R348 Dispatch

Cortical development: Binocular plasticity turned outside-inIan Thompson

Classically, monocular deprivation leaves all layers ofvisual cortex dominated by the non-deprived eye.Unexpectedly, the changes first appear in the outerlayers, not the central input layer. Do thalamocorticaland corticocortical synapses differ in their plasticity andcould the outer layers drive input plasticity?

Address: University Laboratory of Physiology, University of Oxford,South Parks Road, Oxford OX1 3PT, UK.E-mail: ian.thompson@physiol.ox.ac.uk

Current Biology 2000, 10:R348–R350

0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.

Binocular plasticity in developing visual cortex has beenone of the key paradigms in developmental neuroscience.The discovery by Wiesel and Hubel that early depriva-tion of vision in one eye altered the representation of thetwo eyes in the cortex led to a plethora of experimentsdesigned to reveal the underlying mechanisms and opera-tional characteristics of the process. In normal visualcortex, most neurons respond to visual stimulation ineither eye. Input from the two eyes is relayed via thethalamus to layer IV of the visual cortex, and thence tothe other five cortical layers. After prolonged monoculardeprivation, however, the response of neurons in allcortical layers is dominated by the non-deprived eye.This physiological change is associated with anatomicalchanges in layer IV: the terminals of thalamic axons con-nected to the deprived eye shrink, whereas those con-nected to the non-deprived eye expand. These dramaticchanges are seen only if monocular deprivation occurs rel-atively early in development — there is a sensitive periodfor binocular plasticity.

A common assumption is that the functional developmentof cortex, and hence plasticity, proceeds in an ‘inside-out’manner — that is, changes begin at the thalamocorticalconnections in layer IV, and subsequently spread into thesurrounding cortical layers (I,II,III,V and VI). For instance,orientation-tuned neurons emerge first in layer IV andtend to be monocular [1]. In vitro slice studies of bothvisual and somatosensory cortex have revealed that, whileimmature thalamocortical synapses do display plasticity —both long-term potentiation, LTP, and depression, LTD— corticocortical synapses in the outer layers remainplastic for much longer [2,3]. Interestingly, the sensitiveperiod for binocular plasticity is longer in the outer layersof visual cortex and can be modulated by dark rearing, asfound also for in vitro plasticity [2,4]. Thus, a recent studyby Trachtenberg et al. [5], which shows that the effects of

brief monocular deprivation are first seen in the outerlayers of visual cortex, raises some unexpected questions.

In an elegantly designed experiment, Trachtenberg et al. [5]first used optical imaging to map the cortical representa-tions of the two eyes in normal kittens and in kittens thathad received 24 hours of monocular deprivation. Thisshort period of deprivation significantly shifts the balanceof binocularity, but does not leave cortex totally domi-nated by the non-deprived eye. For normal animals, theimaging reveals ocular dominance stripes — regions ofcortex in which activity is biased towards input from eitherthe left or the right eye. Even within one stripe, the ocularbias is not uniform; regularly spaced ‘islands’, where the

Figure 1

A cartoon of the representation of the two eyes across the corticalsurface, based on the optical imaging experiments of Trachtenberget al. [5]. Cortical regions dominated by input from just one eye aredepicted in either blue or yellow. The greater the degree of binocularbalance, the closer the colour code comes to green. (a) The pattern innormal animals, 28–35 days of age. Most of the cortex is activatedbinocularly, but two ocular dominance bands can be seen runningacross cortex; in these bands the balance of activity is towards oneeye or the other. Within each band, however, there are regularlyspaced islands of strong monocular dominance. Overall, normal cortexat this age is equally influenced by the two eyes. (b) The pattern after24 hours of monocular deprivation. The overall balance of activity hasshifted dramatically to the non-deprived eye (activity through this eye isencoded in blue). Some activity can be elicited through the deprivedeye, as seen in the isolated yellow islands. The black crosses in bothpanels indicate typical locations of surface normal electrodepenetrations. These are used to examine the laminar distribution ofocularity. They have been positioned equidistant from the monocularislands to minimise any intrinsic bias towards one eye or the other.

bb10i03.qxd 10/5/00 9:05 am Page R348

Dispatch R349

monocular bias is greatest, are found (Figure 1a). The con-sequence of brief monocular deprivation is to dramaticallyreduce the representation of the deprived eye just to theislands of monocularity, which are isolated in a sea ofcortex dominated by the open eye (Figure 1b). Trachten-berg et al. [5] went on to use electrophysiology to investi-gate how neurons in different cortical layers are affectedby brief monocular deprivation. The optical imagingmaps, however, make it obvious that the degree of binocu-lar plasticity varies with cortical location. By placing themicroelectrodes in cortex equidistant from the islands ofmonocularity (as indicated by the crosses in Figure 1a,b),they were able to focus on what were, initially, the mostbinocular regions of the visual cortex.

Recordings from normal animals confirmed a high level ofbinocularity. Trachtenberg et al. [5] found, on moving theirelectrode down through the cortical laminae, that neurons inall layers are clearly influenced by both eyes, and that, formost neurons, the influence is fairly well balanced. InFigure 2a–c, green shading indicates closely balanced inputfrom the two eyes, and blue or yellow monocular dominancefrom the left or right eye, respectively. In monocularlydeprived animals, recordings from neurons in the upper cor-tical layers were seen to be dominated by input from theopen eye (blue shading in Figure 2d). As expected from theoptical imaging, there has been a significant shift in ocular-ity. A similar shift was also found in the deep cortical layers(Figure 2f). But neurons in layer IV showed little or no plas-ticity: the distribution of ocularity after brief monoculardeprivation was found to be no different from normal(compare Figure 2b and Figure 2d). This lack of binocularplasticity in layer IV following brief monocular deprivationwas robust across animals and penetrations.

Brief monocular deprivation at the height of the sensitiveperiod thus appears to have little effect on thalamocorti-cal connectivity, but significantly alters corticocorticalconnectivity, as assessed by single unit activity. Theseresults confirm the earlier work of Kossut and hercolleagues [6,7] on the effects of brief monocular depriva-tion, which used 2-deoxyglucose autoradiography andcurrent source density analysis: both techniques showedgreater plasticity outside layer IV than inside. What are theimplications of these observations? The findings raisesome intriguing questions about the intrinsic plasticity ofthalamocortical and corticocortical synapses, and the mech-anisms of cortical binocular plasticity.

Is there any reason to think that thalamocortical andcorticocortical connections are intrinsically different in theirpotential for plasticity? In the mature visual cortex,thalamocortical synaptic input to layer IV neurons is bothlarger and less variable than corticocortical input [8].Although both thalamocortical and corticocorticalconnections can display LTP and LTD in slice preparations

of developing cortex, only connections in the upper layersretain this plasticity into adulthood [2,3]. For in vitro induc-tion of LTD in layer IV of juvenile visual cortex, however,blockers of inhibitory neurotransmission mediated by γ-amino butyric acid (GABA) were found to be required; thismay reflect the high level of GABA receptors in layer IV[9]. Finally, there is the question of the sensitive period forthalamocortical synapse plasticity. In the work of Trachten-berg et al. [5], the monocular deprivation took place at theheight of the sensitive period, as defined by studies pooling

Figure 2

A graphical representation of the balance of ocularity in differentcortical layers in normal (a–c) and monocularly deprived (d–f) animals,derived from the electrophysiological data of Trachtenberg et al. [5].The electrode passed surface-normal first through the upper layers IIand III (a,d), then through the central, input layer IV (b,e) and finallythrough the lower cortical layers V and VI (c,f). As in Figure 1, theconvention is that data from neurons exclusively activated by one eyeor the other are shown in either blue or yellow; data from neurons withequally balanced input from the two eyes are indicated in green.Intermediate shades give results from neurons in which the input fromthe two eyes is unbalanced to a greater or lesser degree. Each panelgives the percentage of neurons recorded in given layers that fall intothe different ocularity groups. In (d–f), yellow indicates the influence ofthe deprived eye and blue that of the non-deprived eye (as indicatedbeneath each graph).

bb10i03.qxd 10/5/00 9:05 am Page R349

R350 Current Biology Vol 10 No 9

results from all cortical layers [10]. Is it possible that func-tional thalamocortical plasticity peaks earlier?

At first sight, the consequences of brief monoculardeprivation observed by Trachtenberg et al. [5] are some-what paradoxical. In a region of visual cortex in whichlayer IV neurons are binocular, the neurons in the outerlayers of the same column, activated themselves by thelayer IV neurons, are in fact monocular. How has the inputfrom the deprived eye been lost? This result is only para-doxical on the assumption that the relay of information isstrictly orthogonal to the cortical layers, and this is knownnot to be the case. For instance, widespread horizontalconnections are found in the upper cortical layers. Theseconnections are thought to play an important role in theplasticity seen in adult cortex: could they play a role here?

In such a scenario, the rapid spread of the non-deprivedeye’s dominance in the upper layers would be mediatedthrough increased driving by horizontal connections asso-ciated with that eye. Changes in the balance of horizontaland vertical connectivities in visual cortex followingmonocular deprivation have yet to be explored, but suchchanges have been seen in somatosensory cortex. In astudy published last year, Finnerty et al. [11] looked at theconsequences of sensory deprivation for the whiskerbarrel cortex of young mice. They concluded that the ver-tical pathway from layer IV to the upper layers wasstrengthened in non-deprived regions of cortex, as was thehorizontal input from non-deprived to deprived cortex,whereas the horizontal input from deprived to non-deprived input was weakened. The balance between hori-zontal and vertical connections may be just as important indevelopmental plasticity as in adult plasticity.

Of course, monocular deprivation does produce plasticityat the thalamocortical level: for instance, deprivation for asshort a period as four days causes the terminals of thalamicaxons connected to the deprived eye to shrink. Is thisplasticity simply slower and independent of that in theouter layers, or could it be, as Trachtenberg et al. [5]speculate, that plasticity in the outer cortical layers directsthat in layer IV? How the latter might occur clearly callsfor experimental investigation. But it is worth remember-ing that one of the outer layers, layer VI, provides a majorexcitatory input to layer IV [12]. This input can modulatethalamic transmission in the adult [13,14] — might it playa comparable role in development? So almost forty yearsafter the first description of binocular plasticity in thevisual cortex, the paradigm continues to yield surprises.Not only is there debate about the necessity for interocu-lar competition [15], but there is now uncertainty aboutwhich cortical synapses drive the process. Despite this,the combination of results from careful whole-animalwork, from in vitro slices and from molecular approachespromises a much more integrated view.

References1. Blakemore C, Van Sluyters RC: Innate and environmental factors in

the development of the kitten’s visual cortex. J Physiol (Lond),1975, 248:663-716.

2. Bear MF, Rittenhouse CD: Molecular basis for induction of oculardominance plasticity. J Neurobiol 1999, 41:83-91.

3. Feldman DE, Nicoll RA, Malenka RC: Synaptic plasticity atthalamocortical synapses in rat somatosensory cortex: LTP, LTDand silent synapses. J Neurobiol 1999, 41:92-101.

4. Mower GD: The effect of dark rearing on the time course of thecritical period in cat visual cortex. Dev Brain Res 1991, 58:151-158.

5. Trachtenberg JT, Trepel C, Stryker MP: Rapid extragranularplasticity in the absence of thalamocortical plasticity in thedeveloping primary visual cortex. Science 2000, 287:2029-2032.

6. Kossut M, Thompson ID, Blakemore C: Ocular dominance columnsin cat striate cortex and effects of monocular deprivation: a2-deoxyglucose study. Acta Neurobiol Exp 1983, 43:273-282.

7. Kossut M, Singer W: The effects of short periods of monoculardeprivation on excitatory transmission in the striate cortex ofkittens: a current source density analysis. Exp Brain Res 1991,85:519-527.

8. Stratford KJ, Tarczy-Hornoch K, Martin KAC, Bannister NJ, Jack JJB:Excitatory inputs to spiny stellate cells in cat visual cortex. Nature1996, 382:258-261.

9. Dudek SM, Friedlander MJ: Developmental down-regulation of LTDin cortical layer IV and its independence of modulation by inhibition.Neuron 1996, 16:1097-1106.

10. Olson CR, Freeman RD: Profile of the sensitive period formonocular deprivation in kittens. Exp Brain Res 1980, 39:17-21.

11. Finnerty GT, Roberts LSE, Connors BW: Sensory experiencemodifies the short-term dynamics of neocortical synapses. Nature1999, 400:367-371.

12. Ferster D, Lindstrom S: Synaptic excitation of neurones in area 17of the cat by intracortical axon collaterals of cortico-geniculatecells. J Physiol (Lond) 1985, 367:233-252.

13. Ferster D, Lindstrom S: Augmenting responses evoked in area 17of the cat by intracortical axon collaterals of cortico-geniculatecells. J Physiol (Lond) 1985, 367:217-232.

14. Grieve KL, Sillito AM: Non-length-tuned cells in layers II/III and IV ofthe visual cortex: the effect of blockade of layer VI on responses tostimuli of different lengths. Exp Brain Res 1995, 104:12-20.

15. Kind PC: Cortical plasticity: is it time for a change? Curr Biol 1999,9:R640-643.

bb10i03.qxd 10/5/00 9:05 am Page R350