Copyright © 1999 Massachusetts Institute of Technology
The functional properties of neurons and the functional architecture of the CEREBRAL CORTEX are dynamic, constantly under modification by experience, expectation, and behavioral context. Associated with functional plasticity is a process of modification of circuits, either by altering the strength of a given synaptic input or by axonal sprouting and synaptogenesis. Plasticity has been seen under a number of conditions, including functional recovery following lesions of the sensory periphery of central structures, perceptual learning and learning of object associations, spatial learning, visual-motor adaptation, and context-dependent changes in receptive field properties. This discussion will compare plasticity observed early in development with that seen in adulthood, and then discuss the role of plasticity in recovery of function after lesions, in learning in sensory systems, and in visual-spatial integration.
Much of the original work on neural plasticity in the central nervous system was done in the context of experience-dependent plasticity in the period of postnatal development during which cortical connections, functional architecture, and receptive field properties continue to be refined. Hubel and Wiesel (1977) showed that in the visual system, the balance of input from the two eyes, known as ocular dominance, can be influenced by keeping one eye closed, which shifts the balance toward the open eye, or by induced strabismus (where the two eyes are aimed at different points in the visual field), which blocks the development of binocular cells. The substrate of these changes is an alteration in the extent of thalamocortical axonal arbors, which, immediately after birth, are undergoing a process of collateral sprouting and pruning. The plasticity of these arbors, of ocular dominance columns, and of the ocular dominance of receptive fields, is under experience-dependent regulation for a limited period early in the life of animals that is known as the critical period. The length of the critical period is species-dependent, and can extend for the first few months (cats) or years (humans and nonhuman primates) of life. After the end of the critical period the properties involved become fixed for the rest of the life of the animal.
The model of ocular dominance plasticity has been a prime example of the role of activity and of experience in the formation of the functional properties of neurons and of the refinement of connectivity between neurons. Models of cortical development have shown how spontaneous activity in utero can lead to the formation of cortical functional architecture and receptive field properties in the absence of visual experience, and prenatal patterned activity in the RETINA and cortex has been discovered. It has been shown that the effects of monocular deprivation can be prevented by blockade of retinal activity. Some of the molecular intermediaries implicated in the competition between different populations of cortical afferents and of activity-dependent plasticity include neurotrophins and their receptors and glutamate and N-methyl-d-aspartate (NMDA) receptors. The fundamental rule underlying this plasticity, originating from work in the HIPPOCAMPUS and the ideas of HEBB, is that neurons that fire together wire together, and that LONG-TERM POTENTIATION is involved in the consolidation of connections.
Given this background it had been widely assumed that all cortical areas, at least those involved in early sensory processing, would have fixed properties and connections. Of course, some measure of plasticity would have to accompany the ability to acquire and store new percepts throughout life, but this had been thought to be a special property of higher-order cortical areas, particularly those in the temporal lobe, associated with object memory. A radical change has occurred in this view with the growing body of evidence that experience-dependent plasticity is a universal property of cortex, even primary sensory areas.
Each area of sensory cortex, particularly those at early stages in the sensory pathway, has a representation of the sensory surface on the cortical surface. Somatosensory cortex contains a representation of the body map (somatotopy), auditory cortex of the cochlea (tonotopy), and visual cortex of the retina (visuotopy). The integrity of these maps depends on ongoing stimulation of the periphery. Removal of input from any part of the sensory surface, such as by digit amputation or retinal lesion, leads to a reorganization of the cortical maps. Some of the initial evidence for cortical plasticity in the adult came from changes in somatotopic maps following digit amputation (see Merzenich and Sameshima 1993). Amputation of a body part or transection of a sensory nerve causes the area of cortex initially representing that part to be remapped toward a representation of the adjacent body parts. Retinal lesions lead to a shrinkage of the cortical representation of the lesioned part of retina and an expansion in the representation of the part of retina surrounding the lesion.
The mechanism of the reorganization varies with the sensory pathways involved. Generally, the site of reorganization depends on the existence of exuberant connections linking cells representing widely separated parts of the map. Thus, in the somatosensory system, a measure of lesion-induced plasticity can be observed in the spinal cord, although it is likely that a considerable degree of plasticity is based in the somatosensory cortex. In the plasticity observed in the visual system, most of the changes are intrinsic to the visual cortex, and are likely to involve the long-range horizontal connections formed by cortical pyramidal cells. The unmasking of these connections seen with long-term topographic reorganization involves a sprouting of axonal collaterals and synaptogenesis.
There is a wide range of time scales over which the plasticity of topography takes place. The changes occurring over the largest spatial scales in cortex (topographic shifts of up to a centimeter) require several months or years. Smaller but significant changes can be seen within minutes following a lesion, and this is likely to involve changes in the strength of existing connections. Exuberant connections can be unmasked by a potentiation of excitatory connections or by a suppression of inhibitory connections.
The perceptual consequences of lesion-induced plasticity can include, depending on the site of the lesion, a recovery of function or perceptual distortions. PHANTOM LIMB sensation following limb amputation has been linked by Ramachandran (1993) to experimentally induced somatosensory cortical plasticity. For arm amputations, there is often a sensation of stimulation of the absent hand when stroking the limb stump or the cheek. Human patients suffering a loss of central retinal input (by, e.g., age-related macular degeneration) often adopt a preferred retinal locus in the intact retina for targeting visually guided eye movements. Lesions in area MT, an area that plays a role in the perception of movement and the tracking of moving objects by the eyes, initially leads to a loss of smooth pursuit eye movements, but within a few days this function recovers. It is well known that following stroke there are varying degrees of functional recovery. Though this recovery had been thought to involve a return to health of metabolically compromised but not destroyed tissue, it may well be that intact areas of cortex are taking over the function of adjacent cortical regions that have been destroyed.
The mechanisms available to the cortex for functional recovery following lesions are likely to be used for normal sensory processing. It is likely that perceptual learning involves analogous changes in cortical topography. LEARNING in general has been divided into categories including declarative or explicit learning (including the learning of events, facts, and objects) and nondeclarative or implicit learning (including procedural, classical CONDITIONING, priming, and perceptual learning). These different forms of learning may be distinguished less on the basis of the underlying synaptic mechanisms than on the brain region in which the memory is stored. While one ordinarily associates sensory learning with the acquisition and storage of complex percepts and with the temporal lobe, it has been known for well over a hundred years that it is possible to improve one's ability to discriminate simple sensory attributes. Some characteristics of perceptual learning are suggestive of the involvement of early stages in sensory processing.
Perceptual learning has been shown to apply to a wide range of sensory tasks, including visual acuity, hue discrimination, velocity estimation, acoustic pitch discrimination, and two-point somatosensory acuity. This is a form of implicit learning, generally not reaching conscious awareness or requiring error feedback, but is associated with repetitively performing discrimination tasks.
The evidence supporting the idea that the neural substrate for perceptual learning is found in primary sensory cortex comes from the specificity of the learning and from physiological studies demonstrating cortical changes in animals trained on simple discrimination tasks. Improvement in a visual discrimination task at one position in the visual field, for example, does not transfer to other locations. Since the highest spatial resolution is seen in primary visual cortex, where the receptive fields are the smallest and the topography most highly ordered, one might expect to find the basis for specificity there. The learning also shows no transfer to orthogonal orientations, again indicative of early cortical stages where selectivity for stimulus orientation is sharpest. On the other hand, the learning is also specific for stimulus configuration or the context within which a feature is embedded during the training period, which may point toward a feedback influence from higher-order cortical areas providing information about more complex features.
The physiological studies show changes in cortical magnification factor, or cortical recruitment, associated with training. Training a monkey to do a texture discrimination task with a particular digit will increase the area of primary somatosensory cortex representing that digit. Similarly, it has been suggested that training on an auditory frequency discrimination task increases the representation of that frequency in primary auditory cortex. Not only these forms of implicit learning but associative learning as well causes changes in the receptive fields of cells in primary sensory cortex. When a tone is associated with an aversive stimulus, cells in auditory cortex tend to shift their critical frequency toward that of the tone. The reward component of the training may come from basal forebrain structures, involving the diffuse ascending cholinergic input to cortex.
The storage of more complex information has been identified in the inferior temporal cortex. There, animals trained to recognize complex forms, particularly in association with other forms, have cells that showed elevated activity when a given form is presented. The acquisition of the trained information may depend on input from more medial structures, such as the perirhinal cortex.
A central focus for studies of neural plasticity is the hippocampus. As a consequence of neuropsychological findings showing that persons with medial temporal lesions suffer an inability to acquire and store recent memories, the hippocampus has been an active area of study for neural mechanisms of MEMORY. At the systems level, the hippocampus was shown by O'Keefe (1976) to play a role in spatial learning, with cells being tuned for an animal's position in its external environment, known as a place field. At the synaptic level, the hippocampus has become the prime model for changes in synaptic weight, through the phenomenon of long-term potentiation originally described by Bliss and Lomo (1973), and long-term depression. While it has been presumed that these forms of synaptic plasticity account for the storage of complex information in the hippocampus, the linkage has not yet been established. It is clear, however, that cells in the hippocampus are capable of rapidly changing their place fields as the external environment is altered, and this alteration is associated with changes in effective connectivity between hippocampal neurons.
The functional properties of cells in cerebral cortex and in brain stem have been shown to be modifiable over much shorter time scales, potentially involving neural plasticity in the ongoing processing of sensory information and in sensorimotor integration.
One of the earliest and most active areas of investigation of adult plasticity is the vestibulo-ocular reflex, the compensatory movement of the eyes associated with rotation of the head or body to keep the visual field stabilized on the retina. Melville-Jones and Gonshor (1975) found that if prisms are put on the eyes to reduce the amount of retinal slip associated with a given amount of head rotation, the gain of the reflex is reduced, and eventually settles to a level where once again the world is stabilized on the eyes. Various brain structures have been suggested to be involved in this phenomenon, including the CEREBELLUM and pontine nuclei. Another revealing model of sensorimotor adaptation has been studied in the owl by Knudsen and Brainard (1995). In the tectum of the owl (analogous to the mammalian superior colliculus), there are superimposed maps of visual and auditory space. When prisms are placed over the owl's eyes, shifting the visual map, there is a compensatory shift of the auditory map, so that once again there is a registration between the two maps for a given elevation and azimuth in the external world. This enables the owl to make accurate targeting movements for catching prey, as detected by both visual and sound cues.
Within sensory cortex, rapid changes in receptive field properties have been associated with perceptual fill-in. When an occluder, or artificial scotoma, is placed within a background of a uniform color or a textured pattern, and when this stimulus is stabilized on the retina, the occluder disappears over a few seconds, becoming filled with the surrounding pattern. It is supposed that this phenomenon may be a manifestation of the process of linkage and segmentation of the visual scene into contours and surfaces belonging to particular objects. At the cellular level, at several stages in the visual pathway, cells tend to change their response properties when their receptive fields are placed within the artificial scotoma. Assuming that each cell represents a line label signaling, when active, the presence of a feature centered within its receptive field, the response of a cell whose receptive field is centered within the artificial scotoma is interpreted by the visual system as a shift of stimulus features toward its center.
A growing body of evidence reveals a remarkable degree of mutability of function in primary sensory cortex that is not limited to the first months of life but extends throughout adulthood. It is becoming increasingly clear that rather than performing a fixed and stereotyped calculation on their input, cells in all cortical areas represent active filters. Various components of cortical circuitry have been implicated as the likely source of the changes, including intrinsic horizontal connections and feedback connections from higher-order cortical areas. Neural plasticity serves a wide variety of functional roles, and the extent to which it plays a role in the ongoing processing of sensory information depends on the rapidity with which cells can alter their response properties. To date, it has been shown that substantial changes can be induced within seconds of exposure to novel stimuli. It remains to be seen whether even shorter-term modifications of cortical circuits and receptive field properties may underlie the recognition of objects as the eyes move from saccade to saccade.
Bliss, T. V. P., and T. Lomo. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anesthetized rabbit following stimulation of the perforant path. Journal of Physiology 232:331-356.
Gibson, E. J. (1953). Psychol. Bull. 50:401-431.
Gilbert, C. D. (1994). Early perceptual learning. Proc. Natl. Acad. Sci. U.S.A. 91:1195-1197.
Gilbert, C. D., A. Das, M. Ito, M. Kapadia, and G. Westheimer. (1996). Spatial integration and cortical dynamics. Proc. Natl. Acad. Sci U.S.A. 93:615-622.
Hubel, D. H., and T. N. Wiesel. (1977). Functional architecture of macaque monkey visual cortex. Proc. R. Soc. Lond B. Biol. Sci. 198:1-59.
Knudsen, E. I., and M. S. Brainard. (1995). Creating a unified representation of visual and auditory space in the brain. Annu. Rev. Neurosci. 18:19-43.
Linsker, R. (1986). From basic network principles to neural architecture, III: Emergence of orientation columns. Proc. Natl. Acad. Sci. U.S.A. 83:8779-8783.
Melville-Jones, G., and A. Gonshor. (1975). Goal-directed flexibility in the vestibulo-ocular reflex arc. In G. Lennerstrand and P. Bach-y-Rita, Eds., Basic Mechanisms of Ocular Motility and Their Clinical Implications. New York: Pergamon Press.
Merzenich, M. M., and K. Sameshima. (1993). Cortical plasticity and memory. Current Opinion in Neurobiology 3:187-196.
O'Keefe, J. (1976). Place units in the hippocampus of the freely moving rat. Expt. Neurol. 51:78-109.
Ramachandran, V. S. (1993). Filling in gaps in perception: Part 2. Scotomas and phantom limbs. Curr. Dir.Psychol. Sci. 2:56-65.
Squire, L. (1994). Declarative and nondeclarative memory: Multiple brain systems supporting learning and memory. In D. L. Schachter and E. Tulving, Eds., Memory Systems. Cambridge, MA: MIT Press, pp. 203-231.
Weinberger, N. M. (1995). Dynamic regulation of receptive fields and maps in the adult sensory cortex. Annu. Rev. Neurosci. 18:129-158.
Copyright © 1999 Massachusetts Institute of Technology