Development of the Visual System

The neocortex, or cerebral cortex, is a thick layer of cells (neurons) covering the outer surface of the fore-brain. In humans and some other mammals it is wrinkled, with many folds separated by deep fissures. In humans 75 percent or more of the neurons in the forebrain are in the cerebral cortex, a much higher proportion than in for other animals. Partly for this reason, and also because of clinical observations on people with brain damage, most researchers believe that the cortex plays a significant role in uniquely human behaviors, such as highly complex learning and information processing.

Our knowledge of how the cortex works has grown enormously since the 1950s and 1960s, mostly through experiments in which the activity of cortical neurons is studied one neuron at a time. This procedure may seem paradoxical, since there is great redundancy in the brain and the activity of any particular neuron cannot be crucially important. But through painstaking studies of hundreds of thousands of neurons in different cortical regions, a picture has emerged that allows us to understand many basic operating principles of the cortex.

An important function of the cortex is perception. Just about all the information we receive from the world around us, through our senses, arrives eventually in the cortex. The visual cortex is one of the best-understood cortical regions. It has a complicated, elegant structural and functional organization that appears to underlie many processes of visual perception and recognition.

Researchers have devoted extensive experimentation to answering the intriguing question of how the brain develops these sophisticated mechanisms for analyzing and interpreting the visual world. It is now clear that visual experience early in life is extremely important: The system is highly malleable or plastic because its development depends on the type and amount of visual information the young, developing organism receives; this property may prefigure the kind of plasticity of the adult brain that enables humans and animals to learn new behaviors.

Basic Properties of Visual Cortical Neurons

Information from the two eyes converges on individual neurons in the visual cortex; most of these neurons can respond to visual stimuli presented in either eye and so are termed binocular neurons. Many cells, though, respond more strongly to stimulation of one eye than of the other; the term ocular dominance refers to the relative influence of the two eyes on a cell's response. Visual cortical neurons have another interesting property: They are feature detectors. Almost every neuron responds best to a specific stimulus: usually a line, bar, or straight edge having a particular orientation (e.g., horizontal, vertical, diagonal) and precise location in the visual field. A third principal characteristic of the responses of neurons in the visual cortex is the relationship between a binocular cell's preferred or optimal stimulus in each eye. Many cells show interocular matching; that is, the stimulus orientation that produces the best response is the same in both eyes and has the same location within the visual field. Many other cells have slightly different optimal stimulus orientations or locations in the two eyes. These small differences are called interocular disparities, and neurons that exhibit this property are termed disparity detectors.

Plasticity in Visual Development

The question of how this cortical system for the analysis of visual space develops early in life has been the subject of hundreds of experiments since the 1960s. An important overall conclusion is that many response characteristics of visual cortical neurons depend on early visual experience for their normal development; these include binocularity and ocular dominance, orientation selectivity, and interocular matching of cells' optimal stimulus requirements, especially location and orientation (i.e., disparity detection). Typically, experiments of this sort involve raising young animals (usually kittens or monkeys, whose visual systems resemble those of humans) for days or weeks, controlling or manipulating their visual experience to depart from the norm in various ways. Ensuing microelectrode studies determine whether and how the abnormal experience has altered the activity of visual cortical cells.

During the first weeks or months of postnatal life, called the critical or sensitive period, the developing mammal's visual system undergoes a period of special vulnerability or susceptibility to the effects of many altered rearing conditions. As the critical period evolves into adulthood, the same manipulations of the animal's visual experience generally do not affect the physiological organization of the visual cortex in terms of neurons' response characteristics (although some kinds of plasticity persist).

For instance, early studies examined the effects of completely depriving kittens of experience with visual form or pattern by closing both eyelids during the first few postnatal weeks. Initially it appeared that this manipulation does not produce major changes: Visual cortical cells' binocularity and the distribution of ocular dominance are unaffected, and these kittens' cells also show orientation selectivity resembling that of normally reared kittens, although the precision of orientation detection is slightly reduced. Subsequently, more detailed experiments showed that the system of disparity-detecting neurons does not develop normally in binocularly deprived kittens.

By contrast, kittens raised with just one eyelid closed (monocular deprivation, or MD) show dramatic changes in their visual cortical organization. Nearly all the cells are responsive only to stimulation of the eye that was open, and almost none to stimulation of the formerly closed eye. As might be expected, behavioral tests of visual acuity reveal deficits in the deprived eye.

This cortical ocular dominance shift is not due simply to the presence or absence of patterned visual input but to the absence of simultaneous stimulation of both eyes, as shown by experiments in which kittens were raised with one eye closed for several days or a few weeks, after which that eye was opened and the opposite eye was closed for a comparable period (reverse suture), and also by experiments involving alternating daily MD. Kittens raised using these methods experience the same amount of patterned visual input through both eyes, but at any given moment only one eye is receiving stimulation. When cortical ocular dominance and binocularity are studied in these kittens, almost all neurons are visually responsive, but each responds only to stimulation of one eye; there are almost none of the binocular cells that make up 80 to 90 percent of the visual cortical neurons in normally reared kittens. Furthermore, if the relative amount (time) of stimulation given the two eyes is made unequal in these experiments, there is a corresponding change in the proportion of cells activated by stimulation of each eye.

Not only must patterned visual stimulation be temporally synchronized (simultaneous) between the two eyes in order for cortical binocularity to develop normally, it must also be spatially synchronous; that is, each part of the pattern must stimulate the same point on both retinas (corresponding points). Some humans exhibit an oculomotor disorder called strabismus, in which the two eyes are misaligned. People with this disorder often have poor visual acuity in one eye and almost always have deficient binocular depth perception. An animal model of strabismus, in which some of the muscles that control the position of the eyes are severed, has proved useful in studying the cortical effects of this disorder. Kittens raised with experimental strabismus show a marked loss of binocular neurons; in fact, the physiological organization of the visual cortex resembles that of kittens reared with alternating monocular deprivation.

The system of orientation-detecting neurons in the visual cortex is also susceptible to the effects of early visual experience. Kittens raised viewing contours or edges confined to a single orientation (e.g., vertical stripes) have a preponderance of visual cortical cells whose preferred receptive fields are at or near the experienced orientation; this finding is in marked contrast with the cortical organization of normally reared kittens, in which all possible orientations are represented about equally among cells' receptive fields.

In addition to the dramatic alterations in binocularity and orientation selectivity consequent upon experimental manipulations of early visual experience, there is also a degree of plasticity in the development of visual cortical cells specialized for disparity detection, especially interocular orientation disparity. For instance, there are changes in the visual cortices of kittens raised wearing prism goggles that introduce rotations of the images seen by the left and right eyes. These rotations are around the visual axis (line of sight), and are opposite in the two eyes, producing a controlled amount of interocular orientation disparity; that is, an edge or contour in the field of view does not give rise to parallel images on the two retinas, as is normally the case, but instead is displaced clockwise in one eye and counterclockwise in the other. If these rotations are small (e.g., eight degrees in each eye), there is a corresponding shift in the average disparity of cortical neurons' preferred receptive-field orientations between the two eyes: Most cells show an interocular orientation disparity that matches the experienced image rotation. On the other hand, if the rotations are large (e.g., sixteen degrees or more in each eye), there is a disruption of binocularity: Most cells respond only to stimulation in one eye but not in both eyes. In this respect the effects of large interocular rotations are like those of strabismus or alternating monocular deprivation.

From an evolutionary standpoint, cortical plasticity in the development of interocular relationships has clear adaptive significance. The developing animal undergoes relatively rapid changes in height and in the lateral separation of the two eyes. There is thus a continually changing relation between the interocular image disparity of objects in the environment and the distance of those objects from the observer. The existence of neuronal disparity-detecting mechanisms able to adjust to these changes during early life provides an advantage in capturing prey, eluding predators, and so forth. Detailed reviews of the many studies on this issue have been written by Frégnac and Imbert (1984) and by Shinkman, Isley, and Rogers (1985).

Some Theoretical Considerations

A major problem is identifying underlying mechanisms responsible for plasticity in the developing visual system. One appealing idea involves binocular competition, in which fibers from the thalamus, some carrying information from one eye and some from the other, compete to form synapses on binocular cortical cells. Although the underlying competitive mechanism it is not yet clear, recent work points to a major role for nerve growth factor (NGF): rats with intraventricular injections of NGF fail to show the expected ocular dominance shift when subjected to MD during the critical period (Pizzorusso and Maffei, 1996). In any case, when researchers place one eye at a disadvantage during MD, fibers representing the other eye are more successful at making cortical connections.

At the same time, it is clear that neural activity originating in the deprived eye continues to play a role in animals subjected to MD. Some researchers have raised kittens with MD combined with damage to a small area of the retina in the open eye. Later, these kittens show the usual shift in ocular dominance toward the experienced eye, except that cortical cells that would otherwise have been responsive to stimulation of the damaged retinal area are instead responsive to stimulation of the deprived eye. The important role of intracortical inhibitory mechanisms is consistent with this finding. For example, if a drug that blocks the action of inhibitory neurotransmitters is administered to a previously monocularly deprived kitten while recordings from cells in the visual cortex are in progress, responsiveness to stimulation of the deprived eye increases immediately; this effect continues until the drug wears off and then disappears.

What, exactly, is the role of visual experience in neocortical development? Does it simply maintain the feature-detecting capabilities and the interocular relationships of visual cortical neurons, or does it sharpen and even alter these properties? Some researchers have related this question to the age-old philosophical issue of nature versus nurture; however, it is now clear that both genetic influences (nature) and the influences of the individual's unique visual environment (nurture) are crucial. The real question concerns the relative degree of these influences on the development of the organization of our visual system and of our perceptual capacities. Most researchers believe that early visual experience can, within limits, modify the formation of connections in the central nervous system, thereby exerting substantial control over the final outcome. This conclusion is borne out both by studies using experimental animals and by clinical observations on humans who have experienced visual disorders in early childhood.

Relation of Developmental Plasticity to Learning and Memory in Adults

Candidate Pharmacological and Neurochemical Mechanisms

We are now beginning to understand some mechanisms that may underlie both neural plasticity in early development and plasticity as manifested in adult learning and memory. These may include some dynamic aspects of the synaptic relations within neuronal networks of the cerebral cortex and some neurochemical, especially neurotransmitter, changes that accompany (and may ultimately be responsible for) some of the phenomena of neuronal plasticity described above. For instance, the neurotransmitters norepinephrine (NE) and acetylcholine (ACh) play a critical role in the formation and maintenance of adult memories, and visual cortical plasticity is reduced or abolished when levels of these neurotransmitters are substantially depleted experimentally.

Neurotransmitters exert many of their effects in the brain by acting on their receptors located on post-synaptic cells. There are numerous classes of receptors in the central nervous system; one notable type is N-methyl-D-aspartate (NMDA) receptor, whose action is voltage-dependent; its properties come into play only with some depolarization of the postsynaptic cell. It may therefore be a kind of gate, permitting additional excitation only atop some excitatory effects already present in the postsynaptic neuron. Activation of the NMDA receptor may thus be a neurochemical analogue of the behavioral excitation that ensues from the combination of a conditioned stimulus and an unconditioned stimulus in a learning experiment. Indeed, drugs that block the NMDA receptor interfere with or even prevent the normal acquisition of learned responses in experimental animals. As with NA and ACh depletion, this effect has been obtained using several different kinds of conditioning procedures.

In kittens, the pharmacological blocking of NMDA receptors also blocks the neural plasticity shown in the loss of cortical binocular cells following monocular deprivation. NMDA receptors are composed of subunits that experimentation can selectively alter; recent studies have begun to delimit specific subunits responsible for various plastic changes in cortical neuronal function, both in vivo and also in vitro (Bear and Rittenhouse, 1999; Philpot et al., 2001). Furthermore, plasticity that arises from the actions of neurotransmitters and from the activation of NMDA receptors has been demonstrated at the cellular level through iontophoresis, recording from a single neuron while using minute electrical currents to eject small quantities of neurotransmitter substance or of NMDA from the electrode into the immediate vicinity of the neuron under study. The responses of many visual cortical cells to a visual stimulus presented to the nondominant eye, or at a nonoptimal orientation, show a substantial temporary increase in strength when these stimuli are repeatedly paired with the iontophoretic application of NA and ACh, or of NMDA and glutamate (an excitatory transmitter that acts upon NMDA receptors). This effect is evident in kittens but not in adult cats. Thus these neurotransmitter systems and the NMDA receptors have been clearly implicated in neuronal plasticity early in life and also in adult learning and memory.

Application of Transgenic Models

Some researchers have examined the visual cortical plasticity in genetically altered subjects. Gordon (1997) has provided an excellent review of results obtained using transgenic mice. For instance, ocular dominance shifts consequent upon MD occur in some mice bred to lack the gene encoding a particular form of calcium/calmodulin-dependent protein kinase II. Thus calcium, with its well-known role in presynaptic events, may help to signal activity-dependent plasticity using this particular protein. This novel line of research might yield major advances in our understanding of molecular mechanisms of both adult learning and memory and developmental plasticity.

Behavioral State

Various aspects of animal behavior that contribute to learning and memory appear to enable or enhance cortical plasticity. For instance, the well-studied effects of sleep upon memory consolidation apply in much the same fashion to MD-induced plasticity in visual cortex (Frank, Issa, and Stryker, 2001). Kittens that underwent anesthesia-induced sleep following MD exhibited more pronounced shifts in ocular dominance than kittens treated comparably but kept awake, either in a lighted environment or in darkness. Perhaps other behavioral-state variables can similarly affect cortical plasticity.

Conclusion

There has been a dramatic increase in experimental attention to the relation between brain mechanisms of developmental plasticity and of learning and memory. The search for general mechanisms of neural plasticity is likely to remain a central concern of neuroscience, which will benefit especially from studies of visual cortical neuronal networks.

Bibliography

Bear, M. F., and Rittenhouse, C. D. (1999). Molecular basis for induction of ocular dominance plasticity. Journal of Neurobiology 41, 83-91.

Frank, M. G., Issa, N. P., and Stryker, M. P. (2001). Sleep enhances plasticity in the developing visual cortex. Neuron 30, 275-287.

Frégnac, Y., and Imbert, M. (1984). Development of neuronal selectivity in primary visual cortex of cat. Physiological Reviews 64, 325-434.

Gordon, J. A. (1997). Cellular mechanisms of visual cortical plasticity: A game of cat and mouse. Learning and Memory 4, 245-261.

Philpot, B. D., Weisberg, M. P., Ramos, M. S., Sawtell, N. B., Tang, Y.-P., Tsien, J. Z., and Bear, M. F. (2001). Effect of transgenic overexpression of NR2B on NMDA receptor function and synaptic plasticity in visual cortex. Neuropharmacology 41, 762-770.

Pizzorusso, T., and Maffei, L. (1996). Plasticity in the developing visual system. Current Opinion in Neurology 9, 122-125.

Shinkman, P. G., Isley, M. R., and Rogers, D. C. (1985). Development of interocular relationships in visual cortex. In R. N. Aslin, ed., Advances in neural and behavioral development, Vol. 1. Norwood, NJ: Ablex.

Paul G.Shinkman