-
traction of several pyloric muscles in phase with other, slow-
er neural networks suggests, however, that changing pyloric
5 neural activity will have complex effects on pyloric motor out-
PD put. In particular, these data imply that understanding the
neuron functional effects of changes in pyloric neuron activity may
require consideration not only of pyloric activity on a pyloric
cycle-by-cycle basis, but also the effects of intercontraction
summation over time scales 5- to 10-fold and 60- to 100-fold
5 mV
longer (corresponding to gastric mill and cardiac sac cycle
2s periods, respectively).
Pyloric cycle period varies between 0.5 and 2.0 seconds;
output of the dorsal PD muscle consisted of a tonic contrac- depending on pyloric period, the relaxation times of some
tion, on which were superimposed very small rhythmic con- pyloric muscles may or may not allow intercontraction sum-
tractions in pyloric time but no measurable contractions in mation (T.A. Ellis, P.I. Harness, T.J. Koehnle, M. Rehn, L.G.M.
gastric mill time. In contrast, the faster ventral PD muscle (trace & S.L.H., unpublished observations). For instance, the pyloric
4) was rhythmically active in both pyloric and gastric mill time. network shown in Fig. 2 had a period of 0.7 seconds. At a
Both muscles also showed large contractions during cardiac sac period of 2.0 seconds, this muscles contractions would near-
bursts (dorsal PD muscle, Fig. 3; ventral PD muscle, data not ly or fully relax between neuron bursts, intercontraction sum-
shown). Thus, the relatively fast ventral PD muscle was rhyth- mation and tonic contraction amplitude would dramatically
mically active in pyloric (1 Hz), gastric (0.15 Hz) and cardiac decrease, and the variations of tonic contraction amplitude in
sac (0.01 Hz) time, whereas the slow dorsal PD muscle respond- gastric mill time would be a much smaller component of total
ed to the rapid pyloric and very slow cardiac sac signals, but muscle contraction. Thus one effect of changing pyloric cycle
failed to respond to the intermediate gastric mill signal. Note period may be to alter the expression of gastric mill activity
also the sensitivity of the ventral PD muscle to input in gastric as a variation in tonic contraction amplitude; at very slow
mill time; the small changes in PD neuron overall spike fre- pyloric cycle periods, these variations could be completely
quency (Fig. 4, trace 1) in gastric time were almost indistin- abolished.
guishable in the neuron intracellular recording (trace 5), but
resulted in a pronounced (50% of maximum contraction ampli- Signal extraction from different frequency domains
tude) gastric mill component in the muscle contraction. Considerable evidence indicates that neuronal intracellular
processes are sensitive to the temporal patterning of the input
DISCUSSION the neuron receives. For instance, action potential patterning
These data demonstrate that three slow muscles of the pyloric alters Ca2+ and second messenger concentration, protein expres-
neuromuscular system can respond to slow modulation of the sion and gene regulation12,2023. The specificity of gene activa-
rapid pyloric pattern, and hence express, either partially (the tion by Ca2+ oscillations depends on oscillation frequency24,
PY and ventral PD muscles) or almost exclusively (the dorsal and modeling studies suggest that multiple feedback systems
PD muscle), the rhythmicity of networks of neurons that do sensitive to different time scales49 may be necessary to explain
not innervate the muscles. They further demonstrate that two the response of stomatogastric neurons to temporal variations
targets (the slow ventral and faster dorsal PD muscles) can in input patterns50. Although these observations are intriguing,
respond to different frequency domains in the same neuronal the functional significance of these data is unclear, as the exper-

articles
imental stimulation patterns used are not based on actual ACKNOWLEDGEMENTS

inputs. In contrast, our input exactly matched that produced This work was supported by Ohio University, NSF and NIMH. We thank R.A.
by the systems innervating neurons. DiCaprio for discussion and advice and A. L. Weaver for setting up the CED to
More generally, previous experimental work on sensitivity stimulator interface and writing scripts to transform CED events into
to temporal patterns does not address the issue of whether par- waveforms used to drive the stimulator.
ticular second messengers selectively respond to different fre-
quency domains. Our demonstration that two muscles RECEIVED 2 JULY; ACCEPTED 20 DECEMBER 1999
responded to different frequencies in a single input (Fig. 4)
directly addresses this point and demonstrates a differential
1. Partridge, L. D. Signal-handling characteristics of load-moving skeletal
response as a function of frequency domain. Presumably, this muscle. Am. J. Physiol. 210, 11781191 (1966).
differential response largely stems from the different temporal 2. Smith, S. J., MacDermott, A. B. & Weight, F. F. Detection of intracellular Ca2+
properties of the muscles. Intracellular responses show a wide transients in sympathetic neurones using arsenazo III. Nature 304, 350352
(1983).
kinetic range215, and differential frequency response, therefore, 3. Gamble, E. & Koch, C. The dynamics of free calcium in dendritic spines in
may also be present in neuronal intracellular responses to response to repetitive synaptic input. Science 236, 13111315 (1987).
broad-band input. Our data thus support the hypothesis that 4. Delaney, K. R., Zucker, R. S. & Tank, D. W. Calcium in motor nerve
terminal associated with posttetanic potentiation. J. Neurosci. 9, 35583567
these processes [intracellular messengers] may have...fea- (1989).
tures...that limit their involvement to certain patterns of stim- 5. Breer, H., Boekhoff, I. & Tareilus, E. Rapid kinetics of second messenger
ulation12. formation in olfactory transduction. Nature 345, 6568 (1990).
6. Delaney, K. R. & Tank, D. W. A quantitative measurement of the dependence

The low-frequency response shown here does not require of short-term synaptic enhancement on presynaptic residual calcium.
rhythmicity of the underlying input. Slow processes driven by J. Neurosci. 14, 58855902 (1994).
irregular input would smooth out high-frequency variation and 7. Regehr, W. G., Delaney, K. R. & Tank, D. W. The role of presynaptic calcium
in short-term enhancement at the hippocampal mossy fiber synapse.
would respond primarily to the inputs average frequency10,14. J. Neurosci. 14, 523537 (1994).
However, if this average were slowly modulated, the system 8. Wang, S. S.-H., Alousi, A. A. & Thompson, S. H. The lifetime of inositol
would be expected to follow the moving average and, hence, to 1,4,5-triphosphate in single cells. J. Gen. Physiol. 105, 149171 (1995).
9. Feller, M. B., Delaney, K. R. & Tank, D. W. Presynaptic calcium dynamics at
respond to the modulation. Thus, one consequence of the dis- the frog retinotectal synapse. J. Neurophysiol. 76, 381400 (1996).
tinct dynamics of various intracellular processes in a biologi- 10. Helmchen, F., Imoto, K. & Sakmann, B. Ca2+ buffering and action potential-
evoked Ca2+ signaling in dendrites of pyramidal neurons. Biophys. J. 70,
cal system may be that each process responds to specific 10691081 (1996).
frequency domains present in broad-band inputs. Furthermore, 11. Brezina, A., Orekhova, I. V. & Weiss, K. R. Control of time-dependent
as expression of these processes can itself be regulated, cells may biological processes by temporally patterned input. Proc. Natl. Acad. Sci. USA
94, 1044410449 (1997).
also be developmentally and functionally tuned to specific fre- 12. Fields, R. D., Eshete, F., Stevens, B. & Itoh, K. Action potential-dependent
quency domains contained in their input. regulation of gene expression: temporal specificity in Ca2+, cAMP-responsive
element binding proteins, and mitogen-activated protein kinase signaling.
J. Neurosci. 17, 72527266 (1997).
METHODS 13. Chatton, J.-Y., Cao, Y. & Stucki, J. W. Perturbation of myo-inositol-1,4,5-
Stomatogastric neuromuscular systems were dissected and prepared for trisphosphate levels during agonist-induced Ca2+ oscillations. Biophys. J. 74,
extracellular nerve recording and stimulation, intracellular neuron 523531 (1998).
recording, and measurements of muscle contraction using standard 14. Wang, X.-J. Calcium coding and adaptive temporal computation in cortical
pyramidal neurons. J. Neurophysiol. 79, 15491566 (1998).
techniques25,34,35,39. Nerve recording and stimulation were performed 15. Pivovarova, N. B., Hangpaisan, J., Andrews, S. B. & Friel, D. D.
using suction electrodes or pin electrodes insulated with petroleum jelly, Depolarization-induced mitochondrial Ca accumulation in sympathetic
intracellular recordings were made with glass microelectrodes (filled neurons: spatial and temporal characteristics. J. Neurosci. 19, 63726384
with 0.55 M K2SO4, 20 mM KCl) and an Axoclamp 2A (Foster City, Cal- (1999).
16. Evans, P. D. & Siegler, M. V. S. Octopamine mediated relaxation of
ifornia), and contractions were measured with a Harvard Apparatus maintained and catch tension in locust skeletal muscle. J. Physiol. (Lond.)
(South Natick, Massachusetts) 60-3000 isotonic muscle transducer. 324, 93112 (1982).
Muscle rest length was maintained at approximately physiological 17. Weiss, K. R. et al. Peptidergic cotransmission in Aplysia: functional
levels. Muscle loading was determined by inducing single contrac- implications for rhythmic behaviors. Experientia 48, 456463 (1992).
tions and adjusting muscle load to achieve the maximum contraction 18. Clarac, F. & Cattaert, D. Invertebrate presynaptic inhibition and motor
control. Exp. Brain Res. 112, 163180 (1996).
amplitude consistent with full relaxation after the stimulation. A sup- 19. Stevenson, P. A. & Meuser, S. Octopaminergic innervation and modulation
port bar was then placed under the transducer arm to prevent muscle of a locust flight steering muscle. J. Exp. Biol. 200, 633642 (1997).
overstretching. In all cases, the lateral ventricular nerve, which con- 20. Fields, D. R. & Nelson, P. G. Resonant activation of calcium signal
tains both the PD and PY neuron axons, was stimulated to induce transduction in neurons. J. Neurobiol. 25, 281293 (1993).
21. Sheng, H. Z., Fields, R. D. & Nelson, P.G. Specific regulation of immediate
muscle contractions. There are two PD neurons and six to eight PY early genes by patterned neuronal activity. J. Neurosci. Res. 35, 459467
neurons. To ensure activation of all axons innervating a muscle, stim- (1993).
ulation amplitude was incrementally increased using single-burst 22. Itoh, K., Stevens, B., Schachner, M. & Fields, R. D. Regulated expression of the
stimulations until increases in contraction amplitude ceased. No neu- neural cell adhesion molecule L1 by specific patterns of neural impulses.
romodulatory axons are known to innervate the PD and PY muscles Science 270, 13691372 (1995).
23. Li, W., Llopis, J., Whitney, M., Zlokarnik, G. & Tsien, R. Y. Cell-permeant
and no pyloric neurons are known to contain modulatory co-trans- caged InsP3 ester shows that Ca2+ spike frequency can optimize gene
mitters, Furthermore, comparisons of contractions before and after expression. Nature 392, 936941 (1998).
long stimuli revealed no modulation by the nerve. We therefore believe 24. Dolmetsch, R. E., Xu, K. & Lewis, R. S. Calcium oscillations increase the
that our data reflect simple classical neuromuscular innervation and efficiency and specificity of gene expression. Nature 392, 933936
muscle contraction. (1998).
25. Selverston, A. I., Russell, D. F., Miller, J. P. & King, D. G. The stomatogastric
Data were digitized with a Cambridge Electronic Design (CED, Cam- nervous system: Structure and function of a small neural network. Prog.
bridge, UK) 1401 plus and analyzed using CED and Kaleidagraph (Read- Neurobiol. 7, 215290 (1976).
ing, Pennsylvania) software; figures were prepared in Canvas (Miami, 26. Govind, C. K., Atwood, H. L. & Maynard, D. M. Innervation and neuro-
Florida). In Fig. 1, the nerve was stimulated using a World Precision muscular physiology of intrinsic foregut muscles in the blue crab and spiny
lobster. J. Comp. Physiol. A 96, 185204 (1975).
Instruments (WPI: Sarasota, Florida) stimulus isolation unit and a Grass 27. Maynard, D. M. & Selverston, A. I. Organization of the stomatogastric
(Quincy, Massachusetts) S48 stimulator; in Figs. 24, nerves were stim- ganglion of the spiny lobster. IV. The pyloric system. J. Comp. Physiol. A 100,
ulated using a WPI Pulsemaster A300 stimulator driven by the CED. 161182 (1975).

articles
28. Sigvardt, K. A. & Mulloney, B. Sensory alterations of motor patterns in the 40. Beltz, B. et al. Serotonergic innervation and modulation of the stomatogastric
stomatogastric nervous system of the spiny lobster Panulirus interruptus. ganglion of three decapod crustaceans (Panulirus interruptus, Homarus
J. Exp. Biol. 97, 137152 (1982). americanus, and Cancer irroratus). J. Exp. Biol. 109, 3554 (1984).
29. Kennedy, M. B. & Marder, E. in An Introduction to Molecular Neurobiology 41. Flamm, R. E. & Harris-Warrick, R. M. Aminergic modulation in the
(ed. Hall, Z. H.) 463495 (Sinauer, Sunderland, Massachusetts, 1992). lobster stomatogastric ganglion. I. Effects on the motor pattern and
30. Turrigiano, G. G. & Heinzel, H.-G. in Dynamic Biological Networks. The activity of neurons within the pyloric circuit. J. Neurophysiol. 55, 847865
Stomatogastric Nervous System (eds. Harris-Warrick, R. M., Marder, E., (1986).
Selverston, A. I. & Moulins, M.) 197220 (MIT Press, Cambridge, 42. Hooper, S. L. & Marder, E. Modulation of the lobster pyloric rhythm by the
Massachusetts, 1992). peptide proctolin. J. Neurosci. 7, 20972112 (1987).
31. Meyrand, P. & Moulins, M. Myogenic oscillatory activity in the pyloric 43. Nusbaum, M. P. & Marder, E. A neuronal role for a crustacean red pigment
rhythmic motor system of Crustacea. J. Comp. Physiol. A 158, 489503 (1986). concentrating hormone-like peptide: neuromodulation of the pyloric
32. Meyrand, P. & Marder, E. Matching neural and muscle oscillators: control by rhythm in the crab, Cancer borealis. J. Exp. Biol. 135, 165181 (1988).
FMRFamide-like peptides. J. Neurosci. 11, 11501161 (1991). 44. Dickinson, P. S. & Nagy, F. Control of a central pattern generator by an
33. Jorge-Rivera, J. C. & Marder, E. TNRNFLRFamide and SDRNFLRFamide identified modulatory interneurone in Crustacea. II. Induction and
modulate muscles of the stomatogastric system of the crab Cancer borealis. modification of plateau properties in pyloric neurones. J. Exp. Biol. 105,
J. Comp. Physiol. A 179, 741751 (1996). 5982 (1983).
34. Morris, L. G. & Hooper, S. L. Muscle response to changing neuronal input in 45. Nusbaum, M. P. & Marder, E. A modulatory proctolin-containing neuron
the lobster (Panulirus interruptus) stomatogastric system: Spike number vs. (MPN). II. State-dependent modulation of rhythmic motor activity.
spike frequency dependent domains. J. Neurosci. 17, 59565971 (1997). J. Neurosci. 9, 16001607 (1989).
35. Morris, L. G. & Hooper, S. L. Muscle response to changing neuronal input in 46. Katz, P. S. & Harris-Warrick, R. M. Neuromodulation of the crab pyloric
the lobster (Panulirus interruptus) stomatogastric system: Slow muscle central pattern generator by serotonergic/cholinergic proprioceptive
properties can transform rhythmic input into tonic output. J. Neurosci. 18, afferents. J. Neurosci. 10, 14951512 (1990).
34333442 (1998). 47. Norris, B. J., Coleman, M. J. & Nusbaum, M. P. Pyloric motor pattern
36. Mulloney, B. Organization of the stomatogastric ganglion of the spiny lobster. modification by a newly identified projection neuron in the crab
V. Coordination of the gastric and pyloric systems. J. Comp. Physiol. A 122, stomatogastric nervous system. J. Neurophysiol. 75, 97108 (1996).
227240 (1977). 48. Blitz, D. M. et al. Different proctolin neurons elicit distinct motor patterns
37. Russell, D. F. & Hartline, D. K. A multiaction synapse evoking both EPSPs from a multifunctional neuronal network. J. Neurosci. 19, 54495463
and enhancement of endogenous bursting. Brain Res. 223, 1938 (1981). (1999).
38. Claiborne, B. J. & Selverston, A. I. Localization of stomatogastric IV neuron 49. Liu, Z., Golowasch, J., Marder, E. & Abbott, L. F. A model neuron with
cell bodies in lobster brain. J. Comp. Physiol. A 154, 2732 (1984). activity-dependent conductances regulated by multiple calcium sensors.
39. Maynard, D. M. & Dando, M. R. The structure of the stomatogastric J. Neurosci. 18, 23092320 (1998).
neuromuscular system in Callinectes sapidus, Homarus americanus, and 50. Turrigiano, G., Abbott, L. F. & Marder, E. Activity-dependent changes in the
Panulirus argus (decapoda crustacea). Philos. Trans. R. Soc. Lond. B Biol. Sci. intrinsic electrical properties of cultured neurons. Science 264, 974977
268, 161220 (1974). (1994).

articles
Microsaccadic eye movements and

firing of single cells in the striate
cortex of macaque monkeys
Susana Martinez-Conde, Stephen L. Macknik and David H. Hubel
Dept. of Neurobiology, Harvard Medical School, 220 Longwood Avenue, Boston, Massachusetts 02115, USA
Correspondence should be addressed to S.M.-C. (smart@hms.harvard.edu)
When viewing a stationary object, we unconsciously make small, involuntary eye movements or
microsaccades. If displacements of the retinal image are prevented, the image quickly fades from
perception. To understand how microsaccades sustain perception, we studied their relationship to

the firing of cells in primary visual cortex (V1). We tracked eye movements and recorded from V1
cells as macaque monkeys fixated. When an optimally oriented line was centered over a cells recep-
tive field, activity increased after microsaccades. Moreover, microsaccades were better correlated
with bursts of spikes than with either single spikes or instantaneous firing rate. These findings may
help explain maintenance of perception during normal visual fixation.
Our interest in microsaccades and burst firing grew out of an became clear that the cells firing and the monkeys microsac-
attempt to study the activity of single cortical cells in an awake cades were correlated, and that the eye movements tended to
monkey during free viewing of a visual scene. We made con- be associated with bursts of spikes.
tinuous recordings of eye position while recording spikes from This paper is devoted to analyzing this relationship between
a cortical cell. We marked on a video screen the position of the microsaccades and spikes. The results are based on recordings
eyes a suitable length of time before each spike (eye-position- from 258 cells in 3 monkeys. We asked how often, with or with-
corrected reverse correlation). If the scene viewed by the mon- out a stimulus, a microsaccade was followed by spike activity and
key consisted, for example, of a large bright circle on a dark whether any suppression of activity accompanied or followed a
background, we expected the cell to fire whenever its receptive microsaccade. We also asked the converse question: how often
field was crossed by an appropriately oriented contour 1. To were spikes preceded by microsaccades, and how would the cor-
sample the scene evenly, we had the animal fixate on a spot relation be influenced by whether one looks at single spikes,
whose position changed at random every few seconds. For instantaneous firing rates or bursts of spikes. To ask these ques-
some cells, we could see a clear correlation between the stim- tions meaningfully, we began by defining the terms microsac-
ulus and the eye positions that were spike related, but for other cade and burst.
cells, the constellation of spots marking spike-associated eye
positions showed little or no relationship to the scene Analysis of microsaccades
(Fig. 1a and b). During our recordings, we noticed that for cer- In most studies of longer-range eye movements (movements out-
tain gaze locations, cells tended to fire in bursts rather than side the scope of this study), determining length, direction and
trains of random spikes. When we filtered our records to exam- velocity of the movements is relatively simple. Large saccades
ine only burst firing, we saw a much clearer correlation between reach velocities in the hundreds of degrees per second, and so a
the cells activity and the stimulus (Fig. 1c and d). Evidently, straightforward velocity detector can be used to determine when
any tendency for cells to fire in high-frequency bursts was and where each saccade begins and ends. Moreover, the starting
enhanced by the contours of the stimulus. This project and end positions of the eye movements are generally determined
stemmed in part from our suspicion that the bursts were pro- in advance. With microsaccades, starting and end positions and
duced largely in response to microsaccades that occurred dur- timing are relatively random, and their velocities are low. Previ-
ing the intermittent periods of visual fixation. Here we examine ous studies identify microsaccades with instantaneous velocity
the relationship between microsaccades and various firing pat- thresholds2,3 set to fairly high levels (10 per second) to protect
terns of V1 cells, specifically, single spikes, instantaneous fir- the analysis from noise. By using the fastest portion of the
ing rates and spike bursts of various lengths. microsaccade to categorize them in terms of size and speed, these
studies neglected portions of the microsaccade at speeds of less
RESULTS than 10 per second, a considerable portion of a microsaccade
We stimulated each V1 cell with an optimally oriented bright, because of its slow speed and small size. In addition, these stud-
stationary bar centered over the cells receptive field while the ies measured either the horizontal or vertical directional com-
monkey fixated on a spot of light on another part of the ponent of microsaccades, but not both. Thus the size of obliquely
screen. From simply listening to the cells firing while watch- directional microsaccades was underestimated by as much as 2
ing the eye position on the computer screen, it immediately (or 41% of the actual size).

articles
The problem is made more diffi-

cult by confusion as to what, exactly,
constitutes a microsaccade. Our new
algorithm depends on both freedom
from bias as to microsaccade direction
and accurate assessment of total size
and speed of eye movements to help
separate out drifts and blinks. Our
algorithm therefore expresses eye
position in polar coordinates, uses low
velocity thresholds and is free of
directional bias (see Methods). We
defined microsaccades to be those eye
movements that had a total angular
subtense between three minutes and
two degrees of arc, had an instanta- Fig. 1. Responses of a cell located within the operculum of striate cortex to a white 8 circle.
neous velocity (measured each mil- The upper left inset shows the fovea (small circle) in relationship to the cells receptive field. The
lisecond) that never fell below 3 per fixation spot (fovea) is surrounded by a fixation window, 2 2. (a, b) Each pixel represents the
eye position associated with a single spike. (c, d) The same data, filtered so that each pixel rep-
second and never varied from a
resents a burst of 4 or more spikes occurring within an interval of 20 ms at a delay of 35 ms.
straight trajectory by more than 15 (b, d) Same data as (a, c), except that the circle is removed to reveal the pixels. Duration of the
in any direction. Microsaccades were recording was about 40 min.
recorded from two monkeys during
several two-second intervals (Fig. 2).
Probability of discharges following microsaccades records to see if the effect of microsaccadic suppression had
We set out first to determine whether microsaccades were corre- been drowned out by the averaging process, and found that
lated with modulations in neural activity, and if the modulations only 6 of the 246 cells showed clear suppression of firing asso-
were stimulus driven or caused by the microsaccades themselves ciated with microsaccade onset. In all six cells, the suppression
and unrelated to the visual stimulus. preceded a period of excitation.
In the main portion of our analysis, we examined 246 cells Average probability of spikes recorded from 45 cells after
from 2 monkeys. After the beginning of a microsaccade when microsaccades without any visual stimulus (the monitor was
an optimal stimulus was centered over the cells receptive field, black except for the fixation point) revealed no correlation
average spike probability showed a clear but small increase of between the microsaccades and modulation of neural activity
about 0.5%, peaking at about 70 ms (Fig. 3a, upper curve). (Fig. 3a, lower curve), indicating that microsaccade-induced
The average did not show suppression of activity associated activity in V1 was caused by microsaccades driving V1 neu-
with the onset of microsaccades. We examined the individual rons by moving their receptive fields over stationary stimuli,
and not by direct excitation from
the motor system.
We found similar increases in
neural firing correlated with the
end (as opposed to the beginning)
Fig. 2. Spikes and eye positions

recorded during several two-second tri-
als from two monkeys. In each panel,
the upper (green) and middle (blue)
traces represent x and y eye position
(raw data); the lowermost record (red)
represents amplitude of each microsac-
cade after processing to the same scale
as the x and y plots. Each peak in the
lower record thus represents a
microsaccade. (Its height indicates the
size of the microsaccade.) On the
whole, there is a good agreement
between the deflections in the eye posi-
tion traces (green and blue) and the
deflections in the microsaccade plot
(red). Where possible discrepancies
occur, we placed a red dot under the
panel. Notice that frequency of eye
movements can vary between monkeys.
Vertical black lines represent spikes.

articles
Fig. 3. Correlation between microsaccades and

spikes. (a) Probability of a spike after the start of a a
microsaccade. Starts of all microsaccades were aligned
With stimulus
Spike probability
(vertical line). The probability of spikes after microsac-
cades increased in the presence of the stimulus,
whereas it did not change if the stimulus was absent.
(b) Probability that a microsaccade preceded any given
Without stimulus
spike. (All spikes were aligned at the vertical line.)
b
Time after start of microsaccade (ms)
Microsaccade probability
of microsaccades. The result held true even
when we correlated neural activity with each
individual millisecond of each microsaccade in Spike
flight (data not shown).
We ruled out the possibility that the
observed excitation after microsaccades was
caused by flicker on the stimulus monitor by
recording from 12 additional neurons from a
third monkey using a bar projected from a Time before spike (ms)
tungsten incandescent bulb as a visual stimu-
lus. We again found excitation after microsac-
cades, ruling out the possibility that microsaccade-related activity perpendicular to the orientation of the cells receptive field might
was due to flicker in our video display. then preferentially activate the cell. We found, however, no clear
relationship between microsaccade direction and response mag-
Probability that a spike was preceded by a microsaccade nitude. This finding may have been due to the aperture problem:
We next computed the reverse correlation between spikes and a V1 cell cannot distinguish between a slow microsaccade that
preceding microsaccades, that is, the probability of a microsac- moves the stimulus perpendicularly across its receptive field and
cade preceding the discharge for the 246 cells from the main por- a fast microsaccade that moves the stimulus in an oblique direc-
tion of our study (Fig. 3b). The peak probability of spikes tion to the receptive field. Thus two unlikely events would need
correlated with previous microsaccades was enhanced by 9%, to occur simultaneously for microsaccade direction to be
many times the peak probability seen with the forward correlation involved: the stimulus would need to be perfectly aligned with
(compare with Fig. 3a). the angle of the receptive field, and the microsaccade direction
We presume that the difference arose because the reverse cor- would need to be perfectly parallel to the angle of the receptive
releation (Fig. 3b) considered only those microsaccades that were field. In addition, it may be that only a fraction of microsaccades
effective, whereas the forward correlation (Fig. 3a) included all activated area V1 cells effectively because of imperfectly main-
microsaccades, regardless of
their effectiveness. We con-
sidered whether a microsac- a b
cade would be rendered
Probability of microsaccade
ineffective when its direction

was parallel to the stimulus
before spike pair
orientation. Microsaccades
Fig. 4. Instantaneous firing-

rate analysis. (a) Hypothetical Instantaneous firing-rate
train of spikes. The interspike analysis (ISIs in ms)
intervals (ISIs) were calculated
for each successive pair of Interval between spike pair (ms) Time before spike pair (ms)
spikes and expressed in mil-
liseconds (below each pair). c d
Spikes are represented as verti-
cal lines. (b) Three-dimensional
plot of the probability of a

microsaccade before a pair of
before spike pair
spikes, as a function of latency

)
and ISI, for a single V1 cell. Int ms
er v a ir (
(c) Average probability of a al
b ep
pa etwe pik
microsaccade before a pair of ir ( en r es
spikes for all cells, as a function ms sp efo
) ik e eb
of ISI, at the average peak Tim
latency for the population of
cells (65 ms). (d) Normalized
Interval between spike pair (ms) Probability of microsaccade before spike pair
contour plot of the data in (b).

articles
Fig. 5. Burst analysis. (a) The same train of

a Burst analysis
(with an ISI of 30 ms)
spikes as in Fig. 4a, grouped into bursts. An ISI
of 30 ms has been chosen arbitrarily for this
example. Because the first spike in the train
was separated from the second one by 20 ms,
and this interval was smaller than 30 ms, both
spikes were assigned to the same burst. The
third spike occured 60 ms after the second
spike, and so was part of another burst, and so
on. (b) Three-dimensional plot of burst corre-
lations using bursts of one spike (lone spikes)
Burst of 2 Burst of 1 Burst of 5 for the same V1 cell as in Figs. 4b and d.
(lone spike) (c) Three-dimensional plot of burst correla-
b Lone spikes
c Bursts with 5 spikes
tions using bursts of five spikes for the same
cell. The probability of a microsaccade before a
burst was plotted as a function of latency (time
before the first spike in each burst) and ISI.
(d, e) Normalized contour plots of the data in
before burst
before burst
(b, c). For this cell, bursts of five spikes were

best correlated with previous microsaccades.
Bursts of spikes
ISI Time before burst (ms) ISI Because previous experiments (Fig. 1) sug-
Time before burst (ms)
gested that bursts may convey more reliable
d e information about the visual scene than
single spikes1, we were especially interested
s )
s)
in studying the relationship between bursts
m (m
s t( t (of one or more spikes) and preceding
r rs
bu bu microsaccades. We assigned each spike
r e r e
IS fo fo
I be IS
I be uniquely to a burst by determining the
e e
m
Ti Ti
m interval between it and the spikes preced-
ing and following: two spikes with a given
ISI (or less) were thus considered members
Probability of microsaccade before burst Probability of microsaccade before burst
of a single burst, and two spikes with an
interval greater than the given ISI were con-
tained fixation: some microsaccades may have begun and ended sidered elements of separate bursts (S. M. Smirnakis, D. K. War-
with the stimulus outside the receptive field, without ever cross- land, M.J. Berry and M. Meister, Neural Information Coding
ing it. Such occurrences would diminish the average enhance- Conference, Snowbird, Utah, 1997; Fig. 5a). The ISI chosen is cru-
ment of firing by microsaccades but not affect the incidence of cial in determining the overall distribution of burst lengths (num-
microsaccades preceding cell firing. ber of spikes per burst). Because we could not know in advance
the most appropriate ISI for any given cell, we compared, for all
Instantaneous firing rate the records and for all the cells, all ISIs between 1 and 100 ms and
We examined the relationship between the interspike intervals measured the probability of a microsaccade occurring some time
(ISIs) that separated pairs of successive spikes (within the range (latency; 200 ms) before the first spike in each burst. We then
of 1100 ms) and the microsaccades that preceded these spike sorted all bursts according to length (numbers of spikes per burst),
pairs by up to 200 ms (Fig. 4a). We define instantaneous fir- which ranged from 1 (lone spike) to 20. We then constructed 20
ing rate as the inverse of interspike interval (1/ISI)4. Plots of (for each burst length) three-dimensional arrays; from each of
the probability of a microsaccade preceding the first of two these we plotted the probability of a microsaccade preceding a burst
successive spikes with a given ISI (Fig. 4b and d) demonstrate against latency ( 200 ms) and ISIs ( 100 ms). These plots allowed
that microsaccades were more closely linked to high instanta- us to assess the probability of a microsaccade as a function of burst
neous firing rates (small ISIs) than to slow rates. The proba- size, latency and ISI.
bility of a microsaccade was maximal about 44 ms before a We show three-dimensional plots and normalized contour
spike pair. At higher ISIs, the probability of a microsaccade plots for bursts of one (Fig. 5b) and five (Fig. 5c) spikes for the
before a pair of spikes did not drop to zero, presumably because same cell as in Figs. 4b and d. Burst of one, or lone spikes, were
of spontaneous firing. relatively poor indicators of previous microsaccades. Bursts of 5
We plotted average microsaccade probability against ISI were more closely correlated with previous microsaccades in this
for the cell population by averaging all three-dimensional cell, and had a peak correlation of almost one. Peak probability
graphs and slicing the result parallel to the probabilityISI varied with latency in all cells, whereas peak latencies remained
plane at the average peak latency, 65 ms (Fig. 4c). Pairs of stable as burst size varied and were consistent with V1 responses.
spikes were most closely related to preceding microsaccades For this cell, bursts of 5 spikes were the best indicators that a
when the interspike interval was short. Averaged across all microsaccade had occured about 44 ms before.
cells, the peak probability of a microsaccade preceding a pair We next compared the strengths of the associations for all the
of spikes was 0.42. different burst lengths examined, regardless of ISI or latency

articles
Fig. 6. (a) Peak probability that a microsaccade occurred before bursts,

as a function of burst size, for all latencies and ISIs. Same cell as in Figs. 4 a
and 5. (b) Probability that a burst was preceded by a microsaccade, as a
function of ISI, at the peak latency (44 ms; same cell as in a). The two-
Peak probability of a microsaccade

dimensional plot was obtained from the three-dimensional graphs in
Figs. 5b and c.
(Fig. 6a; same cell). The optimal burst size was clearly five, with a
fall-off for shorter and longer bursts. But why did the correlation
decrease for bursts longer than five? The occurrence of longer bursts
during the experiment that were not well correlated with microsac-
cades led us to wonder if larger eye movements omitted from our
analysis were associated with longer bursts. To test this, we examined
all eye movements longer than three minutes of arc and found a
clear correlation with all bursts measured (data not shown).
Burst size (number of spikes)
For the same cell, we then plotted probability against ISI for
bursts of 1 and 5 spikes at a latency of 44 ms (the cells optimal
latency; Fig. 6b). This amounted to cutting Fig. 5b and c by planes b

normal to the latency axis at 44 ms. For single spikes, it showed an
almost constant probability of about 0.4. For bursts of five spikes,
Bursts of 5 spikes
Probability of a microsaccade
the probability variations were presumably explained by the ten-
dency of spikes to occur in clumps separated by periods of rela-
tive quiescence, often lasting hundreds of milliseconds: when ISIs
that defined bursts were small, they tended to cut these clumps
into fragments that were poorly correlated with microsaccades. Single spikes
ISIs greater than the optimum value did not affect the probabil-
ities unless sufficiently large that successive clumps coalesced. To
sum up, burst size and latency were highly correlated with pre-
vious microsaccades, whereas the ISI used to define bursts was Tighter bursts
less crucial, at least for values greater than 2030 ms.
The correlation with preceding microsaccades was always
higher for bursts, defined in terms of ISIs, than for pairs of spikes
ISI used when assigning spikes to bursts
that determined instantaneous firing rates. This is demonstrated
by plotting for all ISIs and latencies the peak probability of a
microsaccade before the optimum burst size against the peak for all ISIs and latencies the distribution of burst sizes associated
probability of a microsaccade before the optimum instantaneous with peak probabilities (Fig. 8). Peak burst sizes showed a wide
firing rate (Fig. 7). The difference in the probabilities obtained spread, with an average of 11 ( 5, s.d.) spikes per burst and a
by the two methods was significant (p < 0.001; t-test). Thus for substantial fall-off for short bursts. The average latency associ-
every cell tested, burst configuration indicated a previous ated with microsaccade activity was 65 46 ms, and the average
microsaccade with more reliability than the best possible instan- peak ISI was 56 29 ms.
taneous spike rate.
From Fig. 6a, it is clear that a burst size of five supplied by far DISCUSSION
the best indicator of a previous microsaccade for the cell under The phenomenon of fading of a stabilized image comes as a great
consideration. We were curious to learn the variability of such a surprise to someone who learns of it for the first time; when first
preference across our sample population.We therefore plotted discovered5,6 in the early 1950s, it was little short of astonishing.
Determined attempts to reproduce the fading by careful fixation
often fail, especially for objects in or near the fovea, and observers
are quite unaware of the microsaccades whose occurrence large-
Peak probability from burst analysis
ly accounts for failure to perceive fading of stationary images.

It was perhaps equally surprising to learn, in the late '50s, that
cells in the mammalian primary visual cortex respond far more
vigorously to moving stimuli than to stationary ones7,8. These
first studies of visual cortex pointed out the possible relationship
between this and the fading of stabilized images. Our sensory
Fig. 7. Peak probability that a spike was preceded by a microsaccade for

each cell. Peak probabilities determined by instantaneous firing rate
were plotted against determinations using optimum burst sizes, across
all latencies and ISIs. For every cell, analysis from the burst pattern pre-
dicted preceding microsaccades better than determinations based on
Peak probability from firing-rate analysis
instantaneous firing rate.

articles
a byproduct of oculomotor mechanisms. Further, its overall effect

is excitatory. Our results are consistent with previous findings that
much of the variability of responses in striate cortex of awake
behaving monkeys during visual stimulation is due to the effects
of microsaccades10. This is not to say that visual responses driven by
Number of cells
microsaccades are likely to be restricted to V1 cells: one would

expect to find similar results both earlier and later in the visual
pathway.
We were surprised to find so little evidence for microsaccade-
related suppression, as there is ample evidence of suppression in
the much faster long-range saccades1116. Visual responses in pri-
mate area V1 are suppressed when the receptive field of a single
cell crosses a stationary bar during a saccade1214. This suppres-
sion occurs only in cells that do not respond during fixation to
stimuli moving at saccadic velocities. The difference between these
Burst size yielding peak probability results and ours is probably related to the ability of most neurons
to respond to the stimulus velocities evoked by microsaccades
Fig. 8. Distribution of burst sizes yielding peak probabilities across the pop- (30 per s) while failing to respond to stimuli moving at large
ulation of cells. Optimal burst sizes tended to be greater than three spikes. saccadic velocities (typically, over 300 per s). Hypothetically, sup-
pression associated with microsaccades could block awareness of
fixational eye movements, which produce displacements of images
systems seem to be organized on the principle that, whether we across the retina sufficient to be easily seen if produced by a mov-
are predators or prey, changing stimuli are essential for survival ing object. We saw microsaccade-related suppression in only 7 of
and that sensory adaptationthe decline with time of a response the 258 cells we studied, and the suppression was always followed
to a maintained stimulusguarantees that a novel stimulus will by excitation. It is possible that our experimental protocol made
stand out, especially if it moves. The normal inability to see ones excitation easier to detect than inhibition. In any case, we found no
own retinal blood vessels or optic disc is presumably also related conclusive mechanism to explain unawareness of microsaccades,
to this process9. Ironically, the process works so well that some at least at the level of the striate cortex.
further mechanism seems to be required to overcome the fading, The enhancement of firing in V1 cells by microsaccades that
so that we can see a stationary scene despite adaptation. we observed is at odds with similar recordings of cells in mon-
If involuntary eye movements during fixation are crucial role key V1 during fixation with a small grating covering the recep-
in maintaining the visibility of stationary scenes, we should be tive field 3. This study revealed enhancement of firing after
able to observe the relationship in awake alert animals by record- microsaccades in only 6 of 35 cells, suppression of firing in 13
ing eye movements while observing the firing of single cells in cells and no changes in the remaining 16 (but enhancement of
the presence of an adequate visual stimulus. This was the pur- firing downstream, in areas V2, V4 and IT). The failure to see
pose of the present study. microsaccade-related activation in V1 surprised us, given the
Quantitatively correlating microsaccades and bursts of spikes marked elevation of firing rate we consistently saw whenever a
has been a challenging problem. Under the term bursts, we stationary bar was placed in the receptive field and the clear asso-
included both clustered activity
(short, high-frequency trains of two to Single session 239 sessions
eight or more spikes) and bursty fir-
ing, consisting of highly irregular
longer-lasting periods of rapid activity
separated by irregular silent periods.
Both types of firing were described in
Microsaccade velocity
1958 using these terms 7 . Here,

(degrees per s)
although recognizing that their origins

may differ, we lumped the two togeth-
er, simply referring to both as bursts.
We found an excitatory effect of
microsaccades on V1 cells. Such effects
were seen in all 258 cells we examined.
Correlating a cells activity with the
beginning of the microsaccades (or its
end, or any point between) showed that
the probability of spike production
increased when a stimulus, a stationary
bar, was centered over the receptive Microsaccade magnitude
(degrees of visual angle)
field. No such correlation was seen in
the absence of the stimulus (Fig. 3a), Fig. 9. The main sequence analysis. The left panel represents 523 microsaccades measured while study-
and we conclude that microsaccade- ing a single neuron; it plots the linear regression of these microsaccades and 95% confidence intervals.
related activity is a direct product of The right panel shows the linear regression of the main sequence from 691,899 microsaccades recorded
visual-sensory mechanisms rather than from 239 cells from a single monkey (99% confidence intervals were obscured by the regression line).

articles
ciation, to simple inspection, between eye movements and fir- ability that a microsaccade precedes a single burst of arbitrary size
ing. The discrepancy in results may be due to differences in meth- (say of three spikes) is greater than the sum of the probabilities
ods. We stimulated our cells with stationary bars of light, whereas associated with three spikes considered separately. In our results,
they used circular patches of oriented gratings that sometimes this seemed, on average, not to be so. For example, if a given spike
were binocularly rivalrous, with orthogonal orientations in the was related to a previous microsaccade with a probability of over
two eyes. Their cell population in V1 was much smaller, and they 0.3 (Fig. 3b), simple addition should predict that a rapid succes-
sampled eye position at 200 Hz, interpolating 80% of their eye sion of three spikes should give a probability of three times 0.3,
movement data to simulate a sampling rate of 1 KHz. In addi- and on average this is roughly what we found. There may thus be
tion, our monkeys fixated continuously during the entire record- nothing magic about bursts: by definition, several spikes in quick
ing session (nothing in our experimental protocol signaled the succession necessarily constitute a burst.
beginning or end of a trial to the monkey), so the visual stimulus We cannot, on the other hand, rule out the existence of presy-
remained on the screen throughout. Their monkeys, on the other naptic facilitation in some synapses. At a presynaptic level, both
hand, performed a visual discrimination task with discrete trials synaptic facilitation and depression are described in vitro21,22.
that were separated by blank screens. Regardless of these effects, however, rapidly sequential EPSPs should
In the present analysis, we did not look for effects of slow drifts sum postsynaptically, so that even when synaptic efficacy is reduced
in eye position on the firing of cells, and cannot rule out a possible in size, as in presynaptic depression, multiple EPSPs should improve
role for them in the prevention of image fading. The relative role of the chances of postsynaptic signaling through classical temporal
microsaccades and drifts has been a subject of considerable dis- summation. For a pair of spikes, a clear example of paired synaptic
cussion in the literature. Several studies17,18 address both types of facilitation has been shown in vivo for retinogeniculate synapses23.
movement but do not isolate microsaccades and drifts from each Thus, although we have no direct evidence pertaining to cortical
other. The strongest claim against a role of microsaccades in pre- synapses, given the simplest assumptions of temporal summation,
serving visual perception19 is mainly based on two ideas. First, visu- it seems reasonable to expect a burst of spikes to be more reliable
al thresholds may be elevated during saccades. This is true of than single spikes in influencing a postsynaptic cell.
long-range saccades, but the evidence for suppression during the Equally interesting is the expectation that each microsaccade
much slower microsaccades is controversial, and we found little will produce synchronous bursts in many cells. Receptive fields of
physiological evidence for it in the present study. Second, subjects neighboring cells in striate cortex overlap extensively and share
can be trained to suppress their microsaccades without experi- the same orientation preference. Consequently, as a stimulus is
encing fading of images. The possibility that drifts alone can pre- swept across their receptive fields, many cells should fire more or
vent fading does not prove that microsaccades play no part. less synchronously, in bursts that can be expected to overlap in
Moreover, if one fixates steadily on a point while paying attention time. The result should be strong spatial summation of messages
to a small (and, preferably, low contrast) object in the periphery carried by axons converging at the next stage. Microsaccades would
of the visual field, the object will disappear almost immediately20. thus seem to represent an ingenious mechanism for enforcing and
If such precise fixation eliminates microsaccades, this would seem refreshing information coming from stationary visual stimuli.
to support their role in preventing fading. A careful analysis of the It is furthermore possible that microsaccades may be impor-
physiological effects of slow drifts in eye position would be useful, tant for allowing us to use latency in our visual discriminations.
and we are currently attempting such an analysis. Latency differences in neural responses are proposed as an encod-
Having determined the probability that a microsaccade is fol- ing mechanism for changes in the magnitude of contrast2426. But
lowed by spike activity, our next step was to ask whether, con- it has remained unclear how the brain could use latency infor-
versely, neural discharges were associated with an increased mation, as the brain cannot know the latency a priori. Microsac-
probability of previous microsaccades. By doing so, we could cades could theoretically be used by the visual system to measure
concentrate on those microsaccades that had been effective in latency, because the brain knows both when the microsaccade
driving the cells but ignore ones that were ineffective. (The stim- motor signal starts and when the visual signal arrives. Relative
ulus can at times drift away from the receptive field, among other latencies of responses to stimuli could then be used to indicate
possibilities.) We found that the probability of microsaccades contrast.
before a spike (Fig. 3b) was much higher than the probability of
a spike after microsaccades (Fig. 3a). The instantaneous firing
rate of each cell, defined as the inverse of the interval between METHODS
the two members of a pair of successive spikes, was still better We used three rhesus monkeys, Macaca mulatta. A midline stainless-steel
head post was mounted by screws over the skull, and a steel recording
than single spikes in indicating a preceding microsaccade.
chamber was mounted over the occipital operculum. We recorded from
Furthermore, bursts of spikes were better correlated with pre- single neurons in area V1 with lacquer-coated electropolished tungsten
ceding microsaccades than single spikes or instantaneous firing electrodes. A visual search coil attached to the sclera of one eye recorded
rates. For most cells, bursts of fewer than four spikes were less the animals eye movements. Standard sterile surgical techniques and
effective than longer bursts, and on average there was no sharp animal care methods are described1,27. The Harvard Medical Area Stand-
optimum size. ing Committee on Animals (protocol #02935) approved all electrophys-
Why did V1 cells show such a marked tendency to fire in bursts iological experimentation.
when microsaccades made their receptive fields move across a
visual stimulus? Here we will not discuss the intrinsic properties of Stimuli and receptive-field mapping. Monkeys were trained to fixate on
a small spot displayed on a monitor at a distance of 58 cm. A monkey
neurons that cause them to fire in bursts, but will instead consid-
was required to keep fixation within a 2 window to receive a drop of
er the teleological significance of such firing. Are bursts useful in juice; eye movements exceeding the windows limits were also recorded.
terms of enhancing the information conveyed by one cell to a The stimuli displayed on the monitor had a luminance of 24.3 cd
postsynaptic cell? Specifically, are spikes arriving at a synaptic ter- per m2, and the monitor background was 3.84 cd per m2. We recorded V1
minal more or less likely to evoke spikes in a postsynaptic cell if over the occipital operculum (eccentricity, 68) or from the folds of
they occur in bursts? One can begin by asking whether the prob- calcarine cortex immediately beneath. Receptive fields were about

articles
0.33 0.5, and thus the same size or larger than many eye movements if a particular burst size and ISI occurred only once and followed a
during fixation. We first mapped the receptive field roughly by hand, microsaccade, the probability of such a unique event was thus 1. We
using computer-generated moving slits whose orientation, size and rate therefore smoothed every three-dimensional array with a boxcar filter
of movement were controlled with a computer mouse. We then used a whose width was 15 elements in the ISI-latency plane; thus every proba-
bar of optimum length, width and orientation that flashed in quick suc- bility was averaged with those corresponding to 7 ms latency and 7
cession in various parts of the visual field to more precisely determine ms ISI. (At the array edges, this boxcar was reduced to fewer elements.)
the position and size by reverse correlation. A stationary bar with these This smoothing removed virtually all the random peaks of probability
optimal parameters of width, length and orientation was then positioned associated with rare bursts.
over the center of the receptive field, and collected spikes were subse-
quently correlated with the monkeys microsaccades. The bar turned on
before data collection and remained on throughout the recording ses-
ACKNOWLEDGEMENTS
We thank Gail Robertson, David Freeman, Michael Lafratta and Frederic Russo
sion for each cell. Within total recording times of 100800 s (mean, 334
s per cell), we recorded eye movements and spikes in consecutive 2-s tri- for technical assistance and Margaret Livingstone, Clay Reid, Richard Born and
als. The beginning and end of each trial were unknown to the animal, Max Snodderly for reading the manuscript and making comments. We also thank
and the fixation point persisted between trials. Murray Sherman, Jonathan Victor, John Maunsell and Guy Orban for their
advice on the analysis and design of the project. This work was supported by
Analysis of microsaccades. We developed an algorithm for determining research (to D.H.H.) and training grants (to S.L.M.) from the National Eye
when the eye was stopped and when it was moving. We regarded an eye Institute. S.M.-C. is a fellow from the MEC-FPI program (Spain).
movement as ended when either its speed dropped below some thresh-
old or its direction changed by more than some predetermined angle.
RECEIVED 25 MAY 1999; ACCEPTED 24 JANUARY 2000

First, we digitally sampled the horizontal (x) and vertical (y) eye tracker
voltages (with 11-bit resolution at a sampling rate of 1 kHz, each A/D unit
corresponding to 0.982 arcmin of visual angle); we then created a table 1. Livingstone, M. S., Freeman, D. C. & Hubel, D. H. Visual responses in V1 of
of change in x and y (dx and dy) computed as the difference between val- freely viewing monkeys. Cold Spring Harb. Symp. Quant. Biol. 61, 2737
ues at successive intervals. After differentiating the data, so that each ele- (1996).
2. Bair, W. & OKeefe, L. P. The influence of fixational eye movements on the
ment then represented the instantaneous velocity of the eye in horizontal
response of neurons in area MT of the macaque. Vis. Neurosci. 15, 779786
and vertical space, we smoothed the processed data with a 31 ms-wide (1998).
unweighted boxcar filter to reduce noise. Varying the width of this filter, we 3. Leopold, D. A. & Logothetis, N. K. Microsaccades differentially modulate
found that averaging each value of dx and dy over a period up to 15 ms neural activity in the striate and extrastriate visual cortex. Exp. Brain Res. 123,
surrounding each sampling interval gave results consistent with catego- 341345 (1998).
4. Bair, W., Koch, C., Newsome, W. & Britten, K. Power spectrum analysis of
rization of microsaccades by inspection of raw data from the eye-move- bursting cells in area MT in the behaving monkey. J. Neurosci. 14, 28702892
ment monitor, the method used in most previous studies. We prepared (1994).
an array of vectors representing the instantaneous direction () and size (r) 5. Riggs, L. A. & Ratliff, F. The effects of counteracting the normal movements
of motion of the eye at each 1-ms interval of the recording. Thus r may of the eye. J. Opt. Soc. Am. 42, 872873 (1952).
be thought of as the instantaneous velocity of the eye in degrees of visual 6. Ditchburn, R. W. & Ginsborg, B. L. Vision with a stabilized retinal image.
Nature 170, 3637 (1952).
space per ms. In addition to and r, we applied a velocity threshold of r < 7. Hubel, D. Cortical unit responses to visual stimuli in non anesthetized cats.
3 per s as a measure of when the eye had stopped, giving a third datum for Am. J. Opthalmol. 46, 110122 (1958).
each interval: a value of 1 (stopped) or 0 (not stopped) was recorded to 8. Hubel, D. H. & Wiesel, T. N. Receptive fields of single neurones in the cats
generate the eye-stopped array. striate cortex. J. Physiol. (Lond.) 148, 574591 (1959).
9. Coppola, D. & Purves, D. The extraordinarily rapid disappearance of entoptic
At each interval, a rate-of-turn indicator checked whether had images. Proc. Natl. Acad. Sci. USA 93, 80018004 (1996).
changed by >15; if so, we considered the eye to have stopped between 10. Gur, M., Beylin, A. & Snodderly, D. M. Response variability of neurons in
intervals and amended the eye-stopped array appropriately. This indi- primary visual cortex (V1) of Alert Monkeys. J. Neurosci. 17, 29142920
cator was important for accuracy of the eye-stopped array, as the eye sel- (1997).
dom had zero velocity even after the data had been smoothedprobably 11. Dodge, R. Visual perception during eye-movements. Psychol. Rev. 7, 454465
(1900).
because of noise and drifts in eye position. During a true saccade, suc- 12. Wurtz, R. H. Visual cortex neurons: response to stimuli during rapid eye
cessive measurements of were similar (aligned in the direction of the movements. Science 162, 11481150 (1968).
overall eye movement). If continually changed, we considered saccades 13. Wurtz, R. H. Comparison of effects of eye movements and stimulus
to be absent. movements on striate cortex neurons of the monkey. J. Neurophysiol. 32,
987994 (1969).
To determine the overall distance and direction over an entire 14. Wurtz, R. H. Response of striate cortex neurons to stimuli during rapid eye
microsaccade (average, 29 ms), we used simple vector addition to gen- movements in the monkey. J. Neurophysiol. 32, 975986 (1969).
erate cumulative distance and directions of the movement (integrated r 15. Bridgeman, B., Hendry, D. & Stark, L. Failure to detect displacement of the
and ) at each interval between eye stops as registered in the eye-stopped visual world during saccadic eye movements. Vision Res. 15, 719722 (1975).
array. At the final 1-ms interval of each microsaccade, values of integrated 16. Macknik, S. L., Fisher, B. D. & Bridgeman, B. Flicker distorts visual space
constancy. Vision Res. 31, 20572064 (1991).
r and integrated were therefore equal to the final distance and direc- 17. Gerrits, H. J. & Vendrik, A. J. The influence of stimulus movements on
tion of the entire eye movement. To distinguish microsaccades from small perception in parafoveal stabilized vision. Vision Res. 14, 175180 (1974).
artifacts or large voluntary saccades or eye blinks, we applied lower and 18. Sansbury, R. V., Skavenski, A. A., Haddad, G. M. & Steinman, R. M. Normal
upper limits of 3 arcmin and 2, to the final value of integrated r. Plots of fixation of eccentric targets. J. Opt. Soc. Am. 63, 612614 (1973).
the microsaccadic main sequence (size versus velocity) showed a linear 19. Steinman, R. M., Haddad, G. M., Skavenski, A. A. & Wyman, D. Miniature
eye movement. Science 181, 810819 (1973).
regression (Fig. 9). 20. Day, E. C. Photoelectric currents in the eye of the fish. Am. J. Physiol. 38,
369398 (1915).
Data processing. We calculated the reverse correlation for microsaccades 21. Magleby, K. L. in Synaptic Function (eds. Edelman, G. M., Gall, W. E. &
preceding either pairs of spikes or bursts of spikes and plotted our results Cowan, W. M.) 2156 (Wiley, New York, 1987).
22. Zucker, R. S. Short-term synaptic plasticity Annu. Rev. Neurosci. 12, 1331
in three dimensions (Figs. 4 and 5). For particularly rare combinations of (1989).
burst length and ISI, small sample size resulted in an artifactually high 23. Usrey, W. M., Reppas, J. B. & Reid, R. C. Paired-spike interactions and synaptic
probability of the burst being preceded by a microsaccade; for instance, efficacy of retinal inputs to the thalamus. Nature 395, 384387 (1998).

articles
Lack of cortical contrast gain

control in human photosensitive
epilepsy
Vittorio Porciatti1, Paolo Bonanni2, Adriana Fiorentini1 and Renzo Guerrini2,3
1 Institute of Neurophysiology, Area Ricerca CNR, 1 via Alfieri, 56100 Pisa, Italy
2 Institute of Developmental Neurology and Psychiatry, University of Pisa and Institute for Medical Research IRCCS Stella Maris, 2 via Giacinti, 56018
Calambrone, Pisa, Italy
3 Division of Neurosciences, Kings College Hospital, University of London, Denmark Hill, London SE5 9RS, UK
Correspondence should be addressed to V.P. (porciatt@in.pi.cnr.it)
Television and video games may be powerful triggers for visually induced epileptic seizures. To
better understand the triggering elements of visual stimuli and cortical mechanisms of
hyperexcitability, we examined eleven patients with idiopathic photosensitive epilepsy by recording
visually evoked potentials (VEPs) in response to temporally modulated patterns of different contrast.
For stimuli of lowmedium, but not high, temporal frequency, the contrast dependence of VEP
amplitude and latency is remarkably abnormal for luminance contrast (blackwhite), but not so for
chromatic contrast (equiluminant redgreen) stimuli. We conclude that cortical mechanisms of con-
trast gain control for pattern stimuli of relatively low temporal frequency and high luminance
contrast are lacking or severely impaired in photosensitive subjects.
Photosensitive epilepsy (PSE) is the most common form of stim- We studied both control subjects and patients with pure idio-
ulus-induced epilepsy1. Its prevalence in children 414 years old pathic PSE whose seizures originated from the occipital lobe9,10.
is substantial (0.5%0.8%), and its incidence is increasing as a VEPs were recorded in response to simple visual patterns
result of the proliferation of television display units and video (black/white and red/green sinusoidal gratings of different con-
games, which may act as triggers2. During a recent television trast), sinusoidally contrast-reversed at various temporal fre-
showing of the Pocket Monsters cartoon in Japan, 685 children quencies. The EEG activity remained in a physiological state
experienced epileptic seizures3,4. The degree of danger inherent throughout the experiment. In patients, the dependence of VEP
in such seizures ranges from nil to potentially life threatening, in amplitude on luminance-contrast for stimuli of relatively low
exceptional cases5. Epileptic seizures induced by photic stimula- temporal frequency (410 Hz) was dramatically altered. Specif-
tion may be primarily generalized (tonic-clonic, myoclonic, ically, at increasing contrast, the VEPs saturated in amplitude
absence seizures) or focal (occipital, with or without spreading and shortened in phase in controls, whereas in patients there was
to other cortical areas). PSE is defined as pure when seizures are little saturation or phase shift. These results suggest that cortical
exclusively photic induced, and is usually idiopathic (with no mechanisms of contrast gain control were lacking or severely
other etiology than a genetic background)6. impaired in PSE.
Because of the complexity and heterogeneity of video-gener-
ated patterns, reports of pattern-induced epilepsies are mainly RESULTS
anecdotal. Oriented lines are considered more powerful than Luminance gratings: effect of temporal frequency
checkerboards in inducing epileptiform EEG changes7, and oscil- In both controls and patients, steady-state VEPs were recorded as a
lating patterns more epileptogenic than static patterns8. Howev- function of temporal frequency of luminance-contrast gratings.
er, the characteristics of the visual stimulus and the cortical Stimulus contrast (90%) and spatial frequency (2 cycles per degree)
dynamics leading to the hypersynchronous neuronal response were chosen to maximize response amplitude11,12. In agreement with
underlying PSE are poorly understood. A better knowledge of previous studies11,12, the form of the temporal function consistent-
the triggering elements of visual stimuli might help in devising ly showed two amplitude peaks (at 38 Hz and 1620 Hz) and a
safer video-generated patterns. local minimum at 1012 Hz in normal subjects (Fig. 1a and c). In
The temporal frequency of an intermittent photic stimulus is PSE patients (Fig. 1b and c), the high temporal-frequency cutoff
crucial for arousing epileptic activity. This suggests that, even for was comparable to that of controls. However, the shape of the tem-
patterned stimuli, the temporal frequency may be critical. How- poral function was more variable. In particular, low- and high-fre-
ever, there is no systematic study on the visually evoked potentials quency amplitude peaks were less defined, and in some subjects,
to temporally modulated patterned stimuli in PSE patients. Anoth- substantial activity was present at intermediate frequencies (1012
er crucial parameter of pattern stimuli is spatial contrast (related to Hz), at which controls responded poorly. Overall, VEP amplitude
the spatial variations in brightness). It is conceivable that contrast was significantly larger in patients than controls
is also a critical parameter for cortical excitability. (Fig. 1c; two-way ANOVA, F1,271 = 6.2, p = 0.012).

articles
a b c contrast, and saturated at about 20%

contrast (Fig. 2a). In patients, the aver-
age contrast function was virtually iden-
tical to that of controls in the low
contrast range (320%; Fig. 2a), indi-
cating comparable contrast threshold
and contrast dependence at low contrast.
Unlike controls, however, the response
did not saturate with increasing contrast
above 20%, and reached abnormally
higher-amplitude values at maximum
d e f contrast (two-way ANOVA, F1,236 = 17.6,
p < 0.001).
The difference in shape of the con-
trast function between the two groups
was evaluated from individual curves
(see below). At higher temporal fre-
quencies (Fig. 2b), the VEP contrast
function was comparable in controls and

patients, and did not show amplitude
saturation.
Average VEP phase showed contrast
dependence in the low temporal-fre-
Fig. 1. Luminance-contrast gratings: effect of temporal frequency. VEP amplitude and phase plots of quency range (Fig. 2c). Because VEPs
individual control subjects (a, d) and patients (b, e). (c, f) Average ( s.e.) VEP amplitude and phase of
were recorded at different temporal fre-
control subjects and patients are shown superimposed.
quencies in individual subjects, data were
normalized by transforming original
phase values into latency values. Phase
The VEP phase progressively lagged with increasing temporal data were first divided by 4 (modulo = 2 rad * 2nd harmonic)
frequency. Both raw data (Fig. 1d and e) and averaged data and then multiplied by the stimulus period in milliseconds.
(Fig. 1f) indicate that the phase plots were similar in controls and Latency values of responses to maximum contrast were normal-
patients. The response latency (in seconds) can be obtained from ized to zero-latency lag, and latencies for all other contrasts were
the slope of the phase plot ( rad / Hz) divided by 4 (modulo = 2 expressed as relative changes. Normalization of latency for
rad * 2nd harmonic; see ref. 12). The average latencies, evaluat- responses to maximum contrast was possible, as VEPs of con-
ed from individual phase plots, did not significantly differ between trols and patients showed comparable latencies at high contrast
controls and patients (106.7 2.5 ms, s. e., versus 107.7 2.2 ms). (see above). The VEP latency of controls progressively lagged
with decreasing contrast (Fig. 2c). In patients, however, the
Luminance gratings: effect of contrast response latency did not slow down with decreasing contrast in
For each subject, contrast functions were evaluated at the low- the medium- to high-contrast range (1090%). The different
and high-frequency peaks of the individual temporal functions latency dependence between controls and patients was evaluated
(Fig. 1). In agreement with previous reports13, the average VEP from individual curves (see below).
amplitude of controls progressively increased with increasing In the higher temporal-frequency range (Fig. 2d), the con-
trast dependence of average VEP phases was virtually identical
a b in controls and patients. The abnormal dependence of VEP
amplitudes and phases on contrast in the low temporal-frequency
range was very consistent among patients. We plotted individual
response curves for both controls and patients (Fig. 3). The indi-
vidual curves of most controls showed amplitude saturation at
high contrasts (Fig. 3a), whereas this was not observed in the
majority of patients (Fig. 3b). To provide a quantitative evaluation
of saturation, a saturation index14 was derived for the individual
amplitude curves. The saturation index is defined as
c d (1 c1/2)/c1/2
where c1/2 is the contrast that elicits responses with half the ampli-
tude of the response at maximum contrast. A value of one corre-
Fig. 2. Luminance-contrast gratings: effect of contrast. Average ( s.e.)

amplitude (a, b) and latency (c, d) for two different ranges of temporal
frequencies. In the 410 Hz range, the contrast dependence differed
remarkably between controls and patients at mediumhigh contrast. In
the 1622 Hz range, the contrast dependence of amplitude (b) and
phase (d) was virtually identical in control subjects and patients.

articles
p < 0.001; PSE > controls). The aver-

a b age amplitude curves actually
showed a slight difference in band-
width (half-height width), but there
was no significant temporal-fre-
e quency group interaction
(F 1,197 = 1.55, p = 0.14), possibly
because of the larger variability of
patient data. The VEP phase (Fig. 4c)
lagged as a function of temporal fre-
d quency, with a comparable slope in
c controls and patients. Average laten-
cies, evaluated from individual phase
plots, tended to be longer in controls
(142.2 7 ms) than in patients
(129.3 6 ms). However, the differ-
ence was not significant.
Contrast functions were evaluat-
ed at individual peak temporal fre-

quencies (range, 36 Hz). In both
controls and patients (Fig. 4b), the
Fig. 3. Luminance-contrast gratings of 410 Hz temporal frequency: effect of contrast in individual sub- VEP amplitude progressively
jects. Amplitude and latency plots of control subjects (a and c) and patients (b and d). (e) Scatter plot of increased with increasing contrast,
saturation and latency indices (see text) for control subjects and patients. and tended to saturate at the highest
(70%90%) contrasts. The average
saturation index of individual
response curves was not significant-
sponds to lack of saturation. The average saturation index was ly different between controls (2.1 0.32) and patients
much higher in controls than in patients (13.45 2.39 versus (1.81 0.34; t21 = 0.61, p = 0.54). The variation in response laten-
2.8 0.63; t21 = 3.91, p = 0.0008). To provide an index of latency cy (after normalization of phase data; see above) was contrast
decrement with increasing contrast, the individual curves dependent (Fig. 4d). As for luminance-contrast stimuli, the VEP
(Fig. 3c and d) were linearly interpolated between 20% and max- latency progressively lagged with decreasing contrast, with a
imum contrast. The average slope of the regression line (ms per somewhat steeper slope for controls than for patients. Latency
unit contrast) was found to be significantly steeper in controls than decrement indices were obtained by linearly interpolating
in patients (19.52 6.18 versus 1.64 5.15; t21 = 2.16; p = 0.04). between 20% and maximum contrast individual curves. The
In a scatter plot of saturation index and latency decrement average slope of the regression line (ms per unit contrast) did
(Fig. 3e), there was little overlap between the two groups, with not significantly differ between controls and patients (27.3 11.4
patients showing less saturation and lower latency decrement than versus 8.0 8.9; t21 = 2.3; p = 0.2).
most controls. Interestingly, the two indices were correlated in con-
trols (R = 0.72, p = 0.008) but not in patients (R = 0.12, p = 0.73). DISCUSSION
The correlation between saturation and latency decrement is con- Intermittent photic stimulation using a stroboscope is exploit-
sistent with proposed models of contrast gain control1416. ed in the EEG laboratory to trigger photoparoxysmal responses in
PSE patients, with 1120 flashes per second representing the most
Chromatic gratings: temporal frequency and contrast
We explored whether steady-state VEPs to chromatic gratings a b
were abnormal in PSE patients. Previous studies report failure
of equiluminant red/green stimuli to induce paroxysmal EEG
activity7. Temporal and contrast functions were evaluated for
equiluminant red/green gratings (see Methods), and the results
were summarized (Fig. 4). The temporal function for high con-
trast chromatic gratings (Fig. 4a) had a simple profile with an
amplitude peak at about 5 Hz and rapid attenuation at higher
frequencies11,17. Both the shape of the temporal function and the
high-frequency cutoff were similar in PSE and controls, differ- c d
ing mainly in amplitude (two-way ANOVA, F 1,197 = 20.9,
Fig. 4. Chromatic-contrast gratings: effect of temporal frequency and

contrast. Average ( s.e.) VEP amplitude and phase as a function of tem-
poral frequency (a, c) and contrast (b, d). The shape of the temporal
and contrast functions (a, b) was comparable in control subjects and
patients. The slopes of the phase and latency plots (c, d) tended to be
steeper in controls than in patients.

articles
effective frequencies2,18. Photoparoxysmal activity may also be gesting that a mechanism normally responsible for contrast gain
provoked by viewing a striped black/white pattern19. Contrast control has been suppressed25. Our VEP data do not allow us to
values above 40%, a spatial frequency of 24 cycles per degree advance suggestions about the site and nature of the gain-control
and 1020 Hz reversal frequency are most epileptogenic19. mechanism1416,26. Note, however, that the pattern ERG in response
Human photosensitivity is a form of reflex epilepsy arising in to both luminance and chromatic gratings does not show signif-
the visual cortex, with high tendency toward generalization19. icant saturation and phase shift with increasing contrast27. The
The efficacy of the triggering stimulus depends on its ability to average saturation index of VEP amplitudes and the average laten-
elicit action potentials in a hyperexcitable visual cortex, on syn- cy decrement with increasing stimulus contrast in controls are in
chronizing effects and on the size of the neural population acti- agreement with data obtained from simple cortical cells under
vated8. similar stimulus conditions14. This encourages us to interpret the
The present findings, obtained with patterned stimuli of high control data as indicating a mechanism of contrast gain control
contrast sinusoidally modulated in time, show that a range of rel- and to speculate that such a mechanism is defective or absent in
atively low temporal frequencies may be critical for cortical hyper- the visual system of PSE patients. VEP responses at the higher
excitability. Indeed, the VEP responses of the patients in this temporal frequencies, at which normal subjects did not show
temporal frequency range tend to be higher in amplitude in com- amplitude saturation, were unaffected in PSE.
parison with controls, in spite of most patients being treated with VEPs to high-contrast red/green equiluminant gratings tend-
antiepileptic drugs, which may reduce the size of the VEPs2,21. ed to be larger than normal and showed a smaller phase advance
To investigate the origin of the different VEP responses of in PSE, although saturation indices were not different from those
patients and controls, we measured the amplitude and phase for controls. The present findings do not contradict the previ-
dependence of VEPs on stimulus contrast at two temporal fre- ous report that stationary equiluminant patterns are ineffective in
quency ranges, 410 Hz and 1622 Hz. Whereas the VEP depen- inducing photoparoxysmal EEG activity7.
dence on contrast is comparable for patients and controls at In conclusion, our results indicate that pattern stimuli of rel-
higher temporal frequency, there is a clear difference at the lower atively low temporal frequency and high contrast may be partic-
temporal frequency. The amplitude saturation at high contrasts ularly effective in uncovering cortical hyperexcitability in PSE,
and the phase advance with increasing contrast typically found possibly because of an impairment of contrast gain-control mech-
in normal subjects22 were absent or much reduced in patients. anisms normally present at these temporal frequencies. High-
Interestingly, the contrast threshold and the contrast dependence contrast stimuli endowed with such temporal characteristics are
in the low contrast range were not affected in PSE. The critical common in TV images and in video games, and may be impor-
range of reversal of square-wave patterns reported to elicit epilep- tant in triggering the abnormal cortical response underlying visu-
tic EEG activity (1020 reversals per second)19corresponds to a ally induced epileptic seizures.
fundamental temporal frequency of 510 Hz, in agreement with
our findings of sine-wave temporal modulation. METHODS
Single cells of the primary visual cortex of the cat and mon- Subjects. Subjects were 12 normal volunteers (mean age, 15.2 3.2 years;
key show response saturation and phase advance in response to n = 6 males) and 11 adolescents and young adults (mean age, 18 3 years;
drifting sinusoidal gratings of increasing contrast14,23. These non- n = 3 males) with PSE9 (Table 1). All patients showed high-amplitude
VEPs (with normal waveform) to both bright flashes and checkerboard
linear response properties could result from a contrast gain con-
patterns21. Nine patients were being treated with antiepileptic drugs dur-
trol stage that scales the input contrast taking into account the ing the study. Antiepileptic drugs (especially VPA) attenuate the pho-
average contrast in the surround16 or the average activity of a large toparoxysmal response and may reduce the VEP amplitude2,21. These
population of surrounding cells14,24. In the cat visual cortex, effects are interpreted as resulting from reduced spread of cortical activ-
removing inhibitory effects by local application of bicuculline ity rather than alteration of mechanisms responsible for its generation28.
abolishes VEP response saturation and reduces phase shift, sug- EEG monitoring throughout the VEP sessions did not show epileptiform
Table 1. Details of the 11 photosensitive patients.

AEDs at VEPs Reported enviromental Photosensitivity range (frequency)
Gender/age Seizure types examination triggers IPS Pattern (reversal rate)
M/19 CPS VPA bright sunlight 1030 Hz 220 Hz
F/20 CPS no TV 1825 Hz NP
F/22 SPS + II gen CBZ TV, computers 1525 Hz 1018 Hz
F/21 SPS + II gen CBZ TV 1518 Hz NP
F/13 SPS VPA TV, bright sunlight 1230 Hz NP
F/21 CPS + II gen VPA, PB TV, bright sunlight, 1540 Hz 218 Hz
video games
M/19 SPS + II gen no video games 1221 Hz 1020 Hz
F/21 CPS VPA, CBZ TV, bright sunlight 515 Hz 520 Hz
F/19 CPS VPA TV 1318 Hz 1018 Hz
F/14 SPS + II gen VPA TV 1521 Hz 1020 Hz
M/15 CPS VPA TV, highly contrasted 450 Hz 418 Hz
patterns
AEDs, antiepileptic drugs; VEPs, visual evoked potentials; Y, years; IPS, intermittent photic stimulation; M, male; CPS, complex partial seizure; VPA, valproic acid;
Hz, hertz; F, female; TV, television; NP, not performed; SPS, simple partial seizure; II gen, secondary generalization; CBZ, carbamazepine; PB, phenobarbital.

articles
activity. Experiments followed the tenets of the Declaration of Helsinki. 3. Sullivan, K. Perplexity and outrage over TV show. 200 remain hospitalized in
Informed consent was obtained after the nature of the technique and the Japan after seizures from viewing cartoon. The Herald International Tribune
aim of the research were explained. December 19, 5 (1997).
4. Ishida, S. et al. Photosensitive seizures provoked while viewing Pocket
Monsters a made-for-television animation program in Japan. Epilepsia 39,
Visual stimuli. Methods for generating luminance contrast (blackwhite) 13401344 (1998).
and chromatic contrast (red-green) stimuli are reported in detail else- 5. Christie, T. Death after an attack precipitated by playing Nintendo game. The
where17,29. Briefly, stimuli were horizontal sinusoidal gratings of 2 cycles Sun January 21, 5 (1993).
6. Commission on Classification and Terminology of the International League
per degree and different Michaelson contrast (390%, in half-octave steps), Against Epilepsy. Epilepsia 30, 389399 (1989).
sinusoidally reversed in contrast at 324 Hz. Both luminance and chro- 7. Wilkins, A. J., Darby, C. E. & Binnie, C. D. Neurophysiological aspects of
matic gratings were generated by a VSG/2 graphic card (Cambridge pattern-sensitive epilepsy. Brain 102, 125 (1979).
Research System, Cambridge, UK) and displayed on a gamma-corrected 8. Binnie, C. D., Findlay, J. & Wilkins, A. J. Mechanisms of epileptogenesis in
color monitor (Barco CCID 7751, Kortrijk, Belgium). To establish equi- photosensitive epilepsy implied by the effects of moving patterns.
Electroenceph. Clin. Neurophysiol. 61, 16 (1985).
luminance, subjects viewed a red/green grating of 50% contrast alternat- 9. Guerrini, R. et al. Idiopathic photosensitive occipital lobe epilepsy
ing at 15 Hz, and adjusted the relative luminance of red and green30 to [published erratum appears in Epilepsia 37, 310, 1996] Epilepsia 36, 9,
minimize perception of flicker. Chromatic patterns were viewed through 88391 (1995).
yellow filters (Kodak Wratten 16, Rochester, New York) to attenuate wave- 10. Guerrini, R., Bonanni, P., Parmeggiani, L. & Belmonte, A. Adolescent onset
lengths lower than 500 nm. CIE coordinates (evaluated by a Minolta Chro- of idiopathic photosensitive occipital epilepsy after remission of benign
rolandic epilepsy. Epilepsia 38, 777781 (1997).
mameter CS 100, Osaka, Japan) were x = 0.637, y = 0.362 (red); x = 0.416, 11. Porciatti, V., Burr, D. C., Morrone, C. & Fiorentini, A. The effects of ageing
y = 0.582 (green). Visual stimuli had a mean luminance of 14 cd per m2 on the pattern electroretinogram and visual evoked potential in humans.
and subtended an area of 14 14 at the viewing distance of 100 cm. Vision Res. 32, 11991209 (1992).
Subjects monocularly fixated a black spot at the center of the screen. 12. Simon, F. The phase of PVEP in Maxwellian view: influence of contrast,
spatial and temporal frequency. Vision Res. 32, 591599 (1992).
Stimuli in a comparable spatiotemporal range provoked abnormal 13. Campbell, F. W. & Maffei, L. Electrophysiological evidence for the existence
EEG activity under diagnostic investigation (Table 1). To prevent trig- of orientation and size detectors in the human visual system. J. Physiol.
gering abnormal EEG activity during the experiment, we limited stim- (Lond.) 207, 635652 (1970).
ulation to one eye and to a relatively small area of the visual field. 14. Carandini, M., Heeger, D. J. & Movshon, J. A. Linearity and normalization in
Moreover, abrupt changes in retinal stimulation were avoided using sine- simple cells of the macaque primary visual cortex. J. Neurosci. 17, 86218644
(1997).
wave contrast reversal and horizontal gratings that reduced the effects of 15. Shapley, R. M. & Victor, J. D. How the contrast gain control modifies the
saccadic eye movements. frequency response of cat retinal ganglion cells. J. Physiol. (Lond.) 318,
161179 (1981).
VEP recording. Steady-state VEPs were recorded from scalp electrodes 16. Geisler, W. S. & Albrecht, D. G. Cortical neurons: isolation of contrast gain
control. Vision Res. 8, 1491410 (1992).
(resistance < 5 k) placed over the occipital (Oz) and frontal (Fz) cor- 17. Fiorentini, A., Porciatti, V., Morrone, M. C. & Burr, D. C. Visual ageing:
tices, and referenced to the vertex (Cz). The left mastoid was grounded. unspecific decline of the responses to luminance and colour. Vision Res. 36,
Signals were filtered (0.3100 Hz, 6 dB per octave) amplified (50,000- 35573566 (1996).
fold), digitized (2 kHz, 12-bit resolution) and averaged (at least 160 18. Newmark, M. E. & Penry, J. K. Photosensitivity and Epilepsy (Raven, New
sums), with rejection of signals exceeding a threshold voltage. This York, 1979).
19. Binnie, C. D. & Wilkins, A. J. in Reflex Epilepsies and Reflex Seizures. Advances
assured that single sweeps occasionally contaminated by epileptiform in Neurology (eds. Zifkin, B., Andermann, F., Beaumanoir, A. & Rowan, J.)
spikes did not contribute to the averaged records. To reduce patients 123138 (Lippicott-Raven, Philadelphia, 1998).
exposure time to the patterns, the stimuli were presented for a short time 20. Binnie, C. D. in Epileptic Seizures and Syndrome (ed. Wolf, P.) 4954 (John
(corresponding to 20 sweeps) and intermingled with brief interruptions. Libbey, London, 1994).
21. Guerrini, R. et al. in Reflex Epilepsies and Reflex Seizures. Advances in
Partial averages (20 sums) of the total averages were used to evaluate
Neurology (eds. Zifkin, B., Andermann, F., Beaumanoir, A. & Rowan, J.)
response consistency11; these were of the same order in patients as in 159178 (Lippincott-Raven, Philadelphia, 1998).
controls. Averaged VEPs were submitted to DFT analysis to evaluate the 22. Spekreijse, H., Van Der Tweel, L. H. & Zuidema, T. Contrast evoked
second-harmonic (major response component) amplitude and phase11. responses in man. Vision Res. 13, 15771601 (1973).
Responses to patterns of zero contrast were frequently recorded to obtain 23. Albrecht, D. G. & Hamilton, D. B. Striate cortex of monkey and cat: contrast
response function. J. Neurophysiol. 48, 217237 (1982).
a measure of residual noise (average, 0.3 V). Responses recorded simul- 24. Carandini, M. & Heeger, D. J. Summation and division by neurons in primate
taneously from Fz were close to the noise level and are not shown. visual cortex. Science 264, 13331336 (1994).
25. Morrone, M. C., Burr, D. C. & Speed, H. D. Cross-orientation inhibition in
cat is GABA mediated. Exp. Brain Res. 67, 635644 (1987).
ACKNOWLEDGEMENTS 26. Bernardete, E. A. & Kaplan, E. The dynamics of primate M retinal ganglion
cells. Vis. Neurosci. 16, 355368 (1999).
The authors would like to thank M. C. Morrone for suggestions and discussion 27. Morrone, M. C., Porciatti, V., Fiorentini, A. & Burr, D. C. Pattern
and C. Orsini for technical help. electroretinogram in response to chromatic stimuli: I Human. Vis. Neurosci.
11, 861871 (1994).
28. Wilkins, A. J., Binnie, C. D., Darby, C. E. & Kasteleijn-Nolst Trenit, D. G. A.
RECEIVED 1 DECEMBER 1999; ACCEPTED 19 JANUARY 2000 in Reflex Seizures and Reflex Epilepsies (eds. Beaumanoir, A., Gastaut, H. &
Naquet, R.) 153162 (Editions Mdicine & Hygine, Geneva, 1989).
29. Porciatti, V., Di Bartolo, E., Nardi, M. & Fiorentini, A. Responses to
1. Kasteleijn-Nolst Trenit, D. G. A. in Reflex Epilepsies and Reflex Seizures. chromatic and luminance contrast in glaucoma: a psychophysical and
Advances in Neurology (eds. Zifkin, B. Andermann, F. Beaumanoir, A. & electrophysiological study. Vision Res. 37, 19751987 (1997).
Rowan, J.) 99121 (Lippincott-Raven, Philadelphia, 1998). 30. Mullen, K. T. The contrast sensitivity of human colour vision to red-green
2. Harding, G. F. A. & Jeavons, P. M. Photosensitive Epilepsy (Mac Keith, and blue-yellow chromatic gratings. J. Physiol.(Lond.) 359, 381400
London, 1994). (1985).

articles
Learning to find a shape

M. Sigman and C. D. Gilbert
The Rockefeller University, 1230 York Avenue, New York, New York 10021-6399, USA
Correspondence should be addressed to C.D.G. (gilbert@rockvax.rockefeller.edu)
We studied the transition of stimuli from novel to familiar in visual search and in the guidance of atten-
tion to a particular object. Ability to identify an object improved dramatically over several days of train-
ing. The learning was specific for the objects position in the visual field, orientation and configuration.
Improvement was initially localized to one or two positions near the fixation spot and then expanded
radially to include the full area of the stimulus array. Characteristics of this learning process may reflect
a shift in the cortical representation of complex features toward earlier stages in the visual pathway.
In a visual search task, a target must be detected within a field of trally positioned fixation spot. A screen, in which the target was
distractors. The target can be defined by various attributes, such either present in a randomized position or absent, was presented
as color, orientation or form. Depending on the combination of every 3 seconds for a duration of 300 milliseconds (Fig. 1a). The
target and distractors, search efficiency may be influenced by the subjects task was to report whether the target was present. We mea-
number of distractors14. sured the percent of correct responses for a fixed presentation time.
According to feature integration theory, the difficulty of visu- To compensate for guessing, 20% of the trials were a null condition
al search is determined by a targets uniqueness in the map of in which no target was present. A separate false-positive rate was
some elementary feature1. Visual input is processed in two stages. calculated for each experiment. This false-positive rate, fp, was used
The first stage uses a set of retinotopically organized maps coding to adjust the percentage of positive responses, p, according to the
for an elementary attribute such as color or orientation. This formula p = (p fp)/(1 fp) to yield p, the true-positive rate, which
stage operates in parallel across visual space, but it produces no we averaged for all subjects performing each experiment. The false-
information about conjunctions of elementary features. Detec- positive rate was below 3% for all subjects. The tests of significance
tion of conjunctions represents a second stage that operates seri- were carried out using a two-tailed t-test over the data collected from
ally to produce the percept of a whole object. all subjects; error bars correspond to standard deviations.
This theory considers a shape to be a conjunction of elemen- In certain types of search tasks, the target attracts attention
tary features or strokes3. Identifying a shape would therefore require even when the observer has no knowledge about its characteris-
serial search, and the ability to identify it should diminish with the tics. For example, even without a previous cue, a red object
number of distractors. However, this degradation of performance embedded in an array of blue distractors will draw the viewers
can be strongly counteracted by familiarity, even when the low- attention6. Our search task, however, required that the viewer
level features of targets and distractors are held constant. For have explicit knowledge of the object sought. Subjects could not
instance, recognizing the digit 2 among an array of the digit 5 perform the task unless they were instructed to find a triangle of
becomes much harder when rotating the entire image by 90 ren- a particular orientation.
ders the characters less familiar5. Experiments were run in blocks of 150 trials with a target of
Under some circumstances, performance in a visual search a single orientation. Before each block, subjects were informed
task can also be improved by priming. The effect of priming is of the orientation of the target. Each session consisted of eight
thought to be limited to simple visual attributes and is passive different blocks. In two sessions before training, performance
and automatic6,7. Here we show that perceptual learning extends levels on detecting triangles of the four different orientations
the range of priming effects and is important in the ability to were tested. Naive subjects showed an average performance below
guide attention to a particular object. 20% for all different orientations.
We chose a visual search task that involved searching for a tri-
angle (a target defined by form) among an array of distractors, Effects of training
in this case triangles of other orientations (Fig. 1a). We show that After having measured performance levels before training, we
perceptual learning dramatically increases the ability to find a chose a single orientation, and the subject was trained by
shape. Moreover, we show specificity of this learning for visuo- repeating blocks in this particular orientation. Different orien-
topic position, object orientation and object configuration. tations were used as targets for different subjects. All subjects
substantially improved performance over the training period.
RESULTS Training stopped when subjects reached threshold, which we
Performance before training arbitrarily set in the range of 7080% correct responses
We used a visual search task in which the observer was required to (Fig. 1b). For different subjects, the time to reach threshold var-
find a target embedded in an array of distractors. The target con- ied between 4 and 6 days, corresponding to 50007000 trials.
sisted of a triangle of one of four possible orientations (up, left, right To measure the change in performance for the trained and
or down), surrounded by triangles of the other three orientations. untrained orientations, we then repeated the two test sessions in
The triangles were presented in a 5 5 stimulus array with a cen- which subjects were tested on triangles of the four orientations.

articles
a Fig. 1. Training on trian-

b gles of a particular orienta-
tion resulted in
improvements in detection
Percent correct
specific to the object at the
trained orientation. (a) A
stimulus consisting of a
5 5 array composed of
triangles of 4 possible ori-
entations (right, left, up or
down) was presented for
300 ms. The target, a trian-
Days trained gle of particular orienta-
tion, was present in 80% of
Trained orientation Untrained orientation
c the cases. The distractors
were triangles in the three
remaining orientations. (b)
One subjects progress
Target (not present)
through the course of
Percent correct
learning. (c) Averaged

responses (four subjects)

for the trained orientation.
Performance improved
fivefold after training. No
change was seen for the
untrained orientations. (d)
Before After Before After
training training training training Improvement (averaged
over two subjects) lasted
d Trained orientation for at least one month
Untrained orientation without practice with no
degradation in perfor-
mance. Two subjects were
Percent correct
trained on a second orien-

tation; subsequently, per-
formance for the first
trained orientation
Target (present) degraded, reflecting a neg-
ative-transfer function in
the orientation domain.
Before After 1 month after After training on a

training training training second orientation
Subjects showed an average 5-fold increase of performance in ty of the training effect by comparing the change in performance
detecting triangles of the trained orientation (p = 15.4 5.3% at specific locations within the array. In particular, we wanted to
before training; p = 74.0 2.9% after training; significance, determine if the learning occurred sequentially in different loca-
p < 106) but no significant increase in detection of triangles of tions of the visual field or if the improvement resulted from a
the untrained orientations (p = 19.5 4.7% before training, globally and uniformly increased ability in all the locations of the
p = 21.3 5.0% after training; significance, p > 0.3, average array. Learning tended to occur sequentially in different loca-
over 4 subjects; Fig. 1c). tions of the visual field, expanding from the fovea to the periph-
Once training for one orientation was completed, we waited ery, and the spatial pattern of performance levels was very similar
one month and repeated the test session to examine retention of for consecutive or nearly consecutive blocks (Fig. 2a). Further-
the improvement. There was no change in the training effect after more, the expansion in the spatial coverage of the learning tend-
the 1-month hiatus (p = 74 2.5% after learning and ed to occur between adjacent sites in the array. This spatial
p = 77 1.9% after hiatus, average over 2 subjects, p > 0.2), indi- correlation was not exclusively a function of eccentricity. That is,
cating that the improvement showed no extinction over time if a subject was more likely to detect a target in one particular
(Fig. 1d). After training on triangles of a second orientation, how- position in the array than in another, he would more easily detect
ever, performance on the initially trained orientation declined, drop- a target in the same and neighboring locations in the subsequent
ping from p = 74.0 2.9% after the first training to p = 57.0 3.5% trials. To quantify this, we measured the Euclidean distance
after subjects were trained for 7 days in a second orientation, aver- between the blocks. Put simply, distance is a quantitative mea-
aged over 2 subjects (significance, p < 0.05; Fig. 1d). sure of the spatial differences in performance level between con-
secutive blocks, D1, or blocks separated by greater intervals (D2,
Spatial dependence of the learning D3,..., Dn). We plotted Euclidean distances between positions as a
The results above were averaged over all spatial locations at which function of block separation under three different conditions,
the stimuli appeared. One can examine the visuotopic specifici- either in the real data, in shuffled data (in which the responses

articles
were scrambled in the a

different positions for
each block of the original
data) or in angularly
scrambled data, in which
positions were shuffled
but all points retained
their initial radial dis-
tance from the fixation
point (Fig. 2b). The
increase in Euclidean
distance as a function of
block separation in the
original data demon-
strated a strong correla-
tion between successive
blocks. The results for
the scrambled data show
that this correlation was

not fully accounted for
by the increase in the
Fig. 2. Learning showed visuotopic speci-
mean rate of correct ficity. It progressed serially from fovea to
Scrambled data
responses, because the periphery; positions showing improve-
scrambling the respons- Angularly scrambled data ment were correlated from trial to trial.
es through different b (a) Percent of correct responses, for one
positions (keeping the Real data subject, as a function of position for differ-
total rate of responses ent blocks during the learning period. Each
constant through blocks) square corresponds to one block of 150
increased the distance trials; within each square, the gray-scale
value of each circle represents the perfor-
between neighboring tri-
Distance
mance level for a particular location within

als considerably. The the 5 5 array. The fixation spot in the
results for the angularly center of the array is indicated with a small
scrambled data showed black circle. (b) The learning showed spa-
that the correlation was tial specificity. Average distance between
not simply radial corre- blocks for 4 subjects (see Methods) were
lation due to the pro- significantly smaller for measured data than
gression of learning from for scrambled data in 2 different condi-
Block separation (number of blocks)
fovea to periphery; tions, either with all 24 positions scram-
rather, they demonstrat- bled or with only those positions
equidistant from the fovea scrambled.
ed correlation of precise
locations throughout the
course of learning.
It is important to remember that the target was presented Form specificity of the learning
with equal probability at each location within the trained array. The last series of experiments were designed to test whether our
Thus, the observed visuotopic specificity was not due to train- search task involved solely what are considered low-level mecha-
ing on particular locations within the array. To exclude the pos- nisms, such as orientation discrimination or texture segmentation.
sibility that the improvement might result merely from an We tested subjects who had been trained on our search task with
increase in speed in deciding whether a particular shape two novel stimulus configurations. In the first configuration, trian-
matched the target, regardless of its position, we tested the gles were replaced by arrowheads, which were still clearly recogniz-
trained subjects in a condition in which both target and a variable as pointing left, right, up or down, but which did not have the
able number of distractors were presented outside the area of property of closure (Fig. 4b). The learning effect was measured as
the training array. We then compared performance levels for the ratio between performance levels using the trained and untrained
various numbers of distractors outside the training array with orientations. The orientation specificity in the levels of performance
that for the same number of distractors presented along with for closed triangles did not transfer to the arrowheads. The means for
targets within the training array. Within the area of the train- arrowheads were p = 36.2 8.0% in the trained orientation and
ing array, performance levels did not change with the number of p = 35.0 5.5% in the untrained orientation (averaged over 3 sub-
distractors when the target was at the trained orientation, but jects; significance, p > 0.5; Fig. 4d). This shows that subjects learned
did change when it was at an untrained orientation. Outside of not to discriminate orientation but actually to find an object at a
the training array, levels of performance for both trained and particular orientation. In the second configuration, the target was
untrained orientations decreased with increasing number of not changed, but the field of distractors was completely novel. This
distractors (Fig. 3). This shows that the improvement was spe- was done to determine whether the learning was specific for the tar-
cific for a particular shape and for a particular region of the get itself or for a more generalized textural difference between fore-
visual field on which subjects were trained. ground and background. The new figures used as distractors did

articles
Fig. 3. Performance as a function of

number of distractors within and out-
a b side the training region. After training in
the 5 5 array (gray square), perfor-
mance was tested outside the training
region. The target and either 7 (a) or 23
(b) distractors were presented at
eccentricities ranging from 3.5 to 4.7.
(c) Performance within the training
region was tested for a 3 3 array (7
distractors) and a 5 5 array (25 dis-
tractors), and was averaged over 3 sub-
jects. Within the training region,
performance for the untrained orienta-
tion, but not for the trained orientation,
declined with number of distractors.
Outside the training region, perfor-
mance for both trained and untrained
orientations declined with number of
c distractors.
Trained orientation
Untrained orientation the failure of training effects to trans-
Trained region fer to arrowheads suggests that this
Percent correct
task was not a simple orientation-

Untrained region
discrimination task.
Interestingly, after training, all
subjects claimed that there was no
conscious perceptual distinction
between different triangles, even
though they performed consider-
ably better for target triangles of one
orientation than for other orienta-
tions. The effect of learning on this
Number of distractors
task represents an extension of find-
ings on the involuntary nature of
not include the triangles in the other three orientations (Fig. 4c) and priming, which is thought to be limited to simple visual (ele-
were presented at different contrasts to increase distractor variabil- mentary) attributes as opposed to form6. The degradation of per-
3
ity, making the task more difficult . We observed that the specificity formance on figures of new orientations after learning one
of the training extended to this new background: for the trained ori- orientation further extends the analogies with priming effects for
entation, p = 68.1 11.0%, compared with p = 37.9 11% for position, which shows distractor inhibition7, and might result
untrained orientation (averaged over 3 subjects; significance, from a difficulty in ignoring targets whose processing has been
p < 105; Fig. 4d). automated as a consequence of perceptual learning1113.
Learning in this task was not just a consequence of perceptu-
DISCUSSION al exposure, but must have involved top-down influences1321.
We studied the effects of training on a search task in which the This is clear because learning occurred only for the target orien-
target was defined by form. In this task, search efficiency was sig- tation, even though the subject was exposed seven times more
nificantly increased as a consequence of learning. This learning often to triangles in each of the untrained orientations during
was object specific and resulted from a progressive acquisition of the whole course of learning. Identification of form demands
the ability to identify the given object in different locations in the attention2226, as is suggested by the decrease in performance with
visual field. The results suggest that learning in visual search can increases in number of distractors. This implies not only that
be targeted to a specific object. Although it is suggested that learn- attention is required to obtain learning, but that, conversely,
ing in visual search involves a general improvement in perform- learning is required to rapidly direct the attentional mechanism
ing searches8, other studies show orientation dependence of toward a particular object.
learning pop-out detection9. Visual search tasks are usually classified as parallel or serial
The task used in our experiments did not involve texture seg- based on whether the performance depends on the number of
mentation or orientation discrimination, but identification of an distractors1, though it is suggested that this classification repre-
oriented object. This is supported by three observations. First, the sents not a real dichotomy, but two extreme cases of a continu-
subjects could not perform the task if they did not have previous um3,27,28. Here we show that increasing the number of distractors
knowledge of the target characteristics. Triangles of a particular ori- within the training region did not change search efficiency for
entation embedded in triangles of other orientations could not be the trained orientation. However, search efficiency diminished
detected as unique objects. Second, we showed that learning in this when distractors were added outside of the training region. This
task was specific for the target and transferred to different back- showed that the dependence of search efficiency on number of
grounds. In contrast, learning effects in texture discrimination are distractors may be a function of distractor position as a conse-
specific for the field of distractors but not for the target10. Third, quence of perpetual learning. Another characteristic of learning

articles
Fig. 4. Learning was specific

a Trained b Arrowheads c New background for object configuration and
transferred to a new back-
ground. (a) The training array.
(b) Test using arrowheads
as target and distractors.
(c) Target used in training (tri-
angle) in a new field of distrac-
tors. For the new background,
we used open figures, circles,
squares, diamonds and semicir-
cles as distractors (c). The tar-
get was presented at the same
luminance (60 cd per m2) used
Target (present) Target (present) Target (present) in the other experiments.
Distractors were randomly
assigned a contrast of 3391 cd
per m2. (d) Performance was
averaged over three subjects
for targets and for arrowheads
in trained and untrained orien-

Trained Arrowheads New background
tations, which lacked the fea-
d ture of closure but were still
oriented figures. Performance
did not differ for arrowheads in
trained and untrained orienta-
tions. Performance in the new
background (average over three
subjects) was better for the
Percent correct
trained orientation, demon-

strating object specificity of the
learning.
sentations of the trained

object may be built repeated-
ly for different positions
across the cortical area. Even
was that the inherent dependence of performance on position within V1, cells are selective for much more complex stimulus
within the visual field was greatly reduced with training, as for configurations than originally believed3134, suggesting a role for
the detection of oriented gratings9. V1 in the identification of complex forms. Connections within
In our experiments, the target could appear at any position V1 are plastic35,36, and modification of these connections may
within the array. Therefore, the spatial dependence of the learn- contribute to the plasticity of elementary-feature maps. Repre-
ing observed within the training region was not due to the loca- sentation of more complex features at earlier levels may enhance
tion of the stimulus, as is the case when localized improvement efficiency and rapidity in recognizing these features in a complex
results from practice in a fixed position of the visual field10,19,29,30. background at the expense of requiring multiple shape repre-
This specificity results, therefore, from intrinsic mechanisms that sentations in areas showing smaller receptive fields and greater
may reflect the sequence of sites targeted by the search strategy. It visuotopic order.
could be argued that the improvement resulted from an increase
in the speed with which the subjects, independent of visuotopic
position, could determine whether a given item was the target.
METHODS
Psychophysical experiments on human observers (male and female, 2327
This, combined with a search strategy in which subjects started years of age) were designed to study the effects of learning in a visual
scanning close to the fovea and proceeded to the periphery, could search task. All subjects gave written informed consent in accordance with
account for the spatial specificity we found within the training procedures and protocols approved by the Rockefeller University Institu-
region. If this were the case, a subject trained for left triangles tional Review Board. Stimuli were presented on a NEC monitor 5FGp
should perform better for left than for right triangles outside of refreshed at a rate of 60 Hz, and were observed a distance of 150 cm with
the area used for training. However, we showed that when target both eyes, with normal pupil apertures and without head restraint.
and distractors were presented outside of the training region, Each trial consisted of a 3000-ms cycle. A 5 5 array consisting of a
detectability was no better for the trained orientation. Poorer central fixation spot and 24 shapes in the remaining locations was pre-
sented for 300 ms; a response was recorded during the subsequent 270-
performance for targets of the trained orientation presented out-
ms interstimulus interval. As an auditory cue to alert the observer, a short
side the training region, therefore, showed that the improvement beep was sounded at the onset of the visual stimulus.
was localized to a particular region of the visual field. The psychophysical experiments investigated the observers ability to
Based on its spatial specificity, one may speculate that early identify a target triangle among an array of distractors. Both target and
cortical processing might be involved in this process. The pro- distractors were presented at high contrast (60 cd per m2) against a uni-
gression of learning across the visual field suggests that repre- form background (2 cd per m2). The target randomly appeared in any

articles
location within the array. After each presentation, the subject indicated 10. Karni, A. & Sagi, D. Where practice makes perfect in texture discrimination:
whether or not a target shape was present by pressing the appropriate evidence for primary visual cortex plasticity. Proc. Natl. Acad. Sci. USA 88,
button of a computer mouse. 49664970 (1991).
11. Schneider, W. & Shiffrin, R. M. Controlled and automatic human
The array subtended 4.2 4.2, and a small fixation spot of one information processing: I. detection, search and attention. Psychol. Rev. 84,
arcmin radius (1) was positioned in its center. Figures used as target or 166 (1977).
distractors in the different experiments were equilateral triangles, squares, 12. Shiffrin, R. M. & Schneider, W. Controlled and automatic human
diamonds or arrowheads. The sides of all shapes were 27 in length and information processing: II. perceptual learning, automatic attending and a
general theory. Psychol. Rev. 84, 127191 (1977).
their centers were separated by 54. 13. Treisman, A., Verira, A. & Hayes, A. Automaticity and preattentive
Average distance as a function of block separation Dn, was calculated as processing. Annu. Rev. Neurosci. 105, 341362 (1992).
follows. From each block we calculated a 5 5 matrix (Ma, is the corre- 14. Ahissar, M. & Hochstein, S. Learning pop-out detection: specificities to
sponding matrix to block a) where each position of the matrix (mai,j) is stimulus characteristics. Vision Res. 36, 34873500 (1996).
defined as the level of performance in the corresponding location of the visu- 15. Braun, J. Vision and attention: the role of training. Nature 393, 424425
(1998).
al field in this block (Fig. 2a). We then have an ordered array of matrixes, 16. Ahissar, M. & Hochstein, S. Attentional control of early perceptual learning.
and we can consider the distances between any two of those matrixes. Proc. Natl. Acad. Sci. USA 90, 57185722 (1993).
17. Ito, M., Westheimer, G. & Gilbert, C. D. Attention and perceptual learning

5 5 modulate contextual influences on visual perception. Neuron 20, 11911197
d(Ma, Mb) = (m
i=1 j=1
a
i,j mbi,j)2 (1998).
18. Fahle, M. & Morgan, M. No transfer of perceptual learning between similar
stimuli in the same retinal position. Curr. Biol. 6, 292297 (1996).
We then define the average distance as block separation to be 19. Crist, R. E, Kapadia, M., Westheimer, G. & Gilbert, C. D. Perceptual learning
of spatial localization: specificity for orientation, position and context. J.
Neurophysiol. 78, 28892894 (1997).

Dn = d(Mi, Mi+n)i 20. Shiu, L. P. & Pashler, H. Improvement in line orientation discrimination is
retinally local but dependent on cognitive set. Percept. Psychophys. 52,
With the exception of one subject (author M.S.), all subjects were naive 582588 (1992).
and were told only what the target was for each block. A session for one 21. Ahissar, M. & Hochstein, S. Task difficulty and the specificity of perceptual
day consisted of eight blocks, each comprising 150 trials. All results cor- learning. Nature 387, 401406 (1997).
respond to the average values for all subjects performing each experi- 22. Bravo, M. J. & Nakayama, K. The role of attention in different visual search
tasks. Percept. Psychophys. 51, 465472 (1992).
ment, and two-tailed t-tests over the data collected from all subjects were 23. Wolfe, J. M. in Current Directions in Psychological Sciences 124128
used as tests of significance. All errors plotted correspond to standard (Cambridge Univ. Press, Cambridge, 1992).
deviations. Individual tests of significance gave comparable results. 24. Wolfe, J. M., Cave, K. R. & Franzels, S. R. Guided Search: an alternative to the
feature integration model of visual search. J. Exp. Psychol. Hum. Percept.
Perform. 15, 419433 (1989).
25. Joseph, J. S., Chun, M. M. & Nakayama, K. Attentional requirements in a
ACKNOWLEDGEMENTS preattentive feature search task. Nature 387, 805807 (1997).
We thank R. Crist for discussions and comments on the manuscript. This work was 26. Chun, M. M. & Jiang, Y. Contextual cueing: implicit learning and memory of
supported by NIH grant EY07968 and a Burroughs Wellcome fellowship to M.S. visual context guides spatial attention. Cognit. Psychol. 36, 2871 (1998).
27. Braun, J. & Sagi, D. Vision outside the focus of attention. Percept. Psychophys.
48, 4558 (1990).
RECEIVED 7 JULY 1999; ACCEPTED 4 JANUARY 2000 28. Nakayama, K. & Joseph, J. S. in The Attentive Brain (ed. Parasuraman, R.)
279298 (MIT Press, Cambridge, Massachusetts 1997).
29. Fiorentini, A. & Berardi, N. Learning in grating waveform discrimination:
1. Treisman, A & Gelade, G. A feature integration theory of attention. Cognit. Specificity for orientation and spatial frequency. Vision Res. 21, 11491158
Psychol. 12, 97136 (1980). (1981).
2. Sagi, D. & Julesz, B. Where and what in vision. Science 228, 12171219 30. Nazir, T. A. & ORegan, J. K. Some results on translation invariances in the
(1985). human visual system. Spat. Vis. 5, 81100 (1990).
3. Duncan, J. & Humphrey, G. W. Visual search and stimulus similarity. Psychol. 31. Kapadia, M. K., Ito, M., Gilbert, C. D. & Westheimer, G. Improvements in
Rev. 96, 433458 (1989). visual sensitivity by changes in local context: Parallel studies in human
4. Rubinstein, B. S. & Sagi, D. Spatial variability as a limiting factor in texture- observers and in V1 of alert monkeys. Neuron 15, 843856 (1995).
discrimination tasks: implications for performance asymmetries. J. Opt. Soc. 32. Posner, M. I. & Gilbert, C. D. Attention and primary visual cortex. Proc. Natl.
Am. A 7, 16321643 (1990). Acad. Sci. USA 96, 25852587 (1999).
5. Wang, Q., Cavanagh, P. & Green, M. Familiarity and pop-out in visual search. 33. Sillito, A. M., Grieve, K. L., Jones, H. E., Cudeiro, J. & Davis, J. Visual cortical
Percept. Psychophys. 56, 495500 (1994). mechanisms detecting focal orientation discontinuities. Nature 378, 492496
6. Maljovic, V. & Nakayama, K. Priming of pop-out detection: role of features. (1995).
Mem. Cognit. 22, 657672 (1994). 34. Das, A. & Gilbert, C. D. Topography of contextual modulations mediated by
7. Maljovic, V. & Nakayama, K. Priming of pop-out: II. role of position. Percept. short-range interactions in primary visual cortex. Nature 399, 655661
Psychophys. 58, 977991 (1996). (1999).
8. Sireteanu, R. & Rettenbach, R. Perceptual learning in visual search: fast, 35. Darian-Smith, C. & Gilbert, C. D. Axonal sprouting accompanies functional
enduring but non-specific. Vision Res. 35, 20372043 (1995). reorganization in adult cat striate cortex. Nature 368, 737740 (1994).
9. Efron, R. & Yund, E. W. Guided search: the effects of learning. Brain Cogn. 31, 36. Gilbert, C. D., Das, A., Ito, M., Kapadia, M. & Westheimer, G. Spatial integration
369386 (1996). and cortical dynamics. Proc. Natl. Acad. Sci. USA 93, 615622 (1996).

articles
Seeing multiple directions of

motionphysiology and
psychophysics
Stefan Treue, Karel Hol and Hans-Jrgen Rauber
Cognitive Neuroscience Laboratory, Dept. of Neurology, University of Tbingen, Auf der Morgenstelle 15, 72076 Tbingen, Germany
Correspondence should be addressed to S.T. (treue@uni-tuebingen.de)
Dot patterns sliding transparently across one another are normally perceived as independently mov-
ing surfaces. Recordings from direction-selective neurons in area MT of the macaque suggested that
this perceptual segregation did not depend on the presence of two peaks in the population activity.
Rather, the visual system seemed to use overall shape of the population response to determine the
number and directions of motion components. This approach explained a number of perceptual
phenomena, including susceptibility of the motion system to direction metamers, motion patterns
combining three or five directions incorrectly perceived by subjects as comprising only two
directions. Our findings offer insights into the coding of multi-valued sensory signals and provide
constraints for biologically based computational models.
Analysis of visual motion is an important aspect of processing fore use the terms tuning curve (which describes a single neu-
visual information. It is therefore not surprising that primates rons response to many directions of motion) and population
have cells and cortical areas specialized for visual motion pro- response (which describes the response of a population of neu-
cessing. The specialized neurons are direction selective: each rons tuned to different directions of motion to one stimulus)
responds only to a particular subset of directions and speeds interchangeably.
of motion within its receptive field. Cortical areas along the The bell-shaped distribution of population activity suggests
dorsal pathway of the primate visual system are specialized for that the visual system might recover the direction of a stimulus by
the analysis of visual motion; most notably, these include area determining the most active neurons. Such an approach of taking
MT, which is dominated by direction-selective cells, and area the preferred direction of the most active of a population of direc-
MST, which also contains cells encoding more complex tion-tuned neurons as the perceived direction may explain the
motions like rotation or expansion1. MT and MST are arguably ability of the visual system to extract the overall direction of
the most intensively studied extrastriate cortical areas. The motion even for displays combining more than one direction5,6.
curve facing out and left in Fig. 1a is an idealized tuning curve Even for these displays, activity is greatest for neurons whose pre-
of an MT cell to a random dot pattern moving in a single direc- ferred directions match the center of the distribution of stimulus
tion. The response distribution is bell shaped and is well fit by directions.
a Gaussian curve. As is characteristic of direction-selective neu- Not all visual motion displays create a perception of unidi-
rons throughout the visual system, responses are broadly tuned, rectional motion, though. Combination of two patterns sliding
with a tuning width of approximately 90, on average 24 . past each other often creates the impression of two separate sur-
Although neurons differ in their preferred directions, their tun- faces, each moving in a distinct direction. If the percept of a dis-
ing curves are otherwise very similar. If typical tuning curves tinct direction is based on the existence of a peak in the
are combined and plotted as a function of direction preference, population activity, then the perception of transparent motion
we obtain a three-dimensional graph of population responses implies the presence of two peaks. This is either assumed in or a
as a function of both stimulus direction and preferred direc- feature of several models of motion perception79 and cortical
tions of the neurons (Fig. 1a). Experimentally, this three- motion processing10,11.
dimensional plot of activity can be derived by determining If we hypothesize that the response to a transparent stimulus
responses of individual cells to movement in different direc- is simply the scaled sum of the activities evoked by the individual
tions (contour lines in Fig. 1a). The brain, however, must cal- components12,13, we can generate a hypothetical curve for pop-
culate this function from information derived from the ulation activity in response to movement in two directions at a
opposite scenario: a single stimulus simultaneously activates a large angle (Fig. 1c). (Several studies suggest that the response
population of neurons, each preferring a different direction. to two motions is the average of the individual responses, giving
We can predict this population response to a single stimulus a scaling factor of 0.5. Note, however, that the exact value is not
from the three-dimensional plot by taking neural responses along critical, as multiplication by any factor gives the same shape for
a line parallel to the axis denoting direction preference (the curve the population activity as the unscaled sum of the two individual
facing right in Fig. 1a). This population response has the same responses.) However, if the two directions of motion form such
shape as the tuning curve for an individual cell. We will there- a small angle that the two peaks merge (Fig. 1d), the percept

articles
Fig. 1. Direction-tuning curves and

b hypothetical population responses to
transparent patterns. (a) A series of
hypothetical tuning curves plotted
three-dimensionally as a function of
stimulus direction (x axis) and the pre-
ferred direction of a cell (y axis). The z
a axis plots responses normalized to the
Hypothetical population maximal response (response to the pre-
response to a single direction ferred direction). The curve facing out
and left is the response curve of a neu-
c ron tuned to 0, and the curve facing to
the right shows the population
Response
response to a stimulus moving at 180.

(b) A hypothetical population response
to a single direction of motion. (c) A
hypothetical response to movement of
two transparent stimuli at an angle large
Hypothetical population response to two
widely spaced stimulus directions
enough to produce a double-peaked
population response. The dashed lines
d indicate the population responses to

each of the two directions. (d)
Cells preferred Response as in (c) but with an angle
Stimulus direction small enough to merge the two curves
direction into a single-peaked population
response.
Hypothetical population response to two

closely spaced stimulus directions
should change from transparent motion to motion in a single RESULTS

direction intermediate to the two actual component directions. We recorded the responses of 152 direction-selective cells in area
Perceptually, the transition from perceiving two separate direc- MT of 3 macaques to moving random dot patterns while the
tions to only one direction occurs when the two superimposed monkeys performed a fixation task. The stimuli moved coher-
transparent random dot patterns move at a relative angle of about ently within a stationary virtual aperture either in one direction
10 or less14. Although several physiological studies examine or in two directions separated by 30, 60, 90 or 120; the 5
responses of direction-selective neurons to transparent patterns, resulting activity profiles are shown for a narrowly tuned exam-
little is known about the population activities evoked by patterns ple cell (Fig. 2a). The top panel plots the responses for single
combining similar directions, as almost all of these experiments motion. The line represents the best fitting Gaussian (tuning
involve only a small subset of directions or directions at large rel- width, 58). Given the tuning width of this neuron, the scaled
ative angles. sum of the responses to the individual components of the trans-
The hypothesis that the profile of responses to motion in mul- parent motion predicts a single-peaked activity profile for the
tiple directions is the scaled sum of the responses to the individ- 30 pattern and two peaks for the 90 and 120 stimuli (Fig. 2a,
ual components makes a clear prediction: the sum of two lower panels). As predicted, two peaks were generated only for
Gaussians shows distinct peaks if the distance between the centers stimuli separated by 90 or 120. To test our prediction for our
of each distribution exceeds their width (which for direction- entire sample of neurons, we aligned every activity profile to the
tuned neurons averages about 90). Conversely, transparent cells preferred direction, normalized its height to the response
motion at any acute angle should yield an activity profile with a to that direction and then averaged all profiles (Fig. 2b). The tun-
single peak. This means that either the response profile to two ing width of the averaged responses to the single-direction stim-
directions is a highly non-linear combination of the individual ulus was 96; therefore, as predicted, the activity profiles did not
directions, resulting in two peaks even for small directional sep- show two peaks for stimuli at acute angles, remaining essential-
arations between the components, or that a percept of transpar- ly flat for the 90 stimulus. Note, though, that because tuning
ent motion does not depend on the presence of two peaks in the width varied somewhat between neurons, we were concerned
population activity. that deviations from our prediction might be occluded by aver-
Here we show that the latter seems to be the case. Perception aging across neurons irrespective of their tuning width. There-
of multidirectional motion does not require distinct peaks in the fore, we recoded the data to test our prediction more directly.
population activity. Rather, the visual system seems to be able to For every cell, responses were expressed as a function of the neu-
interpret the overall shape of the activity profile. This allows the rons tuning width to single directions. Similarly, the angles
visual system to segregate motions differing even by only a small between transparent stimuli were also expressed as fractions of
angle. However, we also show that summing the responses to the the width of a cells tuning curve to a single direction. Averaging
individual motion components makes the motion system sus- across all cells generated a contour plot (Fig. 2c). The x axis plots
ceptible to direction metamers, stimuli that are perceptually indis- the average stimulus direction as a proportion of the tuning
tinguishable even though they contain different direction width, the y axis gives the normalized angle formed by the two
components. directions in the stimulus, and the contour surface represents the

articles
Single cell Aligned, normalized and

pooled responses
Single-
Single-direction
stimulus
stimulus
single-direction
30 transparent
stimulus
30 transparent
Angle between component directions, divided by tuning widths

stimulus
Response (spikes per s)
Normalized response
60 transparent
60 transparent
stimulus
stimulus
90 transparent 90 transparent
stimulus stimulus
120 transparent 120 transparent

stimulus stimulus
Direction of stimulus from preferred direction,

Stimulus direction Stimulus direction divided by tuning width
(degrees) (degrees away from preferred direction)
Fig. 2. Population activity profiles for transparent motion of various angles. (a) Response profiles for an example cell to the single-direction stimulus
(top curve) and the bidirectional stimuli at 30, 60, 90 and 120 (bottom 4 curves). The x axis plots the stimulus direction relative to the cells pre-
ferred direction. For the transparent motion patterns, the stimulus direction plotted is the average of the two direction components. (b) Response
profiles averaged across all cells after normalizing the firing rates to the response to the preferred single direction (determined by fitting with a
Gaussian curve). (c) Contour plot of the responses across all cells. The data from each neuron are normalized to the highest response in each
response profile. The resulting data points form a horizontal line in the graph at the y position defined by the angle between the two directions pre-
sent in the stimulus normalized by the width of the neurons tuning curve for single directions. The x axis similarly expresses the angle of the stimulus
direction relative to the neurons single-direction tuning width. The four horizontal dashed lines indicate data for the example cell from the left col-
umn. From top to bottom, they show data points for the single direction, 30, 60 and 90 stimulus conditions on the contour plot. Because of the
narrow tuning of the example cell, the data from the 120 stimulus condition fell below the scale. The data are in agreement with our prediction that
the population activity splits into two peaks when the stimulus angle exceeds the bandwidth (when y > 1).
normalized firing rate. The graph forms an inverted Y, clearly at the tuning width of 90 or smaller. Nevertheless, the visual
showing that for ratios less than one (when the angle between system can recover the directions underlying transparent motion
stimuli is smaller than the neurons tuning width), the response with much smaller directional separation. How is this remark-
profile is single-peaked, whereas there are two peaks for larger able ability achieved? Note that, although the population activi-
angles. ty was not multi-peaked, it nevertheless contained information
This demonstrates that the simple scaled sum of the respons- about the underlying directions. Single-lobed activity profiles
es to individual components accounts for responses to multidi- generated by bidirectional stimuli are notably broader than tun-
rectional stimuli very well. Thus, the response profile was not ing curves for a single direction (Figs. 1d and 2). This change in
multi-peaked when the angle between component directions was overall shape of the activity profile might be used by the visual

articles
age direction-tuning curve) that must

be summed to produce the observed
activity profile. For motion in a sin-
gle direction, this would be a single
Gaussian positioned at the actual
direction. When movement in two
directions is presented, the resulting
population activity is too broad to be
accounted for by a single Gaussian,
but can be described with two Gaus-
sians. This approach would yield the
correct percept even when the com-
ponent directions form an acute angle
and, therefore, result in a single-
peaked activity profile (Fig. 1d). Such
an approach can be modeled formal-
ly (see Discussion).
Although this approach is nor-
Fig. 3. Psychophysical stimuli. We created two stimuli with three (b1) and five (b2) directions that
should be directional metamers to the 40 bidirectional stimulus (b). The three and five directions mally a robust method of recovering
and the corresponding dot densities were chosen such that the sum of the individual curves (shown the underlying direction(s), it should
along the right) was virtually identical to the activity evoked by the 40 bidirectional stimulus. If the fail to distinguish between different
population activity was indeed the one proposed here and the only information available to the stimuli when they evoke the same
motion system, these three stimuli (b, b1, b2) should appear to contain the same direction compo- activity profile. Such metamers, stim-
nents. In addition to the potentially metameric bidirectional stimulus (b), we created two other bidi- uli that are perceptually indistin-
rectional stimuli (a and c). These cause population activities different from the three potential guishable despite actual differences,
metamers (as shown in the left graph) and should therefore be perceptually distinguishable from the can be used to psychophysically esti-
other stimuli. Note that the directions in the 50 bidirectional stimulus (c) are present as the
mate direction-tuning bandwidth16.
steepest directional components in stimuli b1 and b2. Because the subjects were asked to judge the
most upward component in all stimuli, stimuli b1 and b2 should result in the same responses as the If the visual system indeed recovers
50 stimulus if the motion system correctly encoded the individual motion components. Tuning direction components of transparent
width for the response profiles used here was assumed to be 90, although note that the width of motion by assessing the width or
the summed activity for the three- and five-direction stimuli (b1 and b2) remains close to the activ- overall shape of the population
ity width evoked by the two-directional stimulus (b) for a large range of tuning-curve widths. response, it should perceive only two
Further material and demonstrations of our stimuli can be viewed at directions in patterns containing three
http://neurosci.nature.com/web_specials/. or five directions, as long as the result-
ing profile of the population activity
matches that expected for two direc-
tions of motion (Fig. 3).
system to identify direction(s) evoking the activity. Component In a series of psychophysical experiments, human subjects were
directions of a bidirectional stimulus can even be approximated asked to report the number of directions perceived in these pat-
(to within 10) by simply subtracting a cells tuning width for terns under eccentric fixation. In more than 94% of the trials with
a single-direction stimulus from the width of the response profile stimuli containing 3 directions, subjects reported perceiving two
to the given bidirectional stimulus. directions. Even the stimulus with 5 directions was reported to
Note that such an approach is different from vector averag- contain just 2 directions in more than 73% of the trials. We pro-
ing, which recovers a direction by calculating the vector sum of ceeded to perform a more direct test of our hypothesis. We pre-
the preferred directions of all activated neurons, weighted by the sented the 3 control stimuli (bidirectional stimuli with
respective activities of those neurons. Such an approach would components at 30, 40 or 50) together with a stimulus com-
yield an intermediate direction whenever two directions are comprising 3 directions that contained components moving at 0 and
bined. In contrast, a winner-take-all approach predicts perceived 50 (b1 in Fig. 3). If the visual system indeed finds the best-fitting
direction as the preferred direction of the most active neurons15. sum of two Gaussians for this stimulus, subjects should perceive
Note that, again, transparently moving patterns would be per- two directions of motion moving in the same directions as in the
ceived as having a single direction of motion, and, as our data 40 stimulus rather than the 50 components actually present.
show, this direction would be equivalent to that predicted by the Subjects could reliably discriminate between stimuli with 2 direc-
vector-averaging model if the two actual directions of motion tions at 30, 40 and 50, but the 2 direction components per-
formed an acute angle. At larger angles, the population activity ceived in the stimulus containing 3 directions were indeed the
shows two peaks, and the winner-take-all model would pick one same as for the 40 stimulus with only 2 directions (Fig. 4, left).
of the two peaks. Clearly, this cannot account for the perception We wondered if this inability of the visual system to exclude
of two directions of motion when they form either a large angle the influence of movement in the horizontal direction could be
or a perceptually separable acute angle (larger than 10). overcome if we supplied additional segregation cues. We colored
We suggest that the visual system uses another approach to upward- and downward-moving dots red and the horizontally
recover the direction(s) underlying a given population activity moving ones green. Subjects were asked to judge the direction of
across direction-selective neurons. The perceived direction(s) the red, upward-moving component. Subjects still judged the
represent the position(s) of the smallest number of Gaussian- perceived upward motion in the stimulus comprising 3 direc-
shaped activity profiles (with widths equal to that for the aver- tions to be the same as the one in the 40 stimulus (Fig. 4, cen-

articles
ter). Interestingly, this was not the case

when the horizontal motion was put on a
different stereoscopic disparity plane (Fig. Metameric Metameric Non-metameric
4, right). With this stimulus manipulation,
Perceived directions
subjects veridically perceived the upward
component, and the responses for the
stimulus with 3 directions did not statisti-
cally differ from those for the 50 stimu-
lus.
DISCUSSION
In summary, we show that the response of
direction-selective neurons in MT to trans- Color Disparity
parent motion was well approximated by
the scaled sum of their responses to the Fig. 4. Psychophysical results. The filled bars show the upward directional components reported
individual motion components. Most for the 30 (bar a), 40 (bar b) and 50 (bar c) stimuli. Error bars indicate standard errors. The
white bar indicates the reported upward direction in the three-direction stimulus (bar b ). Left, in
importantly in the context of this study, the standard condition, the difference between the two middle bars was not significant:1 subjects
population activity profiles were single- reported the same upward component in the 40 stimulus as in the 3-direction stimulus (which
peaked for transparent-motion compo- contains a 50 upward component). Center, when oblique pattern components consisting of red
nents moving at acute angles. Our ability dots and the horizontally moving components consisting of green dots were used, the difference
to perceive multidirectional motion is, between the two middle bars was not significant. Right, here the oblique pattern components were
therefore, not dependent on multiple presented with crossed and the horizontal component with uncrossed disparity. In contrast with
peaks in activity profiles across MT neu- the other 2 conditions, subjects reported the same upward component in the 3-direction stimulus
rons; rather, it seems to be based on an as in the 50 stimulus.
analysis of the overall shape of the popu-
lation response, effectively recovering the
Gaussian-shaped profiles of responses to
the individual components underlying the population activity. parent motion patterns, which is strongest for angles of 3040.
This is supported by our demonstration of direction metamers, The population responses to such patterns are somewhat wider
showing an inability of the visual system to recover motion com- than expected from our simple model, but more data and a more
ponents if identical activity profiles could be created by alterna- formal analysis are required to determine whether this could
tive stimuli with fewer directional components. Our finding that account for the perceptual errors observed in motion repulsion.
stereoscopic disparity information but not color information The demonstration of direction metamers also suggests that
helps the visual system to disambiguate these stimuli supports independent access to the direction components in transparent
the proposal that perception of direction in our stimuli involves motion is lost (in the absence of additional cues such as stereo-
signals from area MT, which contains both disparity-tuned and scopic disparity or speed30), even though the front end of the
direction-tuned neurons 1719 . Such neurons would allow segre- motion system seems to require local imbalance of motion sig-
gation of direction components if they lay at different depths, nals for perception of transparency31,32. Stereoscopic disparity,
effectively creating independent population activities for the dif- but not color differences, also help to create perceptual trans-
ferent planes. Some color information is processed in the con- parency in paired random-dot displays33.
text of motion processing20,21, but it seems to be insufficient for The inability to use color as a segregation aid does not seem to
the segregation of motion components in our displays, presum- hold under all circumstances, though. The influence of noise dots
ably because MT neurons are not color-selective enough to sup- on perception by human subjects and monkeys of coherently
port two distinct surface representations based on color moving random dot patterns seems to be reduced if they are of a
information22,23. different color20,21. A number of differences between previous
Our data show that, contrary to previous proposals, the vec- studies and ours might account for these discrepancies; notably,
tors in multidirectional motion do not seem to be encoded indi- these include the requirement of subjects to make precise direc-
vidually by separate populations of neurons. Rather, the broad tion judgments in our experiment and the use of long stimulus
tuning of direction-selective neurons in areas such as MT leads durations and larger dot sizes in the previous experiments, pos-
to a population code in which the overall shape of the popula- sibly allowing eccentric tracking of individual points.
tion activity, and not (at least for acute angles) separable peaks, It is indeed possible to devise a computational model that can
encodes directions of motion. Using the complete population correctly predict the perceived directions34,35, including the two
activity probably also increases signal-to-noise ratios by using all directions perceived when viewing our three-direction stimuli,
available information4,24. from population activities such as those we report here for MT
Such an encoding scheme accounts for a number of motion neurons. Similarly, in some models of motion processing in
phenomena, especially single-direction9 and bidirectional10 after- MT36,37, acute-angled transparent motion results in single-peaked
effects produced by adaptation of subjects with narrow- and activity profiles.
wide-angled bidirectional patterns, respectively. Our data also In summary, using a combination of single-neuron record-
might help account for some results of psychophysical studies ings in monkeys and psychophysics in humans, we demonstrat-
using stimuli that combine many directions or speeds5,2527. Sim- ed a neural code that supports perception of multiple values
ilarly, our data might contribute to an explanation of the per- without requiring separate neuronal populations to encode the
ceptual phenomenon of motion repulsion14,28,29, a tendency of different stimulus values. Given the straightforwardness of this
subjects to overestimate the angle between directions in trans- coding scheme, it is probably not restricted to the encoding of

articles
motion directions. Rather, it might represent a more general In experiment 1 and 3, all dots were black; in experiment 2, the hori-
approach, suggesting that metamers similar to the ones demon- zontally moving dots were either green or an approximately isoluminant
strated here for the domain of visual motion direction might exist red. In experiment 3, RDPs were viewed through a mirror stereoscope.
for other sensory domains. The patterns moving obliquely were presented with 0.25 crossed dis-
parity, whereas the patterns moving horizontally were presented with
0.25 uncrossed disparity.
METHODS
Physiology. All animal procedures were approved by the local animal Subjects. The same twelve subjects served in all experiments. Nine were
research committee and complied with relevant laws and institutional paid volunteers naive to the research aim, and three were from within
guidelines. Standard preparation techniques were used for electrophysi- the laboratory. All had normal or corrected-to-normal visual acuity.
ological recordings from single neurons in area MT38 of three male
macaque monkeys. Cells were characterized as MT neurons based on Task. Psychophysical testing protocols were approved by the ethics com-
their receptive field locations, directional tuning and position in the cor- mission of the University of Tbingen Medical School. The different pat-
tex. MT cells were analyzed if they showed directional tuning (respons- terns were presented in random order in blocks of 160 trials. Subjects
es as a function of stimulus direction were better fit by a Gaussian than by had to report the upward direction component in each pattern. Imme-
a horizontal line, and responses to the preferred direction were at least diately after the moving dot patterns, a line was presented on the screen
fivefold larger than to the null direction). until the subject made a decision. In a two-alternative forced choice pro-
The monkey initiated each trial by depressing a lever, and had been tocol, subjects were required to indicate whether the perceived steepness
trained to respond to a change in luminance of a small spot at the top of of the upward moving component in the RDP was smaller or larger than
the fixation cross by releasing the lever. Throughout a trial, the animals the slope of the line.
gaze had to stay within 11.3 of visual angle from the fixation cross. If
the animal broke fixation or did not respond to the spots luminance Data analysis. After each trial, in a 50%-staircase procedure, the line was
change within the reaction-time window (200500 ms after the change), brought toward the perceived direction of motion. For the value of per-
the trial was terminated without reward. Only data from correctly com- ceived direction, the results of leftward and rightward motion of all sub-
pleted trials were analyzed. jects for each of the four stimuli were averaged.
During a trial, a moving random dot pattern was presented inside the Note: further material and demonstrations of our stimuli can be viewed
classical receptive field. Its speed and size was matched to the preferred on the Nature Neuroscience web (http://neurosci.nature.com/web_specials/).
properties of the cell under study. Dots were bright white squares of 0.1
width presented at a density of 5 dots per square degree on an otherwise
dark computer monitor positioned 57 cm from the animal. For single- ACKNOWLEDGEMENTS
directional patterns, all dots moved coherently in the same direction; for
This work was supported by a grant from the MWF Baden-Wrttemberg. We
bidirectional stimuli, half of the dots were assigned to each direction. Dur-
ing stimulus presentation, direction(s) continuously changed at a rate of are grateful to O. Braddick, N. Qian and R.S. Zemel for comments on previous
100 per s (ref. 39). For our analysis, the value of the spike density profile versions of the manuscript.
of the cells responses (kernel sigma, 30 ms) was taken every 50 ms. Con-
trol measurements were taken to ensure that our results, especially the
absence of two peaks in the population response to acute angled trans- RECEIVED 29 DECEMBER 1999; ACCEPTED 25 JANUARY 2000
parent motion, could be replicated using non-changing patterns.
Psychophysics: stimuli. The experiments were performed using an Apple 1. Graziano, M. S. A., Andersen, R. A. & Snowden, R. J. Tuning of MST neurons
to spiral motions. J. Neurosci. 14, 5467 (1994).
Macintosh computer and a monitor with 74.5 Hz frame rate. The spa- 2. Albright, T. D. Direction and orientation selectivity of neurons in visual area
tial resolution of the display was 33.3 pixels per degree of visual angle. MT of the macaque. J. Neurophysiol. 52, 11061130 (1984).
Our stimuli consisted of random dot patterns (RDPs) presented behind 3. Snowden, R. J., Treue, S. & Andersen, R. A. The response of neurons in areas
a stationary virtual aperture 6 in diameter. Each RDP was made up of V1 and MT of the alert rhesus monkey to moving random dot patterns. Exp.
Brain Res. 88, 389400 (1992).
200 dots (density, 3 dots per square degree; width, 3.6 arcmin) All dots 4. Britten, K. H. & Newsome, W. T. Tuning bandwidth for near-threshold
moved uniformly at 4 per s for a stimulus duration of 500 ms, with dots stimuli in area MT. J. Neurophysiol. 80, 762770 (1998).
that disappeared behind the aperture reappearing at the opposite side. 5. Williams, D. & Sekuler, R. Coherent global motion percepts from stochastic
The subjects viewed the display binocularly in a dimly lit room from a local motions. Vis. Res. 24, 5562 (1984).
constant distance of 57 cm, maintained by a chin rest. To minimize track- 6. Pasternak, T., Albano, J. E. & Harvitt, D. M. The role of directionally selective
neurons in the perception of global motion. J. Neurosci. 10, 30793086
ing eye movements, subjects were instructed to maintain fixation on a (1990).
small cross, 6 to the left of the stimulus center. Tracking eye movements 7. Jasinschi, R., Rosenfeld, A. & Sumi, K. Perceptual motion transparency: the
would change the relative directions of the motion components in the role of geometrical information. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 9,
retinal image, destroying the metameric qualities of the stimulus. 18651879 (1992).
8. Mulligan, J. B. Motion transparency is restricted to two planes. Invest.
The directions in all stimuli were arranged symmetrically around the Ophthalmol. Vis. Sci. Suppl. (1992).
horizontal direction, moving to either the right or left. (Fig. 3 shows only 9. Verstraten, F. A., Fredericksen, R. E. & Van de Grind, W. A. Movement
the rightward moving patterns.) Stimuli with potentially indistinguish- aftereffect of bi-vectorial transparent motion. Vis. Res. 34, 349358 (1994).
able direction components (Fig. 3, b1 and b2) were constructed such that a 10. Grunewald, A. & Lankheet, M. J. M. Orthogonal motion after-effect illusion
summation of the response profiles created by the two, three or five indi- predicted by a model of cortical motion processing. Nature 384, 358360
(1996).
vidual components yielded almost identical single-peaked activity profiles 11. Wilson, H. R. & Kim, J. A model for motion coherence and transparency. Vis.
(assuming a bandwidth of the underlying response profiles of 90). Addi- Neurosci. 11, 12051220 (1994).
tionally, bidirectional stimuli were used with 30 and 50 components. 12. van Wezel, R. J. A. et al. Responses of complex cells in area 17 of the cat to bi-
The strength of each individual motion component was scaled mul- vectorial transparent motion. Vis. Res. 36, 28052813 (1996).
tiplicatively by the number of dots moving in that particular direction. 13. Recanzone, G. H., Wurtz, R. H. & Schwarz, U. Responses of MT and MST
neurons to one and two moving objects in the receptive field. J. Neurophysiol.
The assumption that the neural strength of activation by a particular 78, 29042915 (1997).
direction component is a linear function of dot density is a simplifica- 14. Mather, G. & Moulden, B. A simultaneous shift in apparent directions:
tion, at least for single-direction patterns3. Given that the overall dot den- Further evidence for a distribution-shift model of direction coding. Q. J.
sity of the stimulus was kept constant and that our simple model Exp. Psychol. 32, 325333 (1980).
15. Salzman, C. D. & Newsome, W. T. Neural mechanisms for forming a
predicted approximately identical population activities over a wide range perceptual decision. Science 264 231237(1994).
of component dot densities, this simplification probably allowed the 16. Williams, D., Tweten, S. & Sekuler, R. Using metamers to explore motion
model to predict perception successfully nevertheless. perception. Vis. Res. 31, 275286 (1991).

articles
17. Maunsell, J. H. R. & Van Essen, D. C. Functional properties of neurons in 27. Zohary, E., Scase, M. O. & Braddick, O. J. Integration across directions in
middle temporal visual area of the macaque monkey. II. Binocular dynamic random dot displays: Vector summation or winner take all? Vis. Res.
interactions and sensitivity to binocular disparity. J. Neurophysiol. 49, 36, 23212331 (1996).
11481167 (1983). 28. Marshak, W. M. & Sekuler, R. Mutual repulsion between moving visual
18. Bradley, D. C., Qian, N. & Andersen, R. A. Integration of motion and targets. Science 205, 13991401 (1979).
stereopsis in middle temporal cortical area of macaques. Nature 373, 609611 29. Rauber, H. J. & Treue, S. Revisiting motion repulsion: Evidence for a general
(1995). phenomenon. Vis. Res. 39, 31873196 (1999).
19. DeAngelis, G. C., Cumming, B. G. & Newsome, W. T. Cortical area MT and 30. Van Der Smagt, M. J., Verstraten, F. A. J. & Van De Grind, W. A. A new
the perception of stereoscopic depth. Nature 394, 677680 (1998). transparent motion aftereffect. Nat. Neurosci. 2, 595596 (1999).
20. Croner, L. J. & Albright, T. D. Image segmentation enhances discrimination 31. Qian, N., Andersen, R. A. & Adelson, E. H. Transparent motion perception as
of motion in visual noise. Vis. Res. 37, 14151427 (1997). detection of unbalanced motion signals III: Modeling. J. Neurosci. 14,
21. Croner, L. J. & Albright, T. D. Segmentation by color influences responses of 73817392 (1994).
motion-sensitive neurons in the cortical middle temporal visual area. 32. Qian, N., Andersen, R. & Adelson, E. H. Transparent motion perception as
J. Neurosci. 19, 39353951 (1999). detection of unbalanced motion signals I: Psychophysics. J. Neurosci. 14,
22. Zeki, S. M. The distribution of wavelength and orientation selective cells in 73577366 (1994).
different areas of monkey visual cortex. Proc. R. Soc. Lond. B Biol. Sci. 217, 33. Braddick, O. Local and global representations of velocity: transparency,
449470 (1983). opponency, and global direction perception. Perception 26, 9951010 (1997).
23. Saito, H., Tanaka, K., Isono, H., Yasuda, M. & Mikami, A. Directionally 34. Zemel, R. S., Dayan, P. & Pouget, A. Probabilistic interpretation of population
selective responses of cells in the middle temporal temporal area (MT) of the codes. Neural Comput. 10, 403430 (1998).
macaque monkey to the movement of equiluminous opponent color stimuli. 35. Zemel, R. S. & Dayan, P. Advances in Neural Information Processing Systems
Exp. Brain Res. 75, 114 (1989). Vol. 11 (MIT Press, Cambridge, Massachusetts, in press).
24. Deneve, S., Latham, P. E. & Pouget, A. Reading population codes: A 36. Simoncelli, E. P. & Heeger, D. J. A model of neural responses in visual area
neural implementation of ideal observers. Nat. Neurosci. 2, 740745 MT. Vis. Res. 38, 743761 (1998).
(1999). 37. Wang, R. Y. A network model of motion processing in area MT of primates.
25. Watamaniuk, S. N. J. & Duchon, A. The human visual system averages speed J. Comput. Neurosci. 4, 287308 (1997).
information. Vis. Res. 32, 931941 (1992). 38. Treue, S. & Maunsell, J. H. R. Effects of attention on the processing of motion in
26. Watamaniuk, S. N. J. Ideal observer for discrimination of the global direction macaque visual cortical areas MT and MST. J. Neurosci. 19, 76037616 (1999).
of dynamic random-dot stimuli. J. Opt. Soc. Am. A Opt. Image Sci. Vis. 10, 39. Schoppmann, A. & Hoffmann, K. P. Continuous mapping of direction
1628 (1993). selectivity in the cats visual cortex. Neurosci. Lett. 2, 177181 (1976).

articles
A multimodal cortical network for

the detection of changes in the
sensory environment
Jonathan Downar1, Adrian P. Crawley2, David J. Mikulis2 and Karen D. Davis1,3
1 Institute of Medical Science, University of Toronto, and Toronto Western Research Institute, MP14-322, 399 Bathurst Street, Toronto, Ontario, M5T 2S8, Canada
2 Department of Medical Imaging, University of Toronto, and Toronto Western Research Institute, MP3-404, 399 Bathurst Street, Toronto, Ontario, M5T 2S8, Canada,
3 Department of Surgery, University of Toronto, Division of Neurosurgery, Toronto Western Hospital, and Toronto Western Research Institute, MP14-322, 399
Bathurst Street, Toronto, Ontario, M5T 2S8, Canada
Correspondence should be addressed to K.D.D. (kdavis@playfair.utoronto.ca)
Sensory stimuli undergoing sudden changes draw attention and preferentially enter our awareness.
We used event-related functional magnetic-resonance imaging (fMRI) to identify brain regions
responsive to changes in visual, auditory and tactile stimuli. Unimodally responsive areas included
visual, auditory and somatosensory association cortex. Multimodally responsive areas comprised a
right-lateralized network including the temporoparietal junction, inferior frontal gyrus, insula and
left cingulate and supplementary motor areas. These results reveal a distributed, multimodal
network for involuntary attention to events in the sensory environment. This network contains areas
thought to underlie the P300 event-related potential and closely corresponds to the set of cortical
regions damaged in patients with hemineglect syndromes.
The ability to detect changes in the sensory environment is cru- oddball protocols, in which subjects are presented with a train of
cial to survival. It is necessary to attend to such changes to eval- repeating, standard stimuli occasionally punctuated by a different
uate and modify behavior in the face of developing obstructions, oddball stimulus. Our study used a modified version of this
opportunities or threats. For this reason, changes in the sensory approach. In our protocol, the stimuli in each modality were pre-
environment, especially abrupt changes, tend to draw attention sented continuously, but alternated independently between two
involuntarily. A sensory element undergoing a change also inserts different states, A and B. Each alternation was termed a transi-
itself preferentially into awareness. For example, a hiker might tion regardless of whether the direction of the change was from
not notice the constant sound of birds chirping unless they were A to B or B to A (Fig. 1). We used these counterbalanced transi-
to stop abruptly, at which point the hiker would become aware tions between stimulus states, rather than a stereotypical oddball
of both the birds and the sudden absence of noise. When the abil- stimulus, to ensure that activations were due to a general change in
ity to attend to stimuli in the sensory world is lost, as in patients the quality of the stimulus and were not merely due to a difference
suffering from neglect syndromes, awareness of the stimuli is lost in some specific feature of the oddball stimulus as compared with
as well1,2. Understanding the mechanisms by which the brain the standard. A transition occurred in one of the 3 sensory modal-
detects changes in the sensory environment will provide us with ities every 14 seconds in a randomized sequence to minimize the
a better understanding of the mechanisms of both involuntary effects of expectancy and habituation. To identify activations, we
attention and awareness. treated transitions as the stimulus events. This approach enabled us
We used event-related fMRI to identify the network of neu- to identify cortical areas responsive to transitions within a single
roanatomical structures underlying the detection of changes in sensory modality, as well as a cortical network responsive to tran-
the sensory environment. Visual, auditory and tactile stimuli sitions in multiple sensory modalities.
were used to identify areas responding to changes in multiple
sensory modalities. These multimodal areas are of particular RESULTS
relevance to the understanding of higher-order cognitive process- Unimodally responsive areas
es such as the construction of an integrated, multisensory per- To identify unimodally responsive areas, we used a voxelwise t-
ceptual environment, the directing of attention to salient features test to compare the blood oxygenation level-dependent (BOLD)
of that environment and the selection of those features for entry signal two to eight seconds following a transition within a given
into awareness3. modality with the signal two to eight seconds after a transition
Subjects underwent fMRI while being presented with visual, in the other modalities. Based on previous studies, these time
auditory and tactile stimuli. To avoid activation due to response periods encompassed the peaks of the hemodynamic respons-
selection, planning or working memory, subjects were not required es to brief stimuli46. Brain regions satisfying the criteria for uni-
to make any sort of response during the experiment. Instead, they modal activation (see Methods) are therefore maximally
were merely instructed to observe the stimuli passively. Studies responsive to transitions in one modality as compared with the
concerning the detection of stimulus events frequently involve others (Table 1; Figs. 2 and 3).

articles
Fig. 1. Excerpt of stimulus-presentation protocol. Visual, auditory and tactile stimuli were presented to subjects simultaneously during imaging. The
visual stimulus was either the blue or the red abstract figure shown; the auditory stimulus was either the sound of running water or the sound of
croaking frogs (actual waveforms shown); the tactile stimulus was either circular brushing or rapid tapping of the lower right leg, using a 6 10 cm
shower brush. At 14-s intervals, in pseudo-random order, 1 of the 3 stimulus modalities underwent a transition from one stimulus type to the other.
A total of 44 transitions (15 visual, 15 auditory, 14 tactile) occurred during the course of the experiment. V, visual; A, auditory; T, tactile.
Visual-specific areas comprised bilateral visual association these areas during the 6 seconds before and after each transi-
cortices of both the dorsal and ventral visual-processing streams, tion, with a 2-second delay to accommodate the hemodynamic
including the fusiform gyrus and middle occipital gyrus. The for- response. Transition-related intermodal inhibition thus appeared
mer area includes inferior temporal regions known to be involved to affect only the visual modality.
in the identification of complex figures such as objects and faces7,
whereas the latter corresponds to area V3 of visual association Multimodally responsive areas
cortex. In addition, there was a locus of activation in the right To identify areas responsive to transitions in all three modalities,
superior parietal lobule, on the superior bank of the intrapari- we used a voxelwise correlation to the predicted hemodynamic
etal sulcus. This area activates for shifts of attention in the visu- response to brief stimuli spaced 14 seconds apart, based on pre-
al modality8. vious studies46. Brain regions satisfying the criteria for multi-
Auditory-specific areas comprised Brodmann areas 41, 42 and modal activation (see Methods) are presented in Table 1 and Figs.
22 of the right and left superior temporal gyri. These areas cor- 2, 3 and 4.
respond to primary and secondary auditory cortex. No other Areas activated by transitions in all three sensory modalities
auditory-specific loci of significant activation were noted. comprised a distributed network of frontal, medial, temporopari-
Tactile-specific areas comprised the secondary somatosenso- etal and insular cortical areas (Fig. 4). The network showed a
ry cortex (S2). Although stimulation was applied only to the right greater overall volume of activated voxels in the right hemisphere
leg (Fig. 1), activation was observed bilaterally, consistent with (5596 mm3) as compared with the left (1244 mm3; Table 1). The
the bilateral receptive fields of many S2 neurons911. No activa- most prominent area of activation was located at the right tem-
tion was observed in the primary somatosensory cortex (S1) for poroparietal junction (TPJ), immediately posterior to the sec-
the right leg. ondary auditory and somatosensory cortices (Figs. 2 and 4). A
Inhibitory mechanisms for attention, both within and much smaller activation was identified in left TPJ. Ventral and pos-
between modalities, are topics of considerable interest. An fMRI terior to the activation in the right TPJ was a small activation in
study of visual cortical activity during binocular rivalry found the right middle temporal gyrus (Figs. 2 and 4). Frontal areas
that extrastriate areas responsive to a particular stimulus show responsive to transitions in all modalities were located bilaterally in
deactivations when a competing stimulus becomes perceptual- the inferior frontal gyrus and frontal operculum. As in the TPJ,
ly dominant12. Intermodal inhibition of visual cortical activity frontal activation was more extensive in the right hemisphere than
during a tactile discrimination task is also reported13. To assess in the left. There were also foci of activation in an anterior and a
the possibility of intermodal inhibition in our study, we exam- posterior region of the right insula. Medial multimodal activations
ined the averaged responses of unimodal-specific areas to tran- were located in the left anterior cingulate and the supplementary
sitions in their own versus the other modalities (Fig. 3). motor areas (CMA and SMA).
Although all unimodal areas showed consistent activation for
their particular modalities, only the visual-specific areas showed DISCUSSION
significant deactivation in response to transitions in other The results of our study reveal a distributed cortical network for
modalities (auditory, t88 = 3.69, p < 0.001; tactile, paired t82 = the detection of changes in the sensory environment, with both
6.04, p < 107), as measured by comparing the BOLD signal of unimodal and multimodal components.

articles
The unimodal components of this network corresponded to Visual unimodal areas showed a significant deactivation in
visual association cortex, primary and secondary auditory cor- response to transitions in other modalities. This finding is con-
tex, and S2. It is noteworthy that V1, V2 and S1 did not show a sistent with a study showing that attention to tactile stimuli deac-
significant response to transitions. Because all stimuli were tivates visual association areas 13. Our study indicates that
applied continuously throughout the experiment, it is unlikely attention-drawing transitions in the auditory modality can also
that these areas were simply inactive during scanning. It is more deactivate visual association cortex. Such intermodal inhibition
likely that their overall level of activation did not change signifi- did not seem to affect auditory or somatosensory association
cantly during or following transitions. The activity of these areas areas, however. The reasons for the unique vulnerability of visu-
may thus reflect a more literal representation of sensory input, al association cortex to transition-related intermodal inhibition
less influenced by habituation or attentional modulation than remain to be investigated.
that of secondary and association cortices. The largest of the multimodal nodes of activation was found
The unimodal auditory activations observed in primary and in the right TPJ. The next highest volumes of activation were in
secondary auditory cortex are consistent with recent findings the right inferior frontal cortex, the left SMA and CMA and right
concerning the detection of sensory events in the auditory modal- anterior insula, with smaller activations in the left TPJ and infe-
ity. Deviant or oddball auditory stimuli embedded in a train of rior frontal cortex, the right posterior insula and the right mid-
standard, repeating stimuli normally evoke a mismatch negativ- dle temporal gyrus. Nonetheless, the total volume of activated
ity (MMN) observable in electroencephalographic (EEG) record- voxels in the two hemispheres, the specific areas activated and
ings. An EEG study of patients with focal cortical lesions suggests their order of prominence show remarkable agreement with neu-
auditory association cortex as a major generator of the MMN for roanatomical correlates of sensory neglect. In this syndrome, the
oddballs presented to the contralateral ear14. Our data supports ability to attend to salient stimuli on one side of the body (usu-
this hypothesis by demonstrating the activation of auditory cor- ally the left), whether voluntarily or involuntarily, is lost as a
tex in response to changes in continuous auditory input. result of neurological damage. Patients suffering from neglect
Table 1. Brain regions showing significant activation in response to transitions in visual, auditory, tactile and all three
sensory modalities.
Brain region Brodmann area Coordinates Volume (mm3) t
x y z
Unimodal activations
Visual
R fusiform gyrus 19/37 29 60 14 5835 16.97
L fusiform gyrus 19/37 31 63 15 5689 15.75
R middle occipital gyrus 18/19 31 81 4 2371 16.97
L middle occipital gyrus 18/19 29 81 7 2912 17.53
R superior parietal lobule 18/19 25 62 47 801 12.56
Auditory
R superior temporal gyrus 22/41/42 53 21 5 2185 9.66
L superior temporal gyrus 22/42 51 37 10 1858 10.32
L superior temporal gyrus 41/42 53 15 3 403 8.97
Tactile
R postcentral gyrus (S2) 40 50 28 25 2177 9.15
L postcentral gyrus (S2) 40 59 28 27 1259 8.34
Multimodal activations
R temporoparietal junction 22/39/40 54 42 13 3657 12.16

L temporoparietal junction 22/39 54 48 10 449 8.25
R middle temporal gyrus 21 57 57 2 338 7.03
R inferior frontal gyrus 44/45 43 6 5 822 8.73
L inferior frontal gyrus 6/44 51 8 5 226 7.41
L SMA/CMA 6/32 8 4 41 569 7.55
R anterior insula 27 18 10 564 8.17
R posterior insula 36 16 6 215 5.97
Activations shown are based on a voxelwise p < 0.00032 (unimodal transitions) or p < 0.00036 (multimodal transitions) and a minimum cluster size of 200
mm3. Average time course of all voxels in each cluster were used to calculate t-values. For multimodal areas, t-values are converted from r-coefficients. S2,
secondary somatosensory cortex; SMA, supplementary motor area; CMA, cingulate motor area.

articles
Fig. 2. Surface rendering of brain regions activated by transitions of the

visual, auditory and tactile stimuli. Note that unimodal activations are
bilateral and correspond predominantly to association cortices,
whereas multimodal activations are strongly lateralized to the right
hemisphere and correspond to the temporoparietal junction, inferior
frontal gyrus and insula. On the left medial wall, the supplementary and
cingulate motor areas are also prominently activated, even though sub-
jects were not required to make responses or movements during task
performance. Activations represent averaged data from 10 subjects,
superimposed on the standardized brain of one subject. L, left; R, right.
also show a profound, multimodal loss of awareness of stimuli

in the neglected field, concurrent with their attentional deficits2,15.
Hemineglect is far more commonly observed following damage
Visual transitions
to the right hemisphere than the left. Lesions of the right TPJ are
Auditory transitions All transitions
the most common neuroanatomical correlate of hemineglect1,16,17,
Tactile transitions
followed by lesions of the CMA, SMA or prefrontal cortex, espe-
cially in the vicinity of Brodmann area 44 (ref. 1). It should be
remembered that fMRI signal provides only an indirect measure involved in the planning of motor responses, not in the shifting of
of the neural activity of a given area. Nonetheless, the remark- attention or the detection of sensory events. Previously report-
able degree of overlap between the known anatomical correlates ed CMA and SMA activation during spatial shifts of attention
of hemineglect and the network identified in our study empha- might be attributed to the planning and execution of motor
sizes the close relationship between the detection of salient events responses during task performance8. However, we observed SMA
in the sensory environment and the selection of these salient stim- and CMA activation even though our subjects made no such
uli for entry into awareness. Furthermore, it underscores the responses during the course of the experiment. The activations
importance of the identified cortical regions to the performance we observed in CMA and SMA could be attributed to the plan-
of both functions across multiple sensory modalities. ning of unexecuted, involuntary motor orienting responses to
The observation of SMA and CMA activation in this study is the changing sensory input. A more interesting possibility, how-
of particular interest. These areas are generally considered to be ever, is that these regions may have an attentional role in senso-
Fig. 3. Responses to visual, auditory and tactile transitions. Averaged event-related hemodynamic responses to transitions in each stimulus modality
are shown for representative unimodally and multimodally responsive areas. (a) Bilateral fusiform gyrus. (b) Bilateral superior temporal gyrus.
(c) Bilateral secondary somatosensory cortex (S2). (d) Bilateral TPJ. Unimodal areas (ac) respond only to transition events within visual, auditory
and tactile modalities, respectively, whereas the TPJ (d) responds bilaterally to transitions in all three modalities. A significant (p = 0.01) deactivation
of the fusiform gyrus in response to auditory and tactile transitions is visible in (a). Multimodal activation in the right anterior insula, not shown in Fig.
2, is visible in (d). The Talairach z coordinate of each slice is indicated in the upper left-hand corner. Error bars indicate s.e.
a c
Change in signal (%)
Post-transition time (s) Post-transition time (s)

b d
Post-transition time (s) Post-transition time (s)

articles
Temporoparietal junction Inferior frontal gyrus
Anterior/posterior insula SMA/CMA Middle temporal gyrus

Fig. 4. Multimodally responsive brain regions. The plane coordinate of each slice is indicated in the upper left-hand corner. SMA, supplementary
motor area; CMA, cingulate motor area.
ry processing as well as their more generally accepted role in the affected by lesions of the parietal cortex immediately dorsal to
planning and execution of movements. This view is supported the TPJ21,23,24. These EEG findings support the results of our study
by the observation that lesions of the SMA and CMA can give in suggesting the importance of TPJ and prefrontal cortex in the
rise to attentional deficits such as hemineglect1. In addition, detection of salient stimuli across multiple modalities.
recordings of epicortical field potentials in humans demonstrate There remains the possibility that additional multimodal areas
the involvement of the SMA, CMA and other premotor areas in were not detected in our study because their responses habitu-
the anticipation of and attention to forthcoming stimuli18. Our ated rapidly over the first few transitions. For example, although
study cannot conclusively demonstrate the involvement of pre- activation in the dorsolateral prefrontal cortex was not observed
motor areas in attentional processes independent of response in our study, evidence from EEG studies of patients with focal
planning; however, the possibility bears further investigation. cortical lesions suggests that this area may be involved in detect-
The involvement of the TPJ and inferior frontal gyrus in ing novel sensory events in visual, auditory and tactile modali-
detecting changes in continuous streams of sensory input is con- ties14,21,24. Similarly, a study using fMRI indicates that novel
sistent with EEG studies of the effects of focal cortical lesions on auditory events activate the right prefrontal cortex25. In addition,
P300 event-related potentials (ERPs), which are widely used as
markers of attention to salient sensory stimuli19. Lesions of the
TPJ produce a marked reduction in the amplitude of both the
P3a ERP, which is normally elicited by unexpected, task-irrele-
vant stimuli, and the P3b ERP, which is normally elicited by
expected, task-relevant stimuli20,21. Such reductions are observed
for visual, auditory and tactile stimuli following injury of the

TPJ2123. Lesions of the prefrontal cortex also reduce the P3a ERP
for all three modalities 21,24. However, neither P3a nor P3b is
Fig. 5. Responses of the right TPJ to each direction of transition in each

modality. Averaged event-related hemodynamic responses to transitions
from stimulus A to stimulus B as well as stimulus B to stimulus A are
shown for each sensory modality. Activations for both transition types
exclude the possibility that the event-related activations in Fig. 3d reflect
an increase in stimulus intensity or contrast associated with just one of
the two types of transitions. Similar activation for both types of transi- Post-transition time (s)
tions was verified for all unimodal and multimodal areas reported.

articles
a number of studies implicate a set of rapidly habituating areas, Data processing and analysis. Preprocessing, volumetry and analysis
including prefrontal, posterior parietal and hippocampal areas, in were performed using BrainVoyager 3.5 (Brain Innovation, Frankfurt,
the detection of novel events and their subsequent encoding into Germany), implemented on a 450 MHz Pentium II platform. Function-
memory26,27. Furthermore, evidence from EEG studies using task- al data were corrected for within-frame time of acquisition, high-pass
filtered with a cut-off at 0.0175 Hz to remove slow drifts in signal inten-
relevant target and task-irrelevant nontarget oddballs suggests
sity (period >1 min), coregistered with the 3-dimensional anatomical
that the context in which task-irrelevant stimuli are presented, images, transformed into standard stereotactic space31 and resampled at
more generally than their novelty, may affect the physiological a resolution of 3 3 3 mm. Data were smoothed using a Gaussian ker-
response28. Our study was not designed to identify areas respon- nel of 4 mm full width at half maximum to accommodate anatomical
sive only to novel events. This issue will be the focus of future and functional-anatomical variation between subjects. Individual sub-
investigations. jects data were averaged together for group analysis. During the gener-
A widespread, multimodal cortical and subcortical network ation of statistical maps, data were further interpolated to 1 1 1 mm
for the shifting of attention to salient features of the sensory envi- resolution to facilitate estimation of the volume of activation in each
ronment was first proposed on the basis of neuroanatomical and area. To identify unimodally responsive areas, we used a voxelwise t-test
lesion studies29. The results of our study provide an illustration of to compare the BOLD signal two to eight s after a transition within a
given modality to the signal two to eight s after transitions in the other
the unimodal and multimodal elements of such a network in
modalities. We used an initial threshold t > 3.70 for individual 1-mm3
action. The agreement of these results with previous EEG and voxels (n = 130 frames, uncorrected p = 0.00032). To identify multi-
lesion studies provides converging evidence for a close relation- modally responsive areas, we used a voxelwise correlation to the pre-
ship between the detection of salient stimuli in the multisensory
dicted hemodynamic response to brief stimulus events spaced 14 s apart,

environment and the entry of those stimuli into awareness, a based on previous studies4,6. We used an initial threshold of r > 0.20,
process mediated by a widespread network of cortical areas com- equivalent to t > 3.61 (n = 314 frames, uncorrected p = 0.00036). The
prising the unimodal sensory association cortices and the mul- total volume of each statistical map was 1,492,559 mm3; hence, at the
timodal TPJ, CMA, SMA, insula and inferior frontal gyrus. thresholds used, 400 voxels in each entire map would be expected to
show activation due to type I errors. These voxels would also be expect-
ed to display a certain degree of clustering, as they were interpolated from
3 3 3 mm-resolution, smoothed functional data. We therefore
METHODS required a conservative minimum cluster size of 200 contiguous 1-mm3
Subjects. Subjects were 6 male and 4 female right-handed individuals
voxels for all reported activations. We initially examined the event-relat-
aged 1841 (mean age s.d., 27.3 7.2), with no history of neurological
ed average response of each area to each transition direction separately
injury. All subjects gave written informed consent for the experimental
(transitions from A to B were considered separately from transitions from
procedures, approved by the University of Toronto Human Subjects
B to A) to rule out areas whose responses, ostensibly to transitions per
Review Committee.
se, merely reflected the different physical characteristics of the stimuli
presented (for instance, intensity, contrast). In such a scenario, the areas
Stimuli. Visual, auditory and tactile stimuli were presented simultane-
observed might show a large or rapid activation for the transition from A
ously during functional imaging. Visual stimuli were back-projected onto
to B but only a small or gradual deactivation for the opposite transition
a screen viewed by the subject through a mirror incorporated into the
from B to A. The average response would seem to be an activation for all
head coil. The visual stimuli were two simple, abstract shapes, one red
transitions, even though it truly reflected only the different properties of
and one blue (Fig. 1). Auditory stimuli were generated by two pneumatic
the two stimuli. We excluded this possibility by analyzing the responses of
transducers and conducted through a pair of eight-mm plastic tubes into
all areas to the two kinds of transitions separately, as in Fig. 5. For all uni-
a set of headphones placed over the subjects ears. The auditory stimuli
modal and multimodal areas presented, we verified activation in response
were the sound of running water and the sound of croaking frogs. Tactile
to transitions from A to B as well as B to A. In addition, for each multi-
stimuli delivered with a 6 10 cm shower brush to the subjects lower
modal area, we examined the event-related average response to transi-
right leg consisted of circular brushing with the bristles (1 Hz) and
tions in each modality as in Fig. 3. Areas that did not respond to
steady tapping with the smooth back of the brush (3 Hz). Visual and
transitions in all three sensory modalities were discarded from the list of
auditory stimuli were produced using a VCR connected to the projector
multimodal activations. This additional step was necessary to exclude
and audio transducer. A visual signal, projected outside the subjects field
areas that merely showed an exceptionally strong response to transitions
of view, prompted the experimenter to change the tactile stimulation
in one of the three modalities rather than a consistent response across
type. During the experiment, a total of 15 visual, 15 auditory and 14 tac-
all three.
tile transitions occurred in a pseudo-random sequence, spaced at 14-s
intervals to allow the BOLD signal to return to baseline between transi-
tions. Subjects were instructed simply to observe the stimuli passively. ACKNOWLEDGEMENTS
After imaging, all subjects reported that the transitions had drawn their This study was supported by grants to K.D.D. from the Whitehall Foundation
attention and that they had not counted the transitions. and the Medical Research Council of Canada.
Imaging. A 1.5 T MRI system (Echospeed, GE Medical Systems, Mil-

waukee, Wisconsin) and a standard quadrature head coil were used to RECEIVED 23 AUGUST; ACCEPTED 22 DECEMBER 1999
obtain all images. For anatomical images, a 3-dimensional SPGR sequence
(flip angle = 45, TE = 5 ms, TR = 25 ms) was used to generate 124 sagit- 1. Vallar, G. Spatial hemineglect in humans. Trends Cogn. Sci. 2, 8797 (1998).
tal slices (1.5 mm thick) with 256 256 matrix size and 30 30 cm field 2. Halligan, P. W. & Marshall, J. C. Neglect of awareness. Conscious. Cogn. 7,
of view resulting in an anatomical resolution of 1.5 1.17 1.17 mm. 356380 (1998).
For analysis, the anatomical images were resampled to 1 1 1 mm 3. Mesulam, M. M. From sensation to cognition. Brain 121, 10131052 (1998).
using sinc interpolation. For functional imaging, 25 contiguous axial 4. Pollmann, S., Wiggins, C. J., Norris, D. G., von Cramon, D. Y. & Schubert, T.
Use of short intertrial intervals in single-trial experiments: a 3T fMRI-study.
slices (4 mm thick) were chosen to provide whole-brain coverage. Func- Neuroimage 8, 327339 (1998).
tional images were acquired with a gradient echo sequence using a single- 5. Buckner, R. L. Event-related fMRI and the hemodynamic response. Hum.
shot spiral trajectory through k-space30, flip angle = 85, TE = 40 ms, TR Brain Mapp. 6, 373377 (1998).
= 2000 ms, a 64 64 matrix size and a 20 20 cm field of view for a func- 6. Aguirre, G. K., Zarahn, E. & DEsposito, M. The variability of human, BOLD
hemodynamic responses. Neuroimage 8, 360369 (1998).
tional resolution of 3.125 3.125 4 mm and an inter-frame interval of 7. Sergent, J., Ohta, S. & MacDonald, B. Functional neuroanatomy of face and
2 s. The first 3 frames were discarded to allow for signal equilibration, object processing. A positron emission tomography study. Brain 115, 1536
giving a total of 314 frames used in analysis. (1992).

articles
8. Kim, Y. H. et al. The large-scale neural network for spatial attention displays 20. Yamaguchi, S. & Knight, R. T. P300 generation by novel somatosensory
multifunctional overlap but differential asymmetry. Neuroimage 9, 269277 (1999). stimuli. Electroencephalogr. Clin. Neurophysiol. 78, 5055 (1991).
9. Robinson, C. J. & Burton, H. Somatotopographic organization in the second 21. Swick, D. & Knight, R. T. in The Attentive Brain (ed. Parasuraman, R.)
somatosensory area of M. fascicularis. J. Comp. Neurol. 192, 4367 (1980). 144162 (MIT Press, Cambridge, Massachusetts, 1998).
10. Dong, W. K. et al. Nociceptive responses of trigeminal neurons in SII-7b 22. Knight, R. T., Scabini, D., Woods, D. L. & Clayworth, C. C. Contributions of
cortex of awake monkeys. Brain Res. 484, 314324 (1989). temporal-parietal junction to the human auditory P3. Brain Res. 502,
11. Maldjian, J. A. et al. Mapping of secondary somatosensory cortex activation 109116 (1989).
induced by vibrational stimulation: an fMRI study. Brain Res. 824, 291295 (1999). 23. Yamaguchi, S. & Knight, R. T. Anterior and posterior association cortex
12. Tong, F., Nakayama, K., Vaughan, J. T. & Kanwisher, N. Binocular rivalry and contributions to the somatosensory P300. J. Neurosci. 11, 20392054
visual awareness in human extrastriate cortex. Neuron 21, 753759 (1998). (1991).
13. Kawashima, R., OSullivan, B. T. & Roland, P. E. Positron-emission 24. Knight, R. T. Decreased response to novel stimuli after prefrontal lesions in
tomography studies of cross-modality inhibition in selective attentional tasks: man. Electroencephalogr. Clin. Neurophysiol. 59, 920 (1984).
closing the minds eye. Proc. Natl. Acad. Sci. USA 92, 59695972 (1995). 25. Opitz, B., Mecklinger, A., Friederici, A. D. & von Cramon, D. Y. The
14. Alain, C., Woods, D. L. & Knight, R. T. A distributed cortical network for functional neuroanatomy of novelty processing: integrating ERP and fMRI
auditory sensory memory in humans. Brain Res. 812, 2337 (1998). results. Cereb. Cortex 9, 379391 (1999).
15. Driver, J. & Mattingley, J. B. Parietal neglect and visual awareness. Nat. 26. Knight, R. T. & Nakada, T. Cortico-limbic circuits and novelty: a review of
Neurosci. 1, 1722 (1998). EEG and blood flow data. Rev. Neurosci. 9, 5770 (1998).
16. Ro, T., Cohen, A., Ivry, R. B. & Rafal, R. D. Response channel activation and 27. Tulving, E., Markowitsch, H. J., Kapur, S., Habib, R. & Houle, S. Novelty
the temporoparietal junction. Brain Cogn. 37, 461476 (1998). encoding networks in the human brain: positron emission tomography data.
17. Friedrich, F. J., Egly, R., Rafal, R. D. & Beck, D. Spatial attention deficits in Neuroreport 5, 25252528 (1994).
humans: a comparison of superior parietal and temporal-parietal junction 28. Katayama, J. & Polich, J. Stimulus context determines P3a and P3b.
lesions. Neuropsychology 12, 193207 (1998). Psychophysiology 35, 2333 (1998).
18. Ikeda, A. et al. Cognitive motor control in human pre-supplementary motor 29. Mesulam, M. M. A cortical network for directed attention and unilateral
area studied by subdural recording of discrimination/selection-related neglect. Ann. Neurol. 10, 309325 (1981).
potentials. Brain 122, 915931 (1999). 30. Glover, G. H. & Lee, A. T. Motion artifacts in fMRI: comparison of 2DFT
19. Coull, J. T. Neural correlates of attention and arousal: insights from with PR and spiral scan methods. Magn. Reson. Med. 33, 624635 (1995).
electrophysiology, functional neuroimaging and psychopharmacology. Prog. 31. Talairach, J. & Tournoux, P. Co-Planar Stereotaxic Atlas of the Human Brain
Neurobiol. 55, 343361 (1998). (Thieme Medical, New York, 1988).

articles
The neural mechanisms of top-

down attentional control
J. B. Hopfinger1, M. H. Buonocore2 and G. R. Mangun3
1 Center for Neuroscience and Department of Psychology, One Shields Ave.,University of California, Davis, Davis, California 95616, USA
2 Department of Radiology, University of California, Davis Medical Center, 4701 X St., Sacramento, California 95817, USA
3 Center for Cognitive Neuroscience, LSRC Bldg., Rm. B203, Duke University, Durham, North Carolina 27708, USA
Correspondence should be addressed to G.R.M. (mangun@duke.edu)
Selective visual attention involves dynamic interplay between attentional control systems and sensory brain
structures. We used event-related functional magnetic resonance imaging (fMRI) during a cued spatial-
attention task to dissociate brain activity related to attentional control from that related to selective process-
ing of target stimuli. Distinct networks were engaged by attention-directing cues versus subsequent targets.
Superior frontal, inferior parietal and superior temporal cortex were selectively activated by cues, indicating
that these structures are part of a network for voluntary attentional control. This control biased activity in
multiple visual cortical areas, resulting in selective sensory processing of relevant visual targets.
Selective attention enables us to focus awareness on objects and analysis13,14 allowed us to combine the spatial resolution necessary
events that are relevant to our immediate goals. Spatial attention, for localization of neural activity, which this technique provides,
the selective direction of visual attention toward a location, can with neuroimaging methods that selectively extract components
occur covertly, without overt movements of the head or eyes. of hemodynamic activity15 correlated with distinct aspects of com-
Theoretically, mechanisms of covert, voluntary spatial attention plex-task performance.
can be decomposed into elementary mental operations: disen- Here we used event-related fMRI methods to determine which
gaging attention from the current focus, orienting attention to a brain regions were involved in attentional-control processes and
new locus and selectively modulating new stimulus inputs are to distinguish these from areas involved in subsequent selective
three stages proposed in current models1,2. Attentional disen- processing of target stimuli. This was accomplished by separate-
gagement and voluntary orienting can be considered aspects of ly convolving the onsets of the cue and target stimuli with syn-
top-down attentional control, whereas subsequent selective mod- thetic hemodynamic response functions. Subjects oriented
ulation of sensory inputs reflects the result of this top-down con- attention covertly (without moving their eyes) to spatial locations
trol on sensory information processing. based on instructive cues shown at fixation (Fig. 1). Subsequent-
Studies in neurological patients and physiological studies in ly, reversing checkerboard stimuli were presented at both sides of
humans and animals implicate a network of cortical and subcor- the visual field, and subjects were required to discriminate whether
tical regions that support visual selective attention 36. Neu- the checkerboard at the cued location contained some gray checks
roimaging studies identify areas involved in spatial attention in the or only black and white checks and to respond accordingly. All
frontal, parietal, temporal and occipital lobes as well as in subcor- stimuli in the opposite (uncued) hemifield were to be ignored.
tical structures711. However, although neuroimaging studies impli- Based on previous studies implicating various brain regions in
cate various brain regions in selective attention, inherent limitations attentional control5,7,8,11, we hypothesized that attentional-control
of classical neuroimaging analysis impair efforts to relate neural processes in response to an instructive cue directing spatial atten-
structures/networks to specific attentional operations. Specifical- tion would activate a network of regions in frontal and parietal
ly, such methodologies use averaging of hemodynamic responses cortex. Additionally, based on ERP studies in humans12, regions
over seconds or minutes, a time course that is too long to permit of occipital cortex corresponding to portions of visual space indi-
direct viewing of brain activity related to the subcomponents of cated by the cue should show spatially specific changes in activity
task performance. As a result, human neuroimaging studies to date in anticipation of selective processing of target stimuli. We also
are only partly successful in distinguishing between neural activi- expected lateralized changes in activity of occipitotemporal regions
ty related to top-down attentional-control processes and activity in response to the bilateral target stimuli to reflect spatial atten-
related to selective sensory and motor processing. tion711. In line with these predictions, we found distinct patterns
Recordings of event-related neuroelectric potentials (ERPs) allow of neural activation for attention-directing cues as compared with
the selective averaging of responses to different classes of events those for subsequent target stimuli, thereby distinguishing top-
and, thus, serve to index mechanisms of attentional control sepa- down attentional control from selective modulations of target pro-
rately from selective stimulus processing during spatial attention12. cessing in sensory structures.
However, although ERPs provide information about the timing of
neural processes, the limited spatial resolution of this approach hin- RESULTS
ders identification of the neural structures involved in attentional Subjects discriminated black and white checkerboards from those
control. Advances in functional magnetic resonance imaging (fMRI) containing some gray checks at the cued location with a mean

articles
Fig. 1. Stimuli and timing. For simplicity, we show here the inverse
black/white contrast of the actual stimulus screen; subjects saw a black Cue
background with white outline boxes and fixation cross, and the cues were 500 ms
overlapping isoluminant yellow and blue arrows (pointing in opposite direc- +
tions). Subjects were told which color arrow to attend for the entire ses-
sion and were required to covertly orient their attention to the visual field ISI
location indicated by that arrow. After a variable interstimulus interval 1000 ms (17%) or
8160 ms (83%)
(1000 or 8160 ms), the target checkerboards appeared bilaterally, and sub- +
jects were required to perform a discrimination of the checkerboard stim-
uli at the attended location only.
Time Target
750 ms
4 Hz reversal
+
accuracy of 83.9% correct responses. This corresponded to a ITI

mean d score of 2.99. 7910 ms
We investigated circuitry involved in attentional control by iden- +
tifying regions activated in response to the cue. Multiple cortical
areas were activated following cue presentation, but before target
frontal activations tended to be larger in the left hemisphere, regard-
presentation (Fig. 2; Table 1). These regions were activated regard-
less of cue direction. In addition, cue stimuli generated activity
less of the cue direction: brain regions activated by rightward- and

along the superior temporal sulcus (STS). With the exception of
leftward-pointing cues were highly similar (with the exception ofthe superior parietal lobule, none of these regions were activated
the extrastriate cortex; see below). Regions of the inferior parietal
in separate sensory control scans in which cues were viewed pas-
lobule (in the lateral intraparietal sulcus, or IPS), the superior pari-
sively or were irrelevant to the task (and, therefore, did not instruct
etal lobule (SPL) and the posterior cingulate cortex (PC) in boththe subjects to direct attention to a single location). Posterior occip-
hemispheres were activated in response to the instructive cues. ital regions of cortex were activated in response to the cues, but
Regions of the lateral and medial superior frontal lobes, including
these regions were also activated by cues in the sensory-control
the frontal eye fields (FEF), were also activated bilaterally. These
scans, indicating that they represented simple sensory processing
of the cues. Additionally, the cues pro-
Cues duced bilateral activations in the region of
the insula near the putamen.
Leftward Rightward In contrast, the target stimuli evoked
neural activity in brain areas that were
largely distinct from those activated by the
L SFG L SFG attention-directing cues (Fig. 3; Table 2).
L MFG R MFG L MFG R MFG Bilateral activations in the supplementary
motor area (SMA) extending into regions
of the midcingulate gyrus and activations
L SPL R SPL L SPL R SPL surrounding the central sulcus were found
L IPS R IPS L IPS R IPS in response to targets whether attention
was directed to the right or left. Ventro-
lateral prefrontal areas were also bilater-
ally activated by the targets. None of these
L MFG
regions were activated by the cues. Only
L MFG L SPL L SPL
the superior parietal lobule was found to
L SFG L IPS L SFG
L IPS
Fig. 2. Activity related to attentional control.

Data for brain regions significantly activated in
L LOG response to the cue stimuli were overlaid onto
L LOG
a brain rendered in 3D. Left column, activa-
L STS/STG L STS/STG tions to cues instructing subjects to orient
attention to the left visual field location. Right
column, activations to cues instructing subjects
L SPL L SPL to attend the right visual field location. Top
L SFG L SFG panels, dorsal view of the brain (frontal pole at
L Post. top); middle and bottom panels, lateral and
Cingulate L Post. medial views, respectively, of the left hemi-
Cingulate sphere. Labels indicate the brain regions
referred to in Table 1. The Z-values and stereo-
tactic coordinates for the regional maxima are
given in Table 1. SFG, superior frontal gyrus;
MFG, middle frontal gyrus; SPL, superior parietal
lobule; IPS, intraparietal sulcus; STS, superior tem-
poral sulcus; STG, superior temporal gyrus; LOG,
lateral occipital gyrus.

articles
be activated in response to both targets and cues. Additionally, selective attentional activations in response to the direction of
bilateral targets activated several regions of the ventral and dorsal the instructive cue, but before target presentation, using the atten-
occipital cortex of both hemispheres. Cue and target responses tion-related activations to targets as regions of interest. Direct
were directly statistically compared (Tables 1 and 2; Fig. 4). statistical evaluation of the cue-left versus cue-right effects
To investigate the consequences of lateralized spatial attention revealed relative increases of activity in extrastriate cortex con-
on stimulus processing, the event-related neural activity evoked tralateral to the direction of attention. These contralateral cue-
by the bilateral target stimuli was evaluated for attention to the related activations overlapped closely with the extrastriate regions
left versus right. In line with previous neuroimaging studies in showing attentional modulations in response to the subsequent
humans9,10,1619 and related findings in animals3, attending to the targets (Fig. 6). That is, there was a relative increase in activity
left visual field increased activity in the right ventral occipital cor- in the visual cortical regions representing the spatial locations of
tex, whereas attending to the right increased activity in the left the expected target stimuli before the target stimuli were actual-
occipital cortex (Fig. 5a). In addition, activity in a more dorsal ly presented. These differential activations were not related to
region of visual cortex was also modulated by spatial attention. inherent sensory features of the cues that activated separate cor-
To relate these attention-related activations to specific visu- tical regions in control sessions.
al-field areas, the vertical and horizontal meridian were mapped
in two subjects to identify the borders of the early cortical visual DISCUSSION
areas20,21. Although the spatial smoothing used in the present This study used event-related fMRI methods to distinguish
analysis tended to blend together the attention-related activa- between neural networks involved in top-down attentional-
tions in adjacent visual cortical areas, the contralateral attention- control processes and those participating in the subsequent spa-
related activations in occipital cortex can be seen to cross multiple tially selective attentional processing of target stimuli. A network
visual area borders from V2 through V4 (Fig. 5b), an observa- of cortical areas including superior frontal, inferior parietal and
tion in line with other reports10,18,22. superior temporal brain regions were implicated in top-down
Models of spatial attention propose that attention effects on attentional control because they were found to be active only
target processing result from a gain-control mechanism that in response to instructive cues. In contrast, other regions of the
enhances the excitability of extrastriate neurons coding attend- cortex, including the ventrolateral prefrontal cortex, anterior
ed regions of visual space4,9,23. To investigate this possibility, we cingulate and supplementary motor area, were found to be
further examined extrastriate cortex for evidence of spatially selectively activated by the target stimuli, suggesting that these
Table 1. Event-related activations to cue stimuli and statistical contrasts.

Leftward-directing cue Rightward-directing cue Cues > targets
Coordinates Z-score Coordinates Z-score Coordinates Z-score p
x y z x y z x y z (corrected)
Frontal
Left SFG* 16 48 36 4.56 16 48 36 4.10 16 48 36 7.10 p < 0.001
Left SFG (lateral)* 44 36 16 2.96 44 36 16 4.30 52 32 16 6.93 p < 0.001
Left MFG* 20 4 52 4.36 20 0 48 5.53 32 16 48 7.67 p < 0.001
Right MFG* 24 0 56 3.96 24 4 52 4.37 20 8 48 4.95 p < 0.01
Parietal
Left SPL* 16 52 56 3.27 16 56 56 4.21 12 56 56 4.41 p > 0.05
Right SPL* 24 40 56 3.84 20 40 52 2.76 8 40 56 4.27 p > 0.05
Left IPS* 44 64 32 3.79 40 64 28 4.06 44 64 32 7.01 p < 0.001
44 48 36 3.29 40 48 32 3.88 44 48 32 7.29 p < 0.001
Right IPS* 40 68 28 4.23 40 64 32 3.83 44 60 36 6.63 p < 0.001
36 48 36 3.46 32 48 32 4.23 36 44 32 7.22 p < 0.001
L cingulate (posterior)* 12 44 28 3.53 8 44 32 2.91 0 40 44 7.64 p < 0.001
R cingulate (posterior) 8 36 28 2.94 4 40 28 3.48 4 44 32 7.63 p < 0.001
Temporal
Left STS/STG* 60 24 8 3.53 56 24 8 2.39 56 24 8 6.71 p < 0.001
Right STS/STG 48 12 8 1.85 48 12 8 1.76 48 12 8 5.28 p < 0.005
Occipital
Left LOG* 36 76 8 4.05 36 76 4 4.42 36 76 4 8.12 p < 0.001
Right LOG 32 80 8 2.82 24 84 8 5.77 28 80 0 8.58 p < 0.001
Other regions
L insula/putamen 28 4 12 4.17 28 8 16 4.56 28 8 20 5.53 p < 0.001
R insula/putamen 28 8 16 2.77 24 0 16 3.96 24 8 12 5.28 p < 0.005
SFG, superior frontal gyrus; MFG, middle frontal gyrus; SPL, superior parietal lobule; IPS, intra-parietal sulcus; STS, superior temporal sulcus; STG, superior tem-
poral gyrus; LOG, lateral occipital gyrus. Coordinates: x, left/right; y, anterior/posterior; z, inferior/superior in the reference frame of the MNI brain in SPM97.
*Activity represented in Figs. 2 and 4.

articles
areas may be more involved in Targets

selective stimulus processing
and/or response mechanisms. Attend left Attend right
The activation in the inferior
SMA/cingulate
parietal lobule (specifically, IPS) in SMA/cingulate
response to the cues, but not targets,
suggests that this region of parietal R preCG
cortex is involved in attentional-con- L preCG
L preCG R preCG
trol processes. This finding supports
R postCG
prior evidence from blocked-design L postCG R postCG
imaging studies, which inferred the L postCG R SPL
L SPL
role of inferior parietal cortex based
L SPL
on activity averaged across blocks of R SPL
trials8,11,24,25, as well as work in ani-
mals26, but provides the first direct L postCG
L preCG L postCG L preCG
evidence that these regions of the
inferior parietal lobule are specifi-
L SPL L SPL
cally linked to attentional control in
humans. The present findings in

inferior parietal cortex suggest a role
for this brain region in attentional
control mechanisms, which may L CUN L CUN
include shifting attention 7,24, or L IFG L IFG
working memory processes engaged
to support task performance27.
In contrast, regions in superior
parietal cortex, also proposed to be SMA/cingulate SMA/cingulate
involved in attentional orienting7,
were not only activated by the cue
stimuli, but were also (and to a
greater extent) activated in response
to the target stimuli. This finding
agrees with neuroimaging studies L LG/FG L LG/FG
reporting that regions of SPL are
active both during shifts of attention
in the visual periphery and when
attention is focused on target stimuli
Fig. 3. Activity related to target processing. Brain regions significantly activated in response to the target
but no attentional shifts are stimuli, overlaid onto a 3D rendering of a brain, as in Fig. 2. Left column, activations to target stimuli when
28
required . These findings fit well attention was focused on the left visual field location. Right column, shows activations to target stimuli
with neurological lesion studies when attention was focused on the right visual field location. The labels indicate the brain regions referred
finding that inferior regions of pari- to in Table 2. The Z-values and stereotactic coordinates for the regional maxima are listed in Table 2. SMA,
etal cortex (extending into the tem- supplementary motor area; preCG, pre-central gyrus; postCG, post-central gyrus; IFG, inferior frontal gyrus;
poroparietal junction) are more MFG, middle frontal gyrus; SPL, superior parietal lobule; CUN, cuneus; LG, lingual gyrus; FG, fusiform gyrus.
critical than superior parietal
regions for normal performance in
spatial cueing tasks29.
Neuroimaging studies in humans reveal activations in the pos- the cortex in the vicinity of the FEF was activated by the cues, but
terior STS during the performance of attentional tasks, suggest- not by the target stimuli, suggesting its participation in attentional-
ing that the STS may be involved in attentional processing in control processes in the absence of eye movements. Given that eye
humans11,24, a role supported by induction of attentional neglect movements needed to be suppressed to both cue and target pre-
in lesions of the STS in monkeys30. However, it has remained sentations, the present pattern cannot be attributed solely to sup-
unclear whether the STS is involved in attentional control cir- pression of oculomotor responses33.
cuitry, or instead is involved in attention-related processing of Regions of the dorsolateral prefrontal cortex (superior frontal
relevant target stimuli. We demonstrated that a region of the STS gyrus) that were anterior to the FEF activations were also acti-
anterior to the temporal-parietal junction was selectively activated to the cues but not the targets. These regions of frontal cor-
vated by the attention-directing cues, suggesting a role in atten- tex have been associated with working memory processes in
tional control circuitry. studies of spatial working memory3438. In the present task, such
The activations to the attention-directing cues in the vicinity of processes may have activated these regions as subjects encoded
the frontal eye field in the present study suggest a role for this brain the to-be-attended location in working memory.
region in attentional control. The FEF is proposed to be involved In contrast, although ventrolateral prefrontal regions are
in attentional processes, specifically as they relate to overt eye move- sometimes considered to be involved in retention of information
ments5,31. Imaging studies, however, suggest that FEF may be in working memory34,35, in the present study, the strongest ven-
involved in covert shifts of attention as well24,32. In the present study, trolateral prefrontal activations were not observed in response to

articles
the cue stimuli when one might expect working memory process- coding to-be-attended locations were differentially activated in
es to be engaged. Rather, these cortical regions were activated response to the instruction to orient attention by the cue. Single-
upon presentation of the target stimuli. This finding may be con- unit studies in non-human primates show an increase in back-
sistent with a role in inhibitory filtering of information from the ground firing rates of neurons coding an attended region of space
ignored hemifield when the bilateral targets appeared3638. in the absence of stimuli45, and proposals based on neuroimag-
Previously, the supplementary motor area and the anterior ing data suggest that increased background firing rates may medi-
cingulate cortex were hypothesized to be part of attentional con- ate attentional facilitation in visual cortex46,47. The present data
trol systems2,8. The present finding that these areas were activat- show a precise spatial correspondence between areas of visual cor-
ed in response to the target stimuli but not the attention-directing tex activated by top-down processes and those showing subse-
cues, however, suggests a role in either the selective analysis of quent selective modulations of target processing. Moreover, these
target features or decisional processes engaged once the relevant priming effects in visual cortex were shown to follow the retino-
stimulus is presented, such as selecting appropriate motor actions topic mapping of the visual fields onto visual cortex as attention
based on target features39,40. was shifted from left to right locations by the cue instructions.
As predicted by the previous literature, target-evoked activity In summary, a top-down attentional control system was isolat-
in posterior lingual and fusiform regions was enhanced in the ed using event-related fMRI methods. This network included the
hemisphere contralateral to the direction of attention. This sup- superior frontal cortex, inferior parietal cortex, superior temporal
ports the idea that spatial attention-related increases in regional cortex and portions of the posterior cingulate cortex and insula.
blood flow in extrastriate cortex are related to enhanced process- The present evidence indicates that this top-down control system
ing of target stimuli falling within the attentional spotlight, as sug- modulated activity in extrastriate cortex as a function of where spa-
gested by prior ERP4,41,42 and functional imaging studies 9,10,1619 tial attention was directed in the visual field. These spatially selective
in humans and single-unit studies in monkeys3,43. In addition, a effects in extrastriate cortex, observed before the presentation of
more dorsal region of visual cortex was also enhanced by the direc- the target stimuli, resulted in modulations in the activity evoked
tion of attention. This region may be area V3a, which has an upper by subsequent target stimuli in precisely corresponding early cortical
visual field representation in dorsal visual cortex44 and which is areas. Hence, it seems that effects of spatial attention in visual cor-
activated in response to flashing checkerboard patterns18. tex before target onset may reflect changes in sensory-neural
Priming of visual neurons by top-down attentional control excitability as a function of top-down control. These findings pro-
signals is suggested by the finding that regions of visual cortex vide direct evidence for clear distinctions between the neural mech-
Table 2. Eventrelated activations to target stimuli and statistical contrasts.

Attended left target Attended right target Targets > cues
Coordinates Z-score Coordinates Z-score Coordinates Z-score p
x y z x y z x y z (corrected)
Frontal
SMA/cingulate** 4 8 52 7.04 4 8 52 6.10 0 4 52 8.67 p < 0.001
Left preCG** 40 16 56 7.26 36 16 56 7.43 36 16 56 8.51 p < 0.001
Right preCG** 32 16 52 5.64 36 20 56 4.00 36 16 60 8.34 p < 0.001
40 4 52 6.37 36 4 52 4.04 32 4 60 7.11 p < 0.001
Left IFG** 56 12 32 5.01 52 12 32 3.97 56 8 32 6.49 p < 0.001
Left insula/IFG** 36 28 4 5.06 36 28 4 4.39 36 24 4 7.30 p < 0.001
Right IFG/MFG 44 12 24 4.82 44 12 24 4.73 52 4 40 5.28 p < 0.005
Right insula/IFG 32 24 4 4.61 32 24 4 4.87 32 28 0 7.64 p < 0.001
Parietal
Left postCG** 36 36 56 4.53 36 40 56 5.85 28 32 64 6.03 p < 0.001
Right post CG** 28 44 56 6.54 24 48 56 4.52 28 48 56 7.29 p < 0.001
Left SPL** 24 64 48 3.76 20 68 44 4.16 16 72 48 5.37 p < 0.001
Right SPL** 24 60 60 4.87 20 60 60 4.11 12 68 52 5.00 p < 0.01
Occipital
Left CUN** 28 68 32 4.24 28 68 24 7.48 28 72 28 7.77 p < 0.001
Left LG/FG** 12 64 8 7.42 12 68 0 8.13 8 64 4 9.31 p < 0.001
24 48 0 4.40 24 44 4 5.98 24 44 8 5.19 p < 0.005
28 60 4 3.76 28 60 4 7.70 28 60 4 8.36* p < 0.001
Right CUN 24 76 28 7.81 24 76 28 3.86 24 76 28 7.8 p < 0.001
36 68 16 6.29 36 64 12 3.03 36 68 20 6.07 p < 0.001
Right LG/FG 12 60 0 7.75 8 60 8 6.89 4 64 8 9.04 p < 0.001
20 56 4 7.45 24 64 0 2.05 20 56 4 7.78* p < 0.001
28 48 8 4.03 24 40 0 2.29 24 48 16 7.86 p < 0.001
SMA, supplementary motor area; preCG, pre-central gyrus; postCG, post-central gyrus; IFG, inferior frontal gyrus; MFG, middle frontal gyrus; SPL, superior
parietal lobule; CUN, cuneus; LG, lingual gyrus; FG, fusiform gyrus. *Not a regional maximum; selected from regional maximum of target-related activations.
**Activations shown in Figs. 3 and 4

articles
checks) that reversed at 4 Hz over a

Cues > targets Targets > cues 750-ms duration (so that each stimulus
comprised 3 checkerboards lasting 250 ms
each). A random one-half of target stim-
SMA/cingulate
L SFG uli contained between 3 and 9 gray checks
(randomly located) in place of white
L MFG R MFG checks during one phase of the reversals
(the second 250-ms period). Subjects were
L preCG R preCG
required to discriminate the target stimuli
L postCG R postCG at the cued location only, pressing a but-
L IPS ton with one hand if gray checks were pre-
L SPL R SPL
R IPS sent, or pressing a button with the other
hand if there were no gray checks (hand
of response was counter-balanced across
subjects). On 17% of trials, the target stim-
L MFG L postCG uli occurred 1000 ms after the offset of the
L preCG cue. On the remaining 83% of trials, tar-
L SFG L IPS gets occurred 8160 ms following the cue.
L SPL
In order to distinguish the hemodynamic
response to the cues from that to the tar-
gets, only the longer ISI data were analyzed

and reported. The interval from the offset
of the targets to the onset of the next trial
L LOG L CUN cue was 7910 ms. Stimuli were delivered
L STS/STG L IFG through an MRI-compatible goggle sys-
tem (Resonance Technology, Northridge,
California). Subjects were trained before
the MRI session to ensure that the task was
L SFG clearly understood and that subjects were
L Post. SMA/cingulate
cingulate able to perform the task without eye
movements. In the scanner, they were
instructed to covertly orient attention at
the moment the cue was presented and to
diligently maintain fixation throughout all
runs; this was verified in control studies in
L LG/FG two subjects using signal-averaged elec-
trooculography outside the scanner. Speed
of response was not emphasized, in order
Fig. 4. Significant differences between cue and target processing. Areas significantly more active for cues to minimize head-motion artifacts that
than targets (left column) and those active more for targets than for cues (right column) projected on a 3D- might result from requiring a speeded
rendered brain. Labels indicate the brain regions referred to in Tables 1 and 2. The Z-values and stereotac- response. Head motion was minimized by
tic coordinates for the regional maxima are listed in Tables 1 and 2. SFG, superior frontal gyrus; MFG, middle using a bite bar attached to the head coil.
frontal gyrus; SPL, superior parietal lobule; IPS, intra-parietal sulcus; STS, superior temporal sulcus; STG, superior Six healthy adults (ages 2234; 4 female)
temporal gyrus; LOG, lateral occipital gyrus. SMA, supplementary motor area; preCG, pre-central gyrus; postCG, participated and gave written informed
post-central gyrus; IFG, inferior frontal gyrus; CUN, cuneus; LG, lingual gyrus; FG, fusiform gyrus. consent before all sessions. All procedures
were approved by the UC Davis human
subjects review committee. Each subject
performed 5 runs of the task (396 s per
run) while MRI data was acquired.
anisms of attentional control and the resulting modulation of sen- Stimuli and parameters: stimulation of meridia and control sessions. In
sory signals. The challenge for the future is to relate activity in each two of the subjects, visual field mapping and cue control sessions were per-
of these identified brain regions to the specific neurocomputation- formed. Meridian stimuli for visual area mapping were 9.0 2.7 checker-
board patterns aligned on the vertical or horizontal meridia and composed
al operations they subserve during visual spatial attention.
of alternating 0.9 square black and white checks, reversing at 15 Hz. Stim-
uli were presented in alternating 22-s blocks of right versus left meridia
METHODS and in a separate run, upper versus lower meridia. The alternating pattern
Stimuli and task parameters: cueing study. Each trial began with a cue of vertical and horizontal meridia activations in visual cortex were used to
that consisted of two overlapping isoluminant arrows presented at fixa- identify the borders between V1, V2, VP and V4 (refs. 44, 47). For the cue
tion (one blue and one yellow, pointing in opposite directions, each 1.9 sensory-control session, the cues used in the main experiment were flashed
tall 0.96 wide; duration, 500 ms). To eliminate differential activations every 750 ms (duration, 500 ms) for 22 s alternating with 22-s no-stimu-
based on the physical differences in right versus left cues, half the subjects lus blocks. A synthetic hemodynamic response function was convolved
were instructed to orient attention based on the direction of the blue with the box-car function that represented the periodic alternation in the
arrow, and the other half were required to use the yellow arrow. The rel- experimental conditions to yield the model response function that includ-
evant arrow pointed randomly to the left or right. Target locations were ed a temporal derivative. Low-frequency drifts were modeled and filtered
demarcated by white outline rectangles (8.1 wide 7.2 tall, centered 5.6 with a high-pass filter cutoff of 88 s, and the data were temporally smoothed
above the horizontal meridian and 7.7 lateral to the vertical meridian) by convolving the time series with a 4-s full width at half maximum
that remained on the screen throughout all runs. (FWHM) Gaussian kernel. In another cue control, the cue and target stim-
Target stimuli appeared within the white outline rectangles and consisted uli were identical to that used in the main attention study except that the
of black and white checkerboards (8.1 7.2, composed of 0.9 square cues were not instructive as to where to orient attention. (Subjects had to

articles
Fig. 5. Selective processing of target stimuli and retinotopy (single sub- Selective attention (targets)
ject). (a) A T2-weighted coronal slice (y = 72) showing the activations
due to selective visual attention on target processing for one subject.
(single subject)
The left column shows the regions showing greater activity when atten- a
tion was focused on the left visual field. The right column shows those Att. left > att. right Att. right > att. left
areas showing greater activity when attention was focused on the right. L R 6 5 L R
Attention-related activations in the contralateral fusiform and lingual
4
gyri are outlined in black (coordinates of left hemisphere maximum, 4
32, 60, 8; Z = 5.26; right hemisphere maximum, 20, 68, 8; Z = 5.39). 3
A more dorsal region in the cuneus (not outlined in black) also showed 2
2
enhanced activity contralateral to the direction of attention (left hemi-
1
sphere maximum = 32, 64, 28, Z = 4.02; right hemisphere
maximum = 24, 72, 28, Z = 6.27). (b) Left panel, ventral visual areas for 0 0
the same subject shown at top, determined using the meridia scans Z value
(see Methods) and traced onto contrast-inverted, T2-weighted MRI scans
(y = 72). Right panel, effect of attention to target stimuli on responses in
lingual and fusiform gyri (shown as thick black outlines) overlaid onto the b
subjects retinotopically mapped visual areas. Attention-related target pro- Visual areas Selective attention
cessing extends from V2 through V4 in the ventral visual-processing stream. L R L R
V1
V2
detect gray checks in both hemifields). During these sessions, target stim- VP
uli (250 ms) were presented on only 22% of trials with an ISI of 300 ms,
whereas the other 78% of trials (catch trials) consisted only of the cue stim-
V4
uli; only these catch trials were analyzed, using event-related methods (see
below). Attend Attend
right left
Scanning procedures. Functional images were acquired with a Signa resolution proton density and T2-weighted fast spin echo (FSE) images were
Advantage 1.5 Tesla whole-body MRI system (General Electric, Wauke- also obtained in the coronal plane during the same scanning session. The
sha, Wisconsin), with an elliptical end-capped quadrature radio fre- anatomical scans were acquired with the following parameters: TR = 4.2 s;
quency and local gradient head coil (Medical Advances, Milwaukee, TE = 17 and 136 ms; echo train = 8; FOV = 22 cm, matrix = 512 256 (inter-
Wisconsin). Images were acquired using T2*-weighted gradient-recalled polated to 512 512 for display), slice thickness of 6 mm, with an interslice gap
echo, echo planar imaging (EPI) in the coronal plane with a repetition of 2 mm. In each EPI acquisition, 144 images were acquired at each of the
time (TR) of 2.75 seconds, an effective echo time (TE) of 40 ms and a 22 slice locations, for a total acquisition time of 396 s (6 min, 36 s).
flip angle (FA) of 90. Twenty-two interleaved slices were collected with
a 22-cm field of view (FOV), 64 64 matrix, slice thickness of 6 mm and Image processing. The first 16 images at each slice location (initial 44 s of
an interslice gap of 2 mm, yielding a voxel size of 3.43 3.43 6.00 mm3. acquisition) were discarded from the analysis. The remaining 128 images at
Fourier image reconstruction included N/2 ghost correction using image each slice were used for the time-series analyses. Images were realigned and
phase correction48. Neural activation was detected via the blood oxygenation corrected for movement artifacts. The anatomical T2 images were coregis-
level dependent (BOLD) contrast mechanism15, which provided differences tered with the proton density images and the functional images for each
in signal intensity based upon differences in tissue T2* and perfusion. High- subject. Each subjects T2 (and coregistered proton density scans) were then
Fig. 6. Selective attention-related activations to tar-

a Targets
gets and cues (group data). Results of group analysis
showing regions with differential activity due to the
Att. left > att. right (n = 6 subs)
Att. right > att. left direction of attention, overlaid onto averaged (6 sub-
ject) proton density MRI scans (slice at y = 64).
6 (a) Target processing was enhanced in regions of visual
5
cortex contralateral to the attended hemifield. These
4 4 significant (all p < 0.001 corrected) attention-related
Z value
Cuneus
Z value
3 activations were observed in the lingual and fusiform

2 Lingual Fusiform 2 gyri in the right (12, 64, 4; Z = 5.86) and left
1 (28, 60, 8; Z = 6.15) hemispheres, and also in dor-
0 L R L R 0 sal regions of the cuneus in the right (24, 72, 28;
Z = 6.04) and left (28, 68, 24; Z = 5.30) hemispheres.
(b) Contralateral to the direction of attention, multi-
b Cues ple regions of extrastriate cortex showed responses
Cued left > cued right (n = 6 subs)
Cued right > cued left to the cue stimuli that differed based on the direction
of attention, including the lingual and fusiform gyri
4 4 (12, 64, 4, Z = 4.51; 20, 64, 4, Z = 4.25) as well as
a more dorsal region of the cuneus (24,68, 28,
3 3
Z = 3.87; 28, 68, 24, Z = 4.61). The maxima of these
Z value
Z value
Cuneus
2 2 activations were 01 cm from those in response to
Lingual Fusiform targets and were all significant at p < 0.001, uncor-
1 1
rected. (Because regions of interest analyzed for cue
0 L R L R 0 responses were specified a priori by analyses for atten-
tion effects on target responses, we give uncorrected
p values for cue responses.)

articles
spatially normalized to a standard stereotactic space49 (using, however, the 15. Ogawa, S. et al. Intrinsic signal changes accompanying sensory stimulation:
canonical Montreal Neurological Institute template in the SPM97 package), functional brain mapping with magnetic resonance imaging. Proc. Natl.
Acad. Sci. USA 89, 59515955 (1992).
with the origin at the anterior commisure and the xy plane extending 16. Mangun, G. R., Hopfinger, J., Kussmaul, C., Fletcher, E. & Heinze, H. J.
through the anterior and posterior commisures. The functional images were Covariations in ERP and PET measures of spatial selective attention in
then spatially normalized into stereotactic space, resampled to 4 4 4 mm3 human extrastriate cortex. Hum. Brain Mapp. 5, 273279 (1997).
voxels and spatially smoothed with an 8-mm isotropic Gaussian kernel. 17. Mangun, G. R., Buonocore, M., Girelli, M. & Jha, A. ERP and fMRI measures
of visual spatial selective attention. Hum. Brain Mapp. 6, 383389 (1998).
18. Tootell, R. B. et al. The retinotopy of visual spatial attention. Neuron21, 14091422 (1998).
Event-related analyses. The responses of interest in the event-related design 19. Brefczynski, J. A. & DeYoe, E. A. A physiological correlate of the spotlight of
were modeled by convolving a vector containing the onset times of the visual attention. Nat. Neurosci. 2, 370374 (1999).
cues and targets with a synthetic hemodynamic response function com- 20. Engel, S. A., Glover, G. H. & Wandell, B. A. Retinotopic organization in
posed of the sum of two gamma functions and its temporal derivative human visual cortex and the spatial precision of functional MRI. Cereb.
Cortex 7, 181192 (1997).
(accounting for the hemodynamic response and subsequent undershoot 21. Sereno, M. I. et al. Borders of multiple visual areas in humans revealed by
and for slight variations in timing)50. As a result, the present analysis functional magnetic resonance imaging. Science 268, 889893 (1995).
emphasizes transient hemodynamic activity temporally adjacent to the cue 22. Woldorff, M. et al. Retinotopic organization of early visual spatial attention
and target events. Multiple linear regression (using a least-squares approach effects as revealed by PET and ERPs. Hum. Brain Mapp. 5, 280286 (1997).
to estimate the parameters), as implemented in SPM97, was used to model 23. Treue, S. & Maunsell, J. H. R. Attentional modulation of visual motion
processing in cortical areas MT and MST. Nature 382, 539541 (1996).
the effects of interest and other confounding factors (such as session effects 24. Corbetta, M. et al. A common network of functional areas for attention and
and low-frequency drifts of signal intensity) at every voxel simultaneous- eye movements. Neuron, 21, 761773 (1998).
ly. Data from different sessions were proportionally scaled to a grand mean 25. Coull, J. T. & Nobre, A. C. Where and when to pay attention: The neural
of 100 arbitrary units to account for overall differences in the intensity of systems for directing attention to spatial locations and to time intervals as
whole brain volumes across the time series. Temporal smoothing was revealed by both PET and fMRI. J. Neurosci. 18, 74267435 (1998).
26. Bushnell, M. C., Goldberg, M. E. & Robinson, D. L. Behavioral enhancement of
accomplished by convolving the data with a 4-s FWHM Gaussian filter (to visual responses in monkey cerebral cortex. I. Modulation in posterior parietal
account for temporal autocorrelations in the time series). An omnibus cortex related to selective visual attention. J. Neurophysiol. 46, 755772 (1981).
voxel-wise analysis provided an SPM(F), and those voxels with significant 27. LaBar, K. S., Gitelman, D. R., Parrish, T. B. & Mesulam, M. -M. Neuroanatomic
F-ratios (p < 0.001 uncorrected) were retained for further specific statis- overlap of working memory and spatial attention networks: A functional MRI
comparison within subjects. Neuroimage 10, 695704 (1999).
tical tests. Specific hypotheses were then tested with t-tests at those voxels, 28. Corbetta, M. Frontoparietal cortical networks for directing attention and the
and the resulting SPM{t} was transformed to the unit normal distribution eye to visual locations: Identical, independent, or overlapping neural
resulting in the Z-scores and SPM(Z) maps in the figures and tables. systems? Proc. Natl. Acad. Sci. USA 95, 831838 (1998).
29. Friedrich, F. J., Egly, R., Rafal, R. & Beck, D. Spatial attention deficits in
humans: a comparison of superior parietal and temporal-parietal junction
ACKNOWLEDGEMENTS lesions. Neuropsychology 12, 193207 (1998).
This research was supported by grants from the National Institute of Mental 30. Watson, R. T., Valenstein, E., Day, A. & Heilman, K. M. Posterior neocortical
systems subserving awareness and neglect. Arch. Neurol. 51, 10141021 (1994).
Health (MH55714 and MH57138) and Human Frontier Science Program to 31. Henik, A., Rafal, R. & Rhodes, D. Endogenously generated and visually guided saccades
G.R.M., and a National Science Foundation fellowship to J.B.H. We thank Neva after lesions of the human frontal eye fields. J. Cogn. Neurosci. 6, 400411 (1984).
Corrigan for assistance in collecting the MRI data, Jeff Maxwell for assistance in 32. Rosen, A. C. et al. Neural basis of endogenous and exogenous spatial
orienting: a functional MRI study. J. Cogn. Neurosci. 11, 135152 (1999).
collecting the EOG data, Karl Friston and Christian Buechel for advice 33. Petit, L. et al. PET study of the human foveal fixation system. Hum. Brain.
regarding the event-related fMRI analyses and Kevin LaBar, Kevin Wilson, Mapp. 8, 2843 (1999).
Barry Geisbrecht, Marty Woldorff, Daniel Weissman, Steven Hillyard and 34. Jonides, J. et al. Spatial working memory in humans as revealed by PET.
Nature 363, 623625 (1993).
Tamara Swaab for comments on the manuscript. 35. Kojima, S. & Goldman-Rakic, P. S. Delay-related activity of prefrontal neurons
in rhesus monkeys performing delayed response. Brain Res. 248, 4350 (1982).
36. DEsposito, M., Ballard, D., Aguirre, G. K. & Zarahn, E. Human prefrontal
RECEIVED 27 SEPTEMBER 1999; ACCEPTED 12 JANUARY 2000
cortex is not specific for working memory: a functional MRI study.
Neuroimage 8, 274282 (1998).
1. Pashler, H. The Psychology of Attention (MIT Press, Cambridge, 37. Jonides, J., Smith, E. E., Marshuetz, C., Koeppe, R. A. & Reuter-Lorenz, P. A.
Massachusetts, 1999). Inhibition in verbal working memory revealed by brain activation. Proc. Natl.
2. Posner, M. I. & Petersen, S. E. The attention system of the human brain. Acad. Sci. USA 95, 84108413 (1998).
Annu. Rev. Neurosci. 13, 2542 (1990). 38. Smith, E. E. & Jonides, J. Neuroimaging analyses of human working memory.
3. Moran, J. & Desimone, R. Selective attention gates visual processing in the Proc. Natl. Acad. Sci. USA 95, 1206112068 (1998).
extrastriate cortex. Science 229, 782784 (1985). 39. Corbetta, M. in The Attentive Brain (ed. Parasuraman, R.) 95122 (MIT
4. Mangun, G. R. & Hillyard, S. A. Modulation of sensory-evoked brain potentials Press, Cambridge, Massachusetts, 1998).
provide evidence for changes in perceptual processing during visual-spatial 40. Posner, M. I. & DiGirolamo, G. J. in The Attentive Brain (ed. Parasuraman,
priming. J. Exp. Psychol. Hum. Percept. Perform. 17, 10571074 (1991). R.) 401423 (MIT Press, Cambridge, Massachusetts, 1998).
5. Mesulam, M.-M. A cortical network for directed attention and unilateral 41. Luck, S. J. et al. Effects of spatial cuing on luminance detectability:
neglect. Ann. Neurol. 10, 309325 (1981). Psychophysical and electrophysiological evidence for early selection. J. Exp.
6. Posner, M. I., Walker, J. A., Friedrich, F. A. & Rafal, R. D. Effects of parietal Psychol. Hum. Percept. Perform. 20, 887904 (1994).
injury on covert orienting of attention. J. Neurosci. 4, 18631874 (1984). 42. Van Voorhis, S. T. & Hillyard, S. A. Visual evoked potentials and selective
7. Corbetta, M., Miezin, F., Shulman, G. & Petersen, S. A PET study of attention to points in space. Percept. Psychophys. 22, 5462 (1977).
visuospatial attention. J. Neurosci. 13, 12021226 (1993). 43. McAdams, C. J. & Maunsell, J. H. R. Effects of attention on orientation-tuning functions
8. Gitelman, D. R. et al. A large-scale distributed network for covert spatial of single neurons in macaque cortical area V4. J. Neurosci. 19, 431441 (1999).
attention. Brain 122, 10931106 (1999). 44. Tootell, R. et al. Functional analysis of human MT and related visual cortical
9. Heinze, H. J. et al. Combined spatial and temporal imaging of spatial selective areas using magnetic resonance imaging. J. Neurosci. 15, 32153230 (1995).
attention in humans. Nature 392, 543546 (1994). 45. Luck, S. J., Chelazzi, L. Hillyard, S. A. & Desimone, R. Neural mechanisms of
10. Martinez, A. et al. Involvement of striate and extrastriate visual cortical areas spatial selective attention in areas V1, V2 and V4 of macaque visual cortex.
in spatial attention. Nat. Neurosci. 2, 364369 (1999). J. Neurophysiol. 77, 2442 (1997).
11. Nobre, A. C. et al. Functional localization of the system for visuospatial 46. Chawla, D., Rees, G. & Friston, K. J. The physiological basis of attentional
attention using positron emission tomography. Brain 120, 515533 (1997). modulation in extrastriate visual areas. Nat. Neurosci. 7, 671676 (1999).
12. Harter, M. R., Miller, S. L., Price, N. J., LaLonde, M. E. & Keyes, A. L. Neural 47. Kastner, S., Pinsk, M. A., De Weerd, P., Desimone, R. & Ungerleider, L. G.
processes involved in directing attention. J. Cogn. Neurosci. 1, 223237 Increased activity in human cerebral cortex during directed attention in the
(1989). absence of visual stimulation. Neuron 22, 751761 (1999).
13. Buckner, R. et al. Detection of cortical activation during averaged single trials 48. Buonocore, M. H. & Gao, L. Ghost artifact reduction for echo planar imaging
of a cognitive task using functional magnetic resonance imaging. Proc. Natl. using image phase correction. Magn. Reson. Med. 38, 89100 (1997).
Acad. Sci. USA 93, 1430214303 (1996). 49. Talairach, J. & Tournoux, P. Co-Planar Stereotaxic Atlas of the Human Brain
14. McCarthy, G., Luby, M., Gore, J. & Goldman-Rakic, P. Infrequent events (Thieme, New York, 1988).
transiently activate human prefrontal and parietal cortex as measured by 50. Friston, K. J. et al. Event-related fMRI: Characterizing differential responses.
functional MRI. J. Neurophysiol. 77, 16301634 (1997). Neuroimage 7, 3040 (1998).

articles
Voluntary orienting is dissociated

from target detection in human
posterior parietal cortex
Maurizio Corbetta1,2,3, J. Michelle Kincade1,4, John M. Ollinger2, Marc P. McAvoy2 and Gordon L.
Shulman1
1 Department of Neurology and Neurological Surgery, Washington University School of Medicine, 4525 Scott Avenue, St. Louis, Missouri 63110, USA
2 Mallinckrodt Institute of Radiology, Washington University School of Medicine, 4525 Scott Avenue, St. Louis, Missouri 63110, USA
3 Department of Anatomy and Neurobiology, Washington University School of Medicine, 4525 Scott Avenue, St. Louis, Missouri 63110, USA
4 Department of Psychology, Washington University School of Medicine, 4525 Scott Avenue, St. Louis, Missouri 63110, USA
Correspondence should be addressed to M.C. (mau@npg.wustl.edu)
Human ability to attend to visual stimuli based on their spatial locations requires the parietal cortex.
One hypothesis maintains that parietal cortex controls the voluntary orienting of attention toward a
location of interest. Another hypothesis emphasizes its role in reorienting attention toward visual
targets appearing at unattended locations. Here, using event-related functional magnetic resonance
(ER-fMRI), we show that distinct parietal regions mediated these different attentional processes.
Cortical activation occurred primarily in the intraparietal sulcus when a location was attended before
visual-target presentation, but in the right temporoparietal junction when the target was detected,
particularly at an unattended location.
Acute structural damage to the right temporoparietal cortical involved in voluntary orienting, then it should be activated when
junction (TPJ; inferior parietal lobule and superior temporal an observer attends to a location before presentation/detection
gyrus) in humans produces a complex clinical syndrome charac- of a visual target. If TPJ is necessary for reorienting to a visual
terized by the inability to attend and respond to objects positioned target, then its activation should follow the presentation/detec-
in the left visual field (unilateral visual neglect)13. Some symp- tion of the target, particularly when it is presented at an unat-
toms of neglect may reflect a deficit in reorienting attention toward tended location. We tested these predictions using ER-fMRI and
new stimuli in the visual field opposite to the lesion (contrale- an ANOVA-based procedure17. This method has two important
sional field)47. Patients with parietal lesions can detect visual stim- characteristics: it can separate the responses to events presented
uli in the contralesional field when correctly cued to their within the same cognitive trial, and it is sensitive to differences
locations, but they are slow or fail to detect the same stimuli when in both magnitude and timing of responses. Unlike other pub-
attending to other locations4. This deficit is more severe after right lished ER-fMRI methods1821, this method makes no assump-
than left parietal lesions5, and is localized to the TPJ7. These find- tions about the shape of the underlying response function.
ings suggest that the posterior parietal cortex near TPJ may be
critical for reorienting the focus of attention toward visual stimuli RESULTS
appearing at unattended locations (reorienting hypothesis). Normal observers were given a cue indicating the most likely
Other data suggest that posterior parietal cortex near/along location of a subsequent target stimulus they were required
the intraparietal sulcus (IPs), which separates the superior from to detect, according to a protocol modified from a published
the inferior parietal lobule, is involved in voluntarily directing procedure4. The stimulus display consisted of a central fixa-
attention to a spatial location (voluntary orienting hypothesis). tion cross flanked on either side by square boxes. The length of
Neurons in the IPs increase firing rate when a monkey attends each arm of the central fixation cross subtended 16 minutes
to a location while preparing a response811. Human functional of visual angle. The boxes (size, 1) were placed at 3.3 of visu-
brain imaging shows activations in the IPs (and superior pari- al angle to either side of the fixation spot. Accurate fixation
etal lobule) when observers voluntarily pay attention to and detect of the central cross-hair was emphasized throughout the
peripheral visual stimuli, with or without concurrent eye move- experiment. At the beginning of a trial, a cue arrow pointing
ments1216. It is unknown to what extent areas in human and to the left or right box was superimposed on the fixation cross.
monkey IPs are homologous. The arrow indicated the most likely location of a subsequent
These complementary functional anatomical theories (reori- target stimulus, and leftward or rightward arrows were equal-
enting to targets in the TPJ and voluntary orienting in the IPs) ly probable. The cue arrow remained on the screen for one
make specific predictions about which regions should be acti- MR frame (2360 ms; cue period).
vated while attending to a spatial location and, subsequently, The sequence of events following presentation of the cue
when detecting a visual target there. If IPs is preferentially arrow depended on the type of trial. On a cue trial (20% of the

articles
Fig. 1. Display, trial types and MR design. Each trial

lasted between 4 and 7 MR frames, and each MR
frame was 2.36 s long. MR frames are indicated by
elongated rectangles below displays. In a cue trial, a
cue arrow was presented for 1 MR frame (cue
period) at fixation (black rectangle) followed by an
intertrial interval (ITI) period signaled by a change in
the color of the fixation point (from green to red).
The ITI period lasted for 2, 3 or 4 MR frame duration
(white rectangles). A two-frame ITI is shown. In a
valid trial, the cue period was identical. During the
test period (2 MR frames or 4.72 s; crossed rectan-
gles), after a randomly selected time between
15003000 ms, a 100-ms target stimulus (asterisk)
was flashed in the box cued by the arrow. Subjects
indicated target detection with a key press. The ITI
period followed. Invalid trial, same as valid trials
except that the target was flashed at the uncued box
location. Noise trial, same as valid trials, except that
no target was flashed during the target period.
trials), the trial ended immediately after cue presentation. The We used this protocol during whole-brain measurements of
end of the trial was signaled by a change in the color of the fix- blood oxygenation level dependent (BOLD) responses on a
ation cross from green to red. During a cue trial, a subject shift- Siemens Vision 1.5 T magnet. The time course of the BOLD
ed attention toward the location indicated by the cue and response for each trial type and each trial period (cue period, test
maintained it there until the end of the trial. On a noise trial period/noise, test period/valid, test period/invalid) was estimat-
(20% of the trials), the cue period was followed by a test period ed in each subject using linear regression. Pixel-wise and region-
lasting 2 MR frames (4720 ms) in which no target was present- al ANOVAs were used for appropriate statistical contrasts17 (see
ed. During noise trials, the subject presumably shifted and Methods).
maintained attention on the cued location for a longer time
than during a cue trial (7080 versus 2360 ms; Fig. 1). On a valid Behavior
trial (44% of the trials), the cue period was followed by a test Reaction times for target detection were faster on valid than
period during which a target appeared at the location indicated invalid trials (380 ms versus 426 ms, F1,11 = 21.92, p = 0.0007),
by the cue. The target was a white asterisk that appeared in one indicating that subjects used the cue arrow to attend to the loca-
of the square boxes for 100 ms. On an invalid trial (16% of the tion of the target. No responses were recorded during noise and
trials), a target appeared during the test period at the uncued cue trials.
location. The cue arrow correctly predicted the target location
on 73% of the trials in which a target was presented. These four Imaging
trial types (cue, noise, valid, invalid) were randomly intermixed. During the cue period, a series of ventral and dorsal visual
Subjects were instructed to press a button as quickly as possi- regions were active; these included bilateral anterior fusiform
ble upon detection of the target and to withhold responses on (Fus; x, y, z atlas coordinates, 35, 57, 20, right; 31, 55, 16,
cue or noise trials. left), lateral occipital (LO;31, 83, 0, left; 27, 87, 0, right)22,23,
Table 1. List of parietal regions during cue and target periods (averaged over valid and invalid targets) and showing
significant validity effect.
Cue Target Valid invalid
Regions x y z Z-score x y z Z-score x y z Z-score
L pos IPS 25 67 48 7.58 25 65 48 6.94
L ant IPs 25 57 46 7.55 25 57 42 7.27
L vIPS 23 67 32 7.42
27 75 26 7.23 27 77 20 7.16
R ant IPs 27 59 52 6.75 33 51 48 7.84 39 47 48 4.09
R pos IPs 21 65 52 6.62
R vIPS 29 71 22 6.01 27 71 30 6.63
R IPL 53 45 20 8.53 53 49 30 5.12
R STG 51 55 4 5.41 57 45 12 4.44
R PC 7 73 32 7.71 7 75 34 4.28
L PC 9 71 40 7.37 5 71 34 4.27
See Figs. 2 and 3 for anatomical labels. Coordinates (x, y, z) correspond to the Talairach atlas.

articles
a b Cue period
Percent change BOLD signal

Pos IPs
Ant IPs
MT+
LO
c d MR frame number
Cue and target IPS/cue

period IPS/valid

TPJ/cue
TPJ/valid
cue ITI
MR frame number
Fig. 2. BOLD responses during cue and target periods. (a) ANOVA F map transformed to Z map for the cue period (averaged over subjects, cue
direction). Left, sagittal slice 25 mm left of midline. Right, coronal slice 45 mm posterior to center of atlas space. Yellow lines indicate corresponding
planes of section. Parietal regions with sustained responses to cues are labeled in red: posIPs, posterior intraparietal sulcus; antIPs, anterior intrapari-
etal sulcus; vIPs, ventral intraparietal sulcus. The lateral occipital region LO (labeled in white) showed a transient response to the cue. (b) BOLD time
courses in different regions (averaged over subjects, cue direction and hemispheres) during the cue period. Response typically peaked two frames
after onset of cue arrow on frame 1. ITI began on frame 4. (c) Target period (averaged over subjects, valid and invalid targets). Parietal regions with
prevalent responses to targets are shown in red label: TPJ, temporoparietal junction; Precun, precuneus. Several motor-related responses are also
shown: SMcx, sensory-motor cortex; Put, putamen; Cbl, cerebellum. (d) BOLD time courses in IPs and TPJ (averaged over subjects, cue direction,
target field and hemispheres) during cue and target (valid) periods. Target was randomly presented around frame 3. IPs/cue, IPs response during cue
period; IPs/valid, IPs response during valid-target period; TPJ/cue, TPJ response during cue period; TPJ/valid, TPJ response during valid-target period.
MT+ (45, 69, 2, left; 45, 69, 4, right) 24,25 and ventral reflect longer times required for processing cues related to ori-
(vIPs), anterior (ant IPs) and posterior intraparietal sulcus (pos enting toward and maintaining attention at the cued location.
IPs; Fig. 2a, left; see Table 1 for coordinates of parietal foci). No To further test this idea, we compared BOLD responses of these
significant activation was detected in the TPJ during the cue regions during the noise period, in which subjects maintained
period (Fig. 2a, right). The IPs activity did not reach the sur- attention at a peripheral location for 4.72 seconds after the offset
face of the inferior parietal lobule, but spread into the superior of the cue arrow. Across all regions active during the noise peri-
parietal lobule. The vIPs region was located at the junction of od, only antIPs and vIPs showed sustained activity (ANOVA
dorsal occipital and parietal cortex, just dorsal and anterior to regions frame, F77,924 = 7.80, p = 0.0001), with responses of the
the V3A representation26. left hemisphere more sustained than those of the right hemi-
The time course of the BOLD response during the cue peri- sphere (Ant IPs, F 7,84 = 3.76 p = 0.0014; vIPs, F 7,84 = 5.40,
od was more sustained within the intraparietal sulcus (posIPs, p = 0.0001). Hence, after the presentation of the cue arrow, some
antIPs, vIPs) than in occipital regions (Fus, LO, MT+; ANOVA IPs regions (antIPs, vIPs) showed a sustained BOLD response
regions frame, F56,672 = 5.18, p = 0.0001). We compared a tran- that was maintained during the noise period, during which sub-
sient time course in two occipital regions (LO, MT+) with more jects attended to the cued location for almost five seconds while
sustained time courses in antIPs and posIPs (Fig. 2b). The dif- waiting for the target stimulus.
ference in response duration cannot be explained by a differ- During the target period, many visual and motor regions
ence in the peak magnitude. LO and posIPs, for example, were active for both valid and invalid targets (Fig. 2c). All occip-
showed similar peak magnitudes on frame 3, but different ital regions that were active during the cue period also respond-
response duration (ANOVA regions frames 3 and 4 only, ed to target presentation, and these responses were significantly
F1,12 = 24.3, p = 0.0003). Similarly, antIPs and MT+ had similar stronger for targets in the contralateral visual field. In parietal
magnitudes, but the response was more sustained in antIPs cortex, significant responses were recorded in antIPs, posIPs,
(ANOVA regions frames 3 and 4 only, F1,12 = 7.22, p = 0.02). vIPs, precuneus (Precun) and the TPJ, where the activation was
Whereas transient time courses in occipital regions probably much stronger in the right than the left hemisphere (Fig. 2c;
reflect visual processes related to the presentation of the foveal Table 1). BOLD responses in these parietal regions during the
cue, the more sustained time courses in intraparietal cortex may cue and valid-target periods were compared by ANOVAs. Data

articles
L valid
a b c L valid
L invalid L invalid

R valid R valid
R invalid R invalid
MR frame number MR frame number
Fig. 3. BOLD responses for valid and invalid targets. (a) Coronal section through TPJ cortex (~47 mm posterior). ANOVA (validity frame) F map
transformed to Z map. Voxels that show significant validity effect (different BOLD responses for valid and invalid targets) independent of the visual
field of the target. BOLD time courses (averaged over subjects) were estimated in right inferior parietal lobule (R IPL; b) and right superior temporal
gyrus (R STG; c) during the target period for valid and invalid targets in left and right visual field.
for antIPs and posIPs were collapsed, as the borders between the location was voluntarily attended, independent of processes relat-
two regions could not be readily identified. Responses of IPs ed to target detection (for instance, visual responses and their
(Ant + Pos) and vIPs regions were stronger during the cue peri- attentional modulation or motor responses). In contrast, the right
od, whereas those of the TPJ and precuneus region were stronger TPJ responded to target presentation more strongly when the
during the target period (F21,252 = 4.29 p = 0.0001; Fig. 2d). These target occurred at an unattended location.
findings indicate that the voluntary orienting and maintenance
of attention to a location primarily recruited the cortex within Voluntary orienting of attention
the IPs. In contrast, target detection recruited the TPJ (and pre- Two findings support a role for the IPs in the voluntary orient-
cuneus), although a significant target response was also evident ing and maintenance of attention to a target location. First, the
in IPs (Fig. 2d). presentation of a cue arrow indicating the most likely location
To test whether TPJ was preferentially involved in reorient- of a subsequent visual target triggered transient responses in
ing attention toward novel unattended stimuli, the time cours- occipital cortex, but more sustained responses in IPs. Transient
es of the BOLD responses during the target periods for valid responses in occipital cortex may reflect the encoding of the cue,
and invalid trials were compared. The strongest validity effect which was probably completed within a half second27. In con-
(difference in the BOLD response for valid and invalid trials) trast, sustained parietal responses may reflect the required shift
across the whole brain was localized in the right TPJ cortex, toward and maintenance of attention at the cued location for
with separate foci in the inferior parietal lobule (IPL) and supe- the entire cue period (2360 ms). Second, when the delay after
rior temporal gyrus (STG; pixel-wise ANOVA Z-score = 5.12; the cue offset (noise period) was extended to 4.72 seconds, forc-
Fig. 3a; Table 1). A regional ANOVA confirmed that the valid- ing subjects to maintain attention at the cued location for longer,
ity effect was significant and independent of the visual field of IPs was the only brain region that showed a sustained response
the target (ANOVA frame validity, R IPL, F 7,84 = 8.13, during the delay.
p < 0.0001; R STG, F7,84 = 7.53, p < 0.0001). The response in Our results extend findings of increased activity in extras-
right IPL and STG was more sustained for valid than invalid triate, frontal and IPs cortex without sensory stimulation when
targets in each visual field (Fig. 3b and c; ANOVA, frames 5 and subjects attend to a specific object at a specific location during
6 only validity, R IPL, F 1,12 = 5.212, p = 0.041; R STG, an identification task20. Those activations are thought to reflect
F1,12 = 6.51, p < 0.025). Significant validity effects were also an attentional signal-biasing activity in visual cortex before
localized bilaterally in the precuneus and near the intersection stimulus presentation28. Here we show, in a simpler detection
of the right IPs and postcentral sulcus. This latter region did protocol, that IPs was uniquely active when attention was ori-
not overlap with the IPs regions active during the cue period ented toward and maintained at a relevant location, suggesting
(vector distance, 17 mm; Table 1). Significant validity effects that IPs is the source of spatial biases observed in extrastriate
not discussed here were also observed in other regions outside visual cortex.
of parietal cortex. Cue-related activity in IPs may underlie an attentional sig-
nal that marks a location of interest9,11 or an intentional signal
DISCUSSION that prepares a response (eye movement or hand reaching)
We tested two functional-anatomical theories about the role of toward that location 10. However, attending to direction of
posterior parietal cortex in visuospatial attention. One theory, motion also drives IPs regions before the presentation of any
supported by studies of neglect patients with parietal lesions, motion target17. Activations are found in IPs during shifts in an
proposes that the TPJ cortex is necessary for reorienting toward object feature (for instance, color or shape)29 or between per-
visual targets appearing at unattended locations. Another theory, cepts during binocular rivalry30. These results suggest that the
based on single unit and imaging data, proposes that cortex along intraparietal cortex may be involved in visual selection beyond
the IPs is involved in the voluntary orienting of attention toward the selection of locations.
a location. Our results provide direct confirmation of both views, The BOLD response in IPs was time locked to different
showing that IPs was active before target presentation when a processes in the course of a trial. Early in the trial (cue and noise

articles
period), the response was controlled by voluntary orienting critical for visual reorienting, and dissociates this region from
processes, whereas the response was controlled by detection voluntary orienting in nearby IPs.
processes later in the trial (target period; Fig. 2d). The coexis-
tence of different processes in human IPs resembled that observed METHODS
in monkey IPs, where neurons also manifest activity time locked Subjects. Thirteen subjects (6 female, 7 male; aged 1838) were recruit-
to different components of a task (for instance, visual, attentional, ed from the Washington University community following procedures
memory, oculomotor or reaching)911,31,32. Although spatial over- approved by the local human studies committee. All subjects were strong-
lap between regions of activation during different protocols (visu- ly right-handed as measured by the Edinburgh handedness inventory44
and had normal or corrected-to-normal visual acuity and no significant
al attention versus eye movements) in previous imaging studies abnormal neurological history. Informed consent was obtained from
was used to examine colocalization of processes16,33, the enhanced each subject. Before the MR session, subjects participated in one behav-
temporal resolution of new ER-fMRI procedures17,20,34,35 allows ioral session during which they were trained to perform the task while
a richer and more realistic view of the temporal dynamics of acti- maintaining central fixation. Eye movements were monitored with elec-
vation in the human brain. tro-oculography, and all subjects were able to perform the task without
breaks of fixation (resolution, 1.5). Eye movements were not recorded
Target detection during the fMRI session. Although we cannot rule out the occurrence
The right TPJ (and precuneus) was specifically engaged during of small eye movements during the fMRI session, several arguments
target detection. Unlike other parietal regions that showed both diminish the likelihood of this possibility. The visual set-up was identi-
cal to the one used in the psychophysical session, in which eye move-
cue and target responses (antIPs, vIPs), right TPJ and precuneus
ments were monitored and found negligible. The detection task was not
showed little if any response to the cue. demanding in terms of acuity. Subjects reported no problems in main-
The pattern of activation in the TPJ cortex fits two main fea- taining fixation, in agreement with many studies involving detection
tures of unilateral spatial neglect. The much stronger activation tasks27. Additionally, there was no activity in the frontal eye field during
on the right than the left TPJ agrees with the clinical2,3,36 and neu- the noise period, when the tendency to look at the peripheral box was
ropsychological findings5 that neglect is more severe following strongest. Reaction times were not collected from one subject because
right TPJ lesions. The stronger TPJ response following the pre- of equipment malfunction.
sentation of a target at an invalid location is also consistent with
clinical and experimental4,7 observations that neglect is worse for Apparatus. Stimuli were generated by an Apple Power Macintosh com-
objects presented at unattended locations. puter and projected onto a screen at the head of the bore by a Sharp LCD
projector. Subjects viewed the stimuli through a mirror attached to the
Activation of the right TPJ may underlie the process of spa- head coil. Subjects recorded behavioral responses by pressing an MRI-
tially redirecting the focus of attention toward the location of compatible fiber-optic key held in the right hand.
unattended stimuli. This reorienting was indexed by longer reac-
tion times for invalid targets than for valid targets; thus, the sen- Data analysis and fMRI scan acquisition. An asymmetric spin-echo,
sitivity of the TPJ activation to target validity implies that the echoplanar imaging sequence was used to measure blood oxygenation-
BOLD signal, despite its slow temporal resolution (seconds), can level-dependent (BOLD) contrast (TR = 2.36 s, TE = 50 ms, flip
track neuronal processes that yield behavioral differences of a angle = 90). Each scan consisted of 128 frames during which 16 con-
few tens of milliseconds (the difference in reaction time between tiguous 8 mm axial slices were acquired (3.75 3.75 mm in-plane res-
valid and invalid trials). olution). Anatomical images were acquired using a sagittal MP-RAGE
sequence (TR = 97 ms, TE = 4 ms, flip angle = 12, inversion time
Another possibility, however, is that activity in right TPJ cor-
T1 = 300 ms). Functional data were realigned within and across runs to
tex is related to spatially nonselective neural processes triggered correct for head movement and coregistered with anatomical data.
by reorienting to an unattended target. Interestingly, TPJ dam- Whole-brain normalization was applied to equalize signal intensity
age reduces the amplitude of P300 scalp electrical potentials, across subjects. In each subject, hemodynamic responses (8 frames
which are commonly elicited by the detection of infrequent visu- long) were estimated at the voxel level using the general linear model.
al, auditory and somatosensory targets during spatial and non- The design matrix was defined using impulse-basis functions such that
spatial tasks37. The right TPJ is also selectively activated when at each frame, the data were modeled as the sum of the overlapping
observers monitor the environment for infrequent targets (for hemodynamic response produced by each task effect and a linear trend.
example, vigilance38), and is the region most densely innervated The use of catch trials (trials in which only the cue stimulus was pre-
by noradrenergic projections from the locus coeruleus39 that are sent) made it possible to estimate unique responses for the cue, noise,
target-valid, and target-invalid periods even though the cue response
thought to mediate vigilance and arousal. Damage to the right overlaps noise and target responses in each full trial17 (J.M.O. et al.,
TPJ can cause vigilance problems in patients with unilateral Soc. Neurosci. Abstr. 24, 1178, 1998). A random-effects analysis was
neglect40. Changes in the level of vigilance have a slower time performed by entering the individual time points of each hemodynamic
course than shifts of attention41 and, therefore, might produce response into voxel-level ANOVAs45. These ANOVAs had two main
stronger and more sustained right TPJ responses. effects, time and task. The main effect of time was used to determine
A dissociation of function between intraparietal and tem- which voxels were activated. The resulting F-maps were corrected for
poroparietal cortex may explain why a verbal cue directing atten- multiple comparisons using a Gaussian random fields approach46. F-
tion toward the contralesional field can transiently reduce neglect statistics at each voxel were converted to equivalent Z-statistics. These
in some neglect patients, who typically have damage in the right Z-maps were used to delineate regions of interest. Separate ANOVAs
were then run at the regional level to determine the task effect of cue
TPJ region. This effect, extensively used by therapists to amelio-
direction (left, right), noise direction (left, right), visual field of the
rate unilateral visual neglect42,43, may reflect the preserved acti- target (left, right) and target validity (valid, invalid).
vation of the IPs, which mediates the allocation of attention by
cognitive cues. Normal orienting, however, can be maintained ACKNOWLEDGEMENTS
only for a short time in neglect patients based on voluntary strate- This research was supported by NIH EY00379 and EY12148 (M.C.). We thank
gies. Typically, orienting involves both cognitive and sensory-dri- Thomas Conturo, Avi Snyder and Erbil Akbudak for technical support.
ven mechanisms 27. This study, along with lesion analyses of
patients with TPJ damage, indicates that the right TPJ region is RECEIVED 27 SEPTEMBER 1999; ACCEPTED 6 JANUARY 2000

articles
1. Mesulam, M. M. A cortical network for directed attention and unilateral 24. Watson, J. D. et al. Area V5 of the human brain: evidence from a combined
neglect. Ann. Neurol. 10, 309315 (1981). study using positron emission tomography and magnetic resonance imaging.
2. Heilman, K. M., Watson, R. T. & Valenstein, E. in Clinical Neuropsychology Cereb. Cortex 3, 7994 (1993).
(eds. Heilman, K. M. & Valenstein, E.) 243293 (Oxford, New York, 25. Tootell, R. B. H. et al. Functional analysis of human MT and related visual
1985). cortical areas using magnetic resonance imaging. J. Neurosci. 15, 32153230
3. Halligan, P. W. & Marshall, J. C. Toward a principled explanation of unilateral (1995).
neglect. Special Issue: The cognitive neuropsychology of attention. Cognit. 26. Tootell, R. B. H. et al. The retinotopy of visual spatial attention. Neuron 21,
Neuropsychol. 11, 167206 (1994). 14091422 (1998).
4. Posner, M. I., Walker, J. A., Friedrich, F. J. & Rafal, R. D. Effects of parietal 27. Posner, M. I. Orienting of attention. Q. J. Exp. Psychol. 32, 325 (1980).
injury on covert orienting of attention. J. Neurosci. 4, 18631874 (1984). 28. Desimone, R. & Duncan, J. Neural mechanisms of selective visual attention.
5. Morrow, L. A. & Ratcliff, G. The disengagement of covert attention and the Annu. Rev. Neurosci. 18, 193222 (1995).
neglect syndrome. Psychobiology 16, 261269 (1988). 29. Le, T. H., Pardo, J. V. & Hu, X. 4T-fMRI study of nonspatial shifting of
6. Di Pellegrino, G. Clock-drawing in a case of left visuospatial neglect: a deficit selective attention: cerebellar and parietal contributions. J. Neurophysiol. 79,
of disengagement? Neuropsychologia 33, 353358 (1995). 15351548 (1998).
7. Friedrich, F. J., Egly, R., Rafal, R. D. & Beck, D. Spatial attention deficits in 30. Lumer, E. D., Friston, K. J. & Rees, G. Neural correlates of perceptual rivalry
humans: a comparison of superior parietal and temporo-parietal junction in the human brain. Science 280, 19301934 (1998).
lesions. Neuropsychology 12, 193207 (1998). 31. Colby, C. L., Gattass, R., Olson, C. R. & Gross, C. G. Topographic
8. Bushnell, M. C., Goldberg, M. E. & Robinson, D. L. Behavioral enhancement of organization of cortical afferents to extrastriate visual area PO in the
visual responses in monkey cerebral cortex. I. Modulation in posterior parietal macaque: a dual tracer study. J. Comp. Neurol. 238, 12571299 (1988).
cortex related to selective attention. J. Neurophysiol. 46, 755772 (1981). 32. Gnadt, J. W. & Andersen, R. A. Memory related motor planning activity
9. Colby, C. L., Duhamel, J. R. & Goldberg, M. E. Visual, presaccadic, and in posterior parietal cortex of macaque. Exp. Brain Res. 70, 216220
cognitive activation of single neurons in monkey lateral intraparietal area. J. (1988).
Neurophysiol. 76, 28412852 (1996). 33. Culham, J. C. et al. Cortical fMRI activation produced by attentive tracking of
10. Snyder, L. H., Batista, A. P. & Andersen, R. A. Coding of intention in the moving targets. J. Neurophysiol. 80, 26572670 (1998).
posterior parietal cortex. Nature 386, 167170 (1997). 34. Courtney, S. M., Ungerleider, L. G., Keil, K. & Haxby, J. V. Transient and
11. Gottlieb, J. P., Kusunoki, M. & Goldberg, M. E. The representation of visual sustained activity in a distributed neural system for human working memory.
salience in monkey parietal cortex. Nature 391, 481484 (1998). Nature 386, 608611 (1997).
12. Corbetta, M., Miezin, F. M., Shulman, G. L. & Petersen, S. E. A PET study of 35. Cohen, J. D. et al. Temporal dynamics of brain activation during a working
visuospatial attention. J. Neurosci. 13, 12021226 (1993). memory task. Nature 386, 604607 (1997).
13. Nobre, A. C. et al. Functional localization of the system for visuospatial 36. Mesulam, M.-M. Large-scale neurocognitive networks and distributed
attention using positron emission tomography. Brain 120, 515533 processing for attention, language, and memory. Ann. Neurol. 28, 597613
(1997). (1990).
14. Vandenberghe, R. et al. The influence of stimulus location on the brain 37. Knight, R. T. & Scabini, D. Anatomic bases of event-related potentials and
activation pattern in detection and orientation discrimination-a PET study of their relationship to novelty detection in humans. J. Clin. Neurophysiol. 15,
visual attention. Brain 119, 12631276 (1996). 313 (1998).
15. Gitelman, D. R. et al. Functional imaging of human right hemispheric 38. Pardo, J. V., Fox, P. T. & Raichle, M. E. Localization of a human system for
activation for exploratory movements. Ann. Neurol. 39, 174179 (1996). sustained attention by positron emission tomography. Nature 349, 6164
16. Corbetta, M. et al. A common network of functional areas for attention and (1991).
eye movements. Neuron 21, 761773 (1998). 39. Foote, S. L. & Morrison, J. H. Extrathalamic modulation of cortical function.
17. Shulman, G. L. et al. Areas involved in encoding and applying directional Annu. Rev. of Neurosci. 10, 6796 (1987).
expectations to moving objects. J. Neurosci. 19, 94809496 (1999). 40. Robertson, I. H., Mattingley, J. B., Rorden, C. & Driver, J. Phasic alerting of
18. Zarahn, E., Aguirre, G. & DEsposito, M. A trial-based experimental design neglect patients overcomes their spatial deficit in visual awareness. Nature
for fMRI. Neuroimage 6, 122138 (1997). 395, 169172 (1998).
19. Friston, K. J., Josephs, O., Rees, G. & Turner, R. Nonlinear event-related 41. Parasuraman, R., Warm, J. S. & See, J. E. in The Attentive Brain (ed.
responses in fMRI. Magn. Reson. Med. 39, 4152 (1998). Parasuraman, R.) 221256 (MIT Press, Cambridge, Massachusetts,
20. Kastner, S., Pinsk, M. A., De Weerd, P., Desimone, R. & Ungerleider, L. G. 1998).
Increased activity in human visual cortex during directed attention in the 42. Weinberg, J. et al. Visual scanning training effects on reading-related tasks in
absence of visual stimulation. Neuron 22, 751761 (1999). acquired brain-damage. Arch. Phys. Med. Rehabil. 58, 479486 (1977).
21. Chawla, D., Reese, G. & Friston, K. J. The physiological basis of attentional 43. Antonucci, G. et al. Effectiveness of neglect rehabilitation in a randomized
modulations in visual areas. Nat. Neurosci. 2, 671676 (1999). group study. J. Clin. Exp. Neuropsychol. 17, 383389 (1995).
22. Dupont, P. et al. The kinetic occipital region in human visual cortex. Cereb. 44. Raczkowski, D., Kalat, J. W. & Nebes, R. Reliability and validity of some
Cortex 7, 283292 (1997). handedness questionnaire items. Neuropsychologia 12, 4347 (1974).
23. Mendola, J. D., Dale, A. M., Fischl, B., Liu, A. K. & Tootell, R. B. The 45. Braver, T. S. et al. A parametric study of prefrontal cortex involvement in
representation of illusory and real contours in human cortical visual areas human working memory. Neuroimage 5, 4962 (1997).
revealed by functional magnetic resonance imaging. J. Neurosci. 19, 46. Worsley, K. J. et al. A unified statistical approach for detemining significant
85608572 (1999). signals in images of cerebral activation. Hum. Brain Mapp. 4, 5873 (1995).

Nature Neuroscience March 2000 PDF

Uploaded by

Document Information

Original Title

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

Nature Neuroscience March 2000 PDF

Uploaded by

Copyright:

Available Formats

2000 Nature America Inc. http://neurosci.nature.

volume 3 no 3 march 2000

Mice lacking NMDA receptor

an enriched environment, Mysterianism lite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

ChIPping away at potassium channel regulation . . . . . . . . . . . . . . . . . . . . . . . . . 202

Toying with memory in the hippocampus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

Attention - brains at work! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

Signaling dendritic growth in vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

Predicting perception from

nature neuroscience volume 3 no 3 march 2000 i

Rho GTPases regulate distinct aspects of dendritic arbor growth in

Growth cone and dendrite dynamics in zebrafish embryos: early

JD Jontes, J Buchanan and SJ Smith

Enrichment induces structural changes and recovery from nonspatial

Muscles express motor patterns of non-innervating neural networks by

Lack of cortical contrast gain control in human photosensitive epilepsy . . . . . . . 259

Learning to find a shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

Seeing multiple directions of motionphysiology and psychophysics . . . . . . . . 270

The neural mechanisms of top-down attentional control . . . . . . . . . . . . . . . . . . 284

Voluntary orienting is dissociated from target detection in human

nature neuroscience volume 3 no 3 march 2000 ii

nature neuroscience volume 3 no 3 march 2000 199

news and views

Predicting perception from

Treue and colleagues use electrophysiological recordings in monkeys and psychophysical

nature neuroscience volume 3 no 3 march 2000 201

news and views

202 nature neuroscience volume 3 no 3 march 2000

news and views

nature neuroscience volume 3 no 3 march 2000 203

news and views

204 nature neuroscience volume 3 no 3 march 2000

news and views

without NMDA-receptor-dependent LTP?

nature neuroscience volume 3 no 3 march 2000 205

news and views

206 nature neuroscience volume 3 no 3 march 2000

news and views

subjects know where to presentation. Unlike block designs, event-

when attention is really were making attentional shifts dur-

nature neuroscience volume 3 no 3 march 2000 207

news and views

208 nature neuroscience volume 3 no 3 march 2000

Correspondence should be addressed to B.L.M. (bruce@nsma.arizona.edu)

Fig. 1. Escher staircase track. (a) To obtain stimula-

nature neuroscience volume 3 no 3 march 2000 211

a Rat 2, flight day 9 may require either a period of adaptation to microgravity or

the international space station may yield a better understand-

212 nature neuroscience volume 3 no 3 march 2000

NMDA receptor-mediated control of

We demonstrate a rapid and complex effect of N-methyl-D-aspartate receptor (NMDAR) activation on

nature neuroscience volume 3 no 3 march 2000 211

(percent of value at incubation onset)

b resulted in a rapid decrease in synaptic protein

synthesis followed by a prolonged increase in

tions. Filled and open boxes represent the means

incubation (n = 9), respectively. Data for these

212 nature neuroscience volume 3 no 3 march 2000

Fig. 2. Alpha CaMK II synthesis is increased by

planned comparison). A second phase of

and 60 min after NMDAR activation (n = 3).

The numbers correspond to autoradiographs

CaMK II protein levels were monitored from