Age, Experience and the Changing Brain

Age, Experience and the Changing Brain

Neuroscience & Biobehavioral Reviews, Vol. 22, No. 2, pp. 143–159, 1998 䉷 1998 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0149...

959KB Sizes 1 Downloads 35 Views

Neuroscience & Biobehavioral Reviews, Vol. 22, No. 2, pp. 143–159, 1998 䉷 1998 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0149-7634/98/$32.00 + .00

Pergamon

PII: S0149-7634(97)00008-0

Age, Experience and the Changing Brain BRYAN KOLB*, MARGARET FORGIE, ROBBIN GIBB, GRAZYNA GORNY AND SHARON ROWNTREE Department of Psychology, University of Lethbridge, Lethbridge, Alberta T1K 3M4, Canada

KOLB, B., M. FORGIE, R. GIBB, G. GORNY, S. ROWNTREE. Age, experience and the changing brain. NEUROSCI BIOBEHAV REV 22(2) 143–159, 1998.—In this review, various aspects of how environmental experience affects the structure of the cortex at different times in the age of the animal are summarized. The interactions of brain injury and sex on the age-dependent plastic changes in the cortex are also considered. Finally, we have attempted to reach some general conclusions that describe the effects of age, experience, sex, and injury on the cortex. 䉷 1998 Elsevier Science Ltd. All rights reserved. Cortex

Enriched experience

Sex differences

Gonadal hormones

1. INTRODUCTION IN PRINCIPLE, there are two ways that experience could alter the brain: either by modifying existing circuitry or by creating novel circuitry (63). It is reasonable to suppose that the brain makes use of both strategies, although the details of the particular strategy will likely vary with the age of the animal. Indeed, during development of the brain all circuitry is, by definition, novel. One way to examine the experiencedependent changes in the brain is to look at the effects of different experiences on neuronal structure and function. For psychologists, this rationale usually means adopting one of two approaches: either animals are placed in differential environments such as so-called ‘‘enriched environments’’ versus ‘‘impoverished environments’’; or animals are trained in specific types of tasks, such as mazes. In either paradigm, the experience is correlated with some measure of structure such as brain weight or dendritic extent (e.g. (17)). These experiments generally show that particular experiences embellish circuitry relative to the absence of experience, which fails to do so. Although this type of experimental psychological approach would appear to have considerable appeal in understanding experience-dependent changes in the brain, the impact of this type of research has been surprisingly limited. Indeed, Purves (63) noted that for reasons that are as much sociological as scientific, the experimental neuropsychological perspective has not been embraced generally by most neurobiologists and that these psychological experiments are rarely referred to in the mainstream literature. Oddly, again for reasons that are both sociological as much as scientific, the importance of studies of enriched experience also have had limited impact in mainstream psychology where there has been a longstanding bias against structural interpretations of psychological phenomena. Nevertheless, the study of experiencedependent changes in experiments that manipulate external

Cortical development

Aging

Spines

Dendrites

experiences has provided a rich broth of information that is relevant both to basic neurobiological theories of brain function as well as to general theories of behavioral organization. The goal of the current review is to illustrate some of principles that have emerged. The review will begin with a summary of some of experience-dependent changes in the intact brain followed by a consideration of the effects of manipulating factors such as gonadal hormones or neurotrophins and the effects of cortical injury. 2. ASSUMPTIONS As we begin, we must first admit to several biases. Firstly, we assume that the structural properties of the brain are important in understanding its function. Although such an assumption is self-evident to most neuroscientists, it is not as ubiquitously assumed by psychologists who do not study the brain (e.g. (69,83)). An important corollary of this assumption is that changes in the structural properties of the brain reflect changes in the function of neural circuits. Secondly, we assume both that the mechanisms of cortical plasticity are most likely to be found at the synapse and that synaptic changes can be measured by analysis of either pre- or post-synaptic structure. Traditionally, the emphasis in the literature on synaptic plasticity has been upon the presynaptic, or axonal terminal side. For example, in the studies of the effects of unilateral entorhinal cortex in rats, various investigators have shown a major reorganization of the remaining hippocampal afferents (e.g. (85)). Similar inferences have been made in other models, such as in studies of cholinergic outgrowth after cortical injury (e.g. (6)), collateral sprouting after peripheral nerve crush (7), and terminal sprouting after various types of central injuries (e.g. (12)). One difficulty with studying presynaptic changes is that they are very difficult to locate unless one knows a

*Corresponding author. Tel.: 403-329-2405; Fax: 403-329-2555; E-mail [email protected].

143

144

priori where to look. In addition, once found, they are difficult to quantify. The ability to quantify specific morphological features is critical if one is to correlate structural change with behavior. An alternative way to look at synaptic change is to study the postsynaptic or dendritic side. This requires that the complete cell body and dendritic tree be stained, such as in a Golgi-type stain. Since the dendritic surface receives more than 95% of the synapses on a neuron it is, therefore, possible to infer changes in synapse number from measurements of dendritic extent and spine density. One clear advantage of this measure is that one need not know a priori where to look since it is possible to stain, and to examine, the structure of cells throughout the entire brain. In addition, analysis requires only a light microscope (and a lot of time!). Golgi analysis of the postsynaptic side has proven useful in several types of studies of cortical plasticity. For example, various groups have shown that housing animals in ‘‘enriched’’ environments leads to increased dendritic outgrowth (e.g. (17)). Similarly, training animals in specific tasks leads to dendritic changes in specific populations of neurons (e.g. (21)). One inescapable conclusion of postsynaptic studies is just how plastic dendritic (and presumably synaptic) structure is. One example is especially intriguing. Purves and Hadley (64) were able to label cells in vivo in the dorsal root ganglion of mice. The dendritic field was then mapped. The same cell was relabelled at different times ranging from a few days to weeks and it was possible to see obvious qualitative changes in dendritic extent, which can be taken as at least suggestive evidence of synaptic plasticity (Fig. 1). Perhaps the most surprising aspect of the Purves and Hadley study was that the dendritic

KOLB ET AL.

morphology was so changeable in absence of any particular training. One could reasonably expect even greater change in an animal that was given special somatosensory-related training or that had peripheral nerve injury. Thirdly, although the emphasis in most studies of structure–function relationships falls upon the analysis of neurons, there are solid grounds for looking at changes in the structure and number of glial cells. Glial cells play an important role in synaptic modification, and thus, can be a clue to the location and nature of experience-dependent changes in neurons and their synapses. Fourthly, it is implicit in the foregoing discussion that changes in the postsynaptic structure will be visible in the light microscope. Although the final verification of the nature of structural modification must be at the ultrastructural, and thus, electron microscopic (EM) level, EM studies are impractical on a large scale as they are time (and money) consuming, even if one knows where to look. Practically, therefore, our studies are carried out in tissue that is stained with a Golgi-type stain (for neurons) or with other specialized histochemical procedures that identify specific proteins, such as glial fibrillary acidic protein (GFAP), in glial cells. Fifthly, the emphasis of our work and this review will be on the cerebral cortex. As psychologists, our primary interest is in cognitive function and it is our assumption that the changes in the cerebral cortex form the principal mechanism for cognitive change. This assumption comes from several lines of evidence. For instance, it is generally agreed that the relative increase in cortical volume across mammalian evolution is associated with increased cognitive capacity. It follows that changes in cognitive functions in a

FIG. 1. Camera lucida reconstructions of portions of the dendritic arbors of five mouse superior cervical ganglion cells visualized at an interval of 3 months. Changes involving both the extension and retraction of particular branches are evident. Open arrowheads show examples of branches that appear to have retracted; asterisks mark examples of branches that have formed de novo in the interval. The black arrows indicate the appearance after the specified number of days. (After Purves and Voyvodic (68).)

AGE, EXPERIENCE AND THE CHANGING BRAIN

particular mammal likely will involve changes in cortical structure or organization. Furthermore, studies of decorticated rats show that although they are capable of a remarkable behavioral repertoire (e.g. (94)), there is limited functional flexibility under conditions that would normally lead to marked functional and/or structural change in intact animals (e.g. (56,96)). Finally, there are marked interspecies differences in the details of cortical organization, such as in Old world and New world monkeys, and it has been assumed that these differences reflect the clear differences in perceptual and cognitive abilities (36). Finally, since most work on cortical plasticity and behavior is done using rats, we have assumed that changes in cortical structures are likely to be largest, and thus, easiest to study in those structures that play a central role in somatosensory and motor function. Rats are nocturnal animals that rely heavily on tactile sensitivity, especially in the representation of the face and snout, and they have well developed forelimb manipulatory abilities. In fact, rats are skilled climbers and jumpers, and are able to use their forelimbs in ways that are strikingly similar to those seen in primates (95). Thus, our enriched environments are organized to allow considerable motor activity and feature many tunnels, the opportunity to climb, and to manipulate objects with both the forelimbs and mouth (Fig. 2). This contrasts somewhat with the extensive experiments of Greenough

145

and his colleagues who have utilized environments that are richer in visual complexity, and thus, these researchers have focused upon analyses of structural changes in the visual system (e.g. (17–19,90–93)). These two approaches are, thus, complementary and will be considered in this way. 3.

HISTORICAL CONTEXT

Although the idea that experience can modify brain structure can probably be traced back at least to Ramon y Cajal (70), it was Hebb who made this a central feature of his neuropsychological theory (25). Hebb did the first experiments on the consequences of enriched rearing on the behavior of the rat (24), but it was not until the group at Berkeley began to demonstrate changes in brain weight, cortical thickness, acetylcholine levels, and dendritic structure that there was any structural correlate of the behavioral changes related to experience (8,9,10,15,74,75). Later, beginning in the 1970s and continuing still, Greenough and his colleagues initiated a multidisciplinary investigation of the effects of rearing animals in visually or motorically enriched environments (Table 1). These experiments demonstrated that not only were there changes in the dendritic structure but that these changes reflected an increase in the number of synapses that could be identified using EM techniques. Evidence that changes in dendritic structure reflect changes in connectivity is important for it implies that it is possible to make inferences about connectivity from dendritic structure. More recently, Purves and his colleagues have demonstrated that there is a direct relationship between dendritic geometry and synaptic input (63–67,73). These authors have shown that the number of inputs to a sympathetic or parasympathetic ganglion in an adult animal is directly related to the number of dendritic branches (Fig. 3). Importantly, however, these experiments have also shown that this relationship is not found in newborn animals. Rather, the number of inputs to a given neuron in young animals is about the same regardless of the

TABLE 1 PRINCIPAL CELLULAR DIFFERENCES IN THE OCCIPITAL CORTEX BETWEEN RATS RAISED IN ENRICHED CONDITION (EC) AND IMPOVERISHED CONDITION (IC)

FIG. 2. Schematic illustration of the rat condominiums used in studies of the effects of enriched experience.

Cellular variable

Environment

Reference

Neuron size Neuron density Dendritic branching Dendritic spine density Number of unmyelinated axons in splenium of corpus callosum Size of unmyelinated axons in splenium of corpus callosum Number of synapses per neuron Size of synaptic contact Synaptic plate perforations

EC ⬎ IC IC ⬎ EC EC ⬎ IC EC ⬎ IC EC ⬎ IC

(9) (91) (92) (15) (35)

EC ⬎ IC

(35)

EC ⬎ IC EC ⬎ IC EC ⬎ IC

(91) (93) (19)

Percentage of total tissue volume Capillary vessels Astrocytic nuclei Oligodendrocytic nuclei Mitochondria

EC EC EC EC

⬎ IC ⬎ IC ⬎ IC ⬎ IC

(2) (80) (80) (80)

After Juraska (32).

146

KOLB ET AL.

FIG. 3. Correlation between input number and postsynaptic geometry among rabbit ciliary ganglion cells, using the number of primary dendrites as an index of the complexity. Each point represents the mean of measurements on a number of neurons ( ⫾ SE). The straight line was fitted to the data by a least squares linear regression program and has a slope of about 1. (After Purves and Hume (65).)

size of the dendritic field. It is reasonable to presume that experience plays an important role in shaping the adult structure but the details of how this might happen are not yet known. The Purves experiments can be generalized to other parts of the brain, such as the cortex, where such experiments are impractical because of the very large number of inputs, which may range up towards 100 000 per neuron. The accumulating evidence showing that experience could modify cortical structure even in adult animals led many investigators to turn to environmental manipulations as a method for influencing recovery from brain injury (e.g. (82,97)). After all, if experience could modify cortical structure and allow enhanced performance in a variety of cognitive tasks, then one would expect that specific environmental experiences could also enhance the reorganization of the brain after cerebral injury. Although some studies have shown clear beneficial effects of such treatments, others have failed to find enhanced functional recovery. Furthermore, it has become apparent that the direct effects of brain injury interact not only with the experience but also with various factors such as age and sex (see below). 4.

distributed. The disadvantage historically has been that the staining is capricious and sometimes obscured with precipitate. In recent years, there have been several modifications to the procedures that have made the staining more consistent. One assumption of Golgi studies is that the staining of neurons is random and does not reflect some specific feature of particular neurons. That is, since only about 1% of cells stain, it is important that those cells that are stained are not unique in some consistent manner. Several studies have now demonstrated the aselectivity of the Golgi procedures (60). We have modified the Golgi-Cox procedure to allow us to stain consistently and with a high signal-to-noise ratio (14). In addition, our procedure allows us to visualize spines clearly and consistently (Fig. 4). The second type of staining procedure is to inject individual cells with dyes that diffuse throughout the cell, including the dendrites and axons (e.g. (61)). The advantage of this procedure is that specific cell populations can be labelled and the issue of stain selectivity is removed, although there is obvious experimenter selectivity. The disadvantages are that it is very time consuming and that it is only practical if one knows a priori where one wishes to look at cells. However, very often in our experiments we wish to look in many regions of the brain, making this procedure impractical. Once the cells are stained, the dendritic length can be measured in several ways. Cells first are drawn using a microscope, which is typically set at 250–400 ⫻ magnification. The drawing can be done using some type of computerized imaging system or it can be done by using a camera-lucida procedure (Fig. 5) in which cells are drawn with pen and ink. The advantage of the computerized systems is that the precise length of all dendritic segments

ANALYSIS OF DENDRITIC MATERIAL

We indicated earlier that there is a high correlation between the extent of dendritic arborization and the number of synapses. Although synapses can only be studied (and counted) directly with EM procedures, an estimate of synapse number can be made by calculating the total dendritic length. Cells can be stained using one of two different procedures. The oldest procedure was devised by Golgi in the late 19th century and it involves depositing a heavy metal, such as gold or silver, on cell bodies. A more recent recipe, known as Golgi-Cox, uses mercury. The Golgi stains have the advantage that they stain about 1% of the cells in the tissue and these cells are thought to be randomly

FIG. 4. Photograph of a layer-V pyramidal neuron from the parietal cortex. The cell is stained with a variant of the Golgi-Cox procedure. Dendritic spines are clearly visible.

AGE, EXPERIENCE AND THE CHANGING BRAIN

FIG. 5. Illustration of the camera lucida method for tracing neurons. Mirrors in the drawing tube allow the drawer to visualize both the cell through the microscope as well as the tracing page. Cells, therefore, can be traced.

can be calculated and various statistical measurements can be made (e.g. (4)). The disadvantage is that these semi-automated procedures are very slow and, somewhat paradoxically, are labor-intensive. Although only 1% of the neurons are stained in Golgi-type stains, the cells are still close together and dendritic branches from different cells overlap. The human eye can easily distinguish which branches belong to which cells, but computers cannot yet

147

do so. This means that an operator must guide all of the computer drawing. If cells are drawn using pen and ink, then the analysis is normally done by using a procedure that estimates dendritic length. One way is to count the total number of dendritic branches, whereas another way is to place some sort of grid over the drawing and to count the number of intersections of the dendrites with the grid lines (Fig. 6). Although the two procedures give very different statistical views of dendritic arborization, both measures lead to the same conclusions. For example, in a study of the effects of ovariectomy on dendritic growth in the cortex of rats, Stewart and Kolb (87) measured dendritic growth using both methods (Fig. 7) and found essentially the same result: removal of the ovaries in adult female rats led to an increase in dendritic arborization relative to intact female controls. It can be seen in the branching analysis that the increased arborization was seen at the higher order branches, meaning that it reflected growth at the end of the existing branches. It can be seen in the Sholl analysis that the branches crossed more of the circles, meaning that the branches were longer in the ovariectomized animals. A second measure of synaptic density is to look at the distribution of dendritic spines on the dendrites (Fig. 8). About 90% of excitatory synapses are made on dendritic spines so a count of the number of spines allows an estimate of the number of synapses (22). Spine density varies on different parts of the neuron (Fig. 8), so it is typical to measure spine density at several places on the neuron. Measurement of spine density requires a higher magnification than dendritic arbor, with most investigators using a

FIG. 6. An illustration of the two main methods of quantifying dendrites. The concentric rings on the left form the grid for the Sholl analysis. The number of ring crossings of each branch gives an estimate of length of dendrite. The numbers on the branches on the right indicate the branch order, which offers a different estimate of dendrite length.

148

KOLB ET AL.

exemplified by pyramidal cells in the cortex. In our initial studies, we looked at cells in cortical layers II, III, and V but since we have found cells in layer II to be less affected by experience or other factors, we have focused on layers III and V in our more recent studies. Although other investigators have focused upon cells in the occipital areas, we have focused more on cells in the somatosensory, motor and prefrontal cortices. Our selection of the somatosensory cortex has been guided by our recognition that rats are very tactile animals, by the ease in identifying the primary somatosensory cortex in animals with cortical injuries, and by the fact that most of our lesion studies have investigated the effects of prefrontal or motor cortex injuries (e.g. (37,38)). Somatosensory cortex changes in response to these injuries and we have been able to investigate the interaction of these changes with various factors such as experience and sex. 5.

FIG. 7. A comparison of data summaries from Sholl and branch order analyses of the same neurons. Adult female rats were ovariectomized (OVX) and layer-III parietal pyramidal neurons were drawn several months later. The branch order analysis shows that all branch orders after the first were increased in the OVX rats. The Sholl analysis shows that the increased branching was near the cell body, falling at about 30–60 m along the branches. (After Stewart and Kolb (86).)

magnification of at least 1000 ⫻ . Dendritic segments are drawn using camera lucida procedures, the segments are measured, and the number of spines per unit length are calculated. One of the difficulties in measuring spine density with standard light microscopy is that spines that lie either on the backside of the dendrite or spines that are pointed directly up at the viewer cannot be seen. Thus, measurements of spine density with the light microscope will necessarily underestimate the total number of spines. The newly developed confocal microscopes overcome this difficulty, and thus, yield higher estimates of spine density. One of the decisions that must be made in measuring dendrites is which cells to measure. In any slice of brain there will be multiple cell types (e.g. pyramidal versus granule cells). In addition, there is the issue of which brain regions to draw. In our experiments, we have chosen to draw excitatory neurons, which are spiny and are best

ANALYSIS OF GLIA

There are three main types of glia in the cortex: astrocytes, microglia, and oligodendroctyes. The oligodendroctyes form the myelin and have not been studied with respect to experience. The astrocytes are large glia that are found both in the white matter as well as the grey matter of the cortex. Many astrocytes have processes that resemble dendrites and these processes expand in response to various events, including experience. Microglia are normally visible only when the brain is injured and they function as macrophages. Both astrocytes and microglia can be visualized by using immunohistochemical procedures. Histological sections are exposed to antibodies to different proteins expressed by the cells. For example, astrocytes express proteins such as glial fibrillary protein (GFAP) or vimentin and with appropriate reactions the cells can be visualized (Fig. 9). Similarly, microglia can be located with an antibody known as OX-42. Once visualized, the glia can be quantified in two different ways. Astrocytes are best identified by using procedures analogous to the grid crossing systems used for dendrites. In this case, the cells are not drawn but rather a video camera is attached to a microscope, and under relatively high power magnification (200–400 ⫻ ), the cells are visualized on a monitor that has a grid superimposed. The number of astrocytic branch crossings over the grids are counted. When astrocytes are measured in this manner, Greenough and his colleagues have found that there is a consistent increase in astrocytic branching in the brains of animals housed in enriched environments (23,81). The simplest way to measure microglia is to use an imaging system to measure the density of staining in different cortical regions. Microglia are small and it is impractical to count individual cells. Rather, it is more common simply to measure the density of staining, and to infer that there are larger numbers of microglia present. Microglia are not usually present in large numbers unless the brain is injured or if cells are dying for some other reason, such as in early development or in dementia. Thus, they have not been systematically studied in studies of experience-dependent changes. There are studies, however, showing that the density of microglia can be affected by experience. For example, we have found that environmental stimulation in infancy will reduce the density of microglia observed both during infancy and in adulthood (47).

AGE, EXPERIENCE AND THE CHANGING BRAIN

149

FIG. 8. Illustrations of representative layer-III parietal pyramidal neurons from a rat placed in an enriched environment at weaning versus a littermate that was housed in standard laboratory housing. The dendritic branches down the midline are expanded views of terminal (T), oblique (O), and basilar (B) portions, illustrating the dendritic spines. Note that the spine density varies with location on the dendritic tree. Enriched housing produced an increase in branching but a decrease in spine density.

6.

EXPERIENCE-DEPENDENT CHANGE IN THE INTACT BRAIN

As we began our studies of experience-dependent changes in the brain, we used the logic of those before us who had compared animals in laboratory cages to others in ‘‘enriched environments’’ (e.g. (25,18,74)). Thus, we placed animals in same-sex groups of 4–6 in rat condominiums, which are 1 m ⫻ 0.6 m ⫻ 1.8 m high enclosures (Fig. 2). The condominiums feature sawdust floors to allow digging and three hardware cloth walls to allow climbing. The enclosures contain numerous objects, tree branches, and pieces of PVC pipe, and the objects are moved about weekly. The animals are allowed to live in this environment for about 90 days, at which time their brains and behavior are compared to littermate controls that were group-housed in standard hanging cages. In adulthood, the animals housed in the condominiums typically have a decrease in body weight, presumably because they are more active, and an increase in brain weight of about 6%. Although this increased brain weight likely reflects changes throughout the brain, there are large effects upon cortical thickness, with a typical increase in thickness the order of 7%. 6.1. Age, experience and dendritic changes

FIG. 9. Photomicrographs illustrating staining of by antibodies specific for basic fibroblast growth factor (bFCF) and glial fibrillary acidic protein (GFAP).

It has long been assumed in the psychological literature that experiences in early childhood have greater effects upon later behavior than do similar experiences in adulthood. Our analysis of dendritic changes following exposure to enriched environments suggests that there is a structural basis to this differential effect of early experience on behavior. We placed rats in enriched environments for 3 months beginning either at weaning (21 days of age), at young adulthood (4 months of age), or in senescence (2 years or older). The principal finding was that the age at which animals were placed in the enriched environments

150

has qualitatively different effects upon dendritic structure. Rats placed into the condominiums in young adulthood showed effects similar to those reported by others: there was a large increase in dendritic length relative to cagehoused control animals (Fig. 10). In addition, there was an increase in spine density (e.g. (45)). Parallel results were seen in senescent animals, as they showed increases in dendritic length and spine density relative to age-matched control rats (45). In contrast, when we analyzed the changes in animals who were placed in the condominiums as juveniles, we saw an increase in dendritic branching but a consistent decrease in spine density. That is, the young animals showed a qualitatively different change in the distribution of synapses on pyramidal neurons compared to older animals. The differential effect of enrichment in the young versus older animals led us to look at the effects of environmental manipulation even earlier in the animals’ lives. It has been shown that tactile stimulation of premature human babies with a brush leads to faster growth and earlier hospital discharge (11,79,84). In addition, studies in infant rats have shown that similar treatment alters the structure of olfactory

KOLB ET AL.

bulb neurons and has effects on later behavior (e.g. (6,57,88,98)). We therefore stroked infant rats with a camel hair paintbrush three times daily from day 7 to day 21 of life. Animals were subsequently raised in standard lab cages and were sacrificed in adulthood. Golgi analysis revealed that the early experience had no effect upon dendritic length in adulthood but there was a significant drop in spine density (45). These results surprised us and led us to consider other changes in cortical morphology in the tactile-stroking paradigm. For example, in one experiment infant rats were given tactile stimulation and then were sacrificed at different ages. At sacrifice we measured: (1) the density of acetylcholinesterase staining, which allowed an indirect measure of acetylcholine innervation in the cortex; and 2) the density of immunohistochemical staining for OX-42, which is a marker of microglia; and 3) the density of staining for GFAP, which is a marker for astrocytes. The results showed that acetylcholinesterase levels were significantly increased after only 7 days of stimulation and that this increase was still present 6 weeks after the stimulation was ended. Similarly, the number of microglia was significantly decreased after only 7 days of stimulation and remained lower than in controls 6 weeks later. Surprisingly, GFAP-positive astrocytes were virtually unchanged by the experience. These results lead us to several conclusions. First, ‘‘enriched’’ experience can have very different effects upon the brain at different ages. Second, experience not only leads to ‘‘more’’ but can also lead to ‘‘less’’. That is, although there is a temptation to presume that experiences lead to increased numbers of synapses and probably to increases in glia, it appears that there may be either increases or decreases, the details varying with age at experience. Third, changes in dendritic length and dendritic spine density are clearly dissociable. It is not immediately clear what the differences mean in terms of neuronal function but it is clear that experience can alter these two measures independently and in different ways at different ages. 6.2. Task-dependent changes in dendritic arbor

FIG. 10. Summary of the effects of housing in condominiums beginning at weaning (young), young adulthood (adult) or in old age (old). Enriched housing led to an increase in dendritic branching in all groups. In contrast, spine density was decreased in the young group and increased in the adult and old groups. (After Kolb and Gibb (45).)

Greenough and his colleagues probably were the first to look at the effects of training in specific tasks on the structure of cortical neurons. In their first studies, they trained rats in visual mazes and then analyzed the structure of neurons in the visual cortex (e.g. (17–20,91,92)). Their principal finding was that both pyramidal and stellate neurons had significantly increased dendritic arbor in the brains of trained relative to untrained control rats. In fact, in one experiment the investigators took advantage of the fact that the visual projections from the eyes to the visual cortex of the rat are nearly completely crossed. Thus, input from one eye goes largely to the contralateral hemisphere. The authors therefore exposed one hemisphere to a visual task by occluding the ipsilateral eye, and subsequently only one hemisphere had the visual experience (5). As might be expected, there was an effect of training on the visually trained hemisphere but not on the untrained hemisphere. Greenough and colleagues (21) subsequently trained animals to reach for food with a single forelimb and showed growth in pyramidal neurons in layer III and V, but not II, in the motor cortex. More recently, we trained rats

AGE, EXPERIENCE AND THE CHANGING BRAIN

either to reach with a single forelimb to obtain food or to pull up strings with both forelimbs to retrieve food attached to the end (54). We found that there was a significant increase in dendritic branching in layer V pyramidal neurons in the hemisphere controlling the reaching forelimb versus the nonreaching forelimb in the unimanual task. In contrast, neurons in both hemispheres showed increased growth in the bimanual task. These changes were specific to the forelimb region of the motor cortex and were not observed in motor cortex neurons located more laterally. We also measured spine density in these animals and found no changes in any of the treatment groups. Thus, once again it appears that dendritic branch length and spine density are independent processes. 6.3. Experience and the aging brain There have been few systematic studies of experiencedependent changes in the aging brain. One difficulty in putting senescent rats that have lived their lives in laboratory cages into enriched environments is that the animals do not interact with the environment in the same manner as younger animals. For example, in our rat condominiums we have observed that whereas young or middle-aged rats explore the entire complex, and spend much time climbing to the upper levels, old rats do not appear to have the motoric abilities (or inclination?) to explore the upper levels and stay on the ground level. In one experiment we were able to identify ‘‘climbers’’ and ‘‘nonclimbers’’ and showed that climbers had significantly larger brains (45). This observation is important as it demonstrates that mere passive exposure to experiences is not sufficient to ensure neuronal change: there must be active interaction with the environment. This study also underscores the importance of observing the behavior of animals in specific environments for it may account for some of the individual variation in experience-induced brain changes. One interesting result in our study was seen in the comparison of the old adult animals and young adult animals with the same lab-cage housing. The old animals had much simpler cells than did the younger animals. Curiously, a comparison of the enriched old rats and the young adults showed that the enriched housing returned the cells of the old animals to the level of younger impoverished animals. It would be interesting to determine if a longer period of housing might have stimulated growth similar to that seen in the younger enriched rats. Furthermore, it would be interesting to determine if a period of enriched housing during young adulthood or even middle age might prevent the atrophy of the neurons in the aged lab-reared rat. 7.

EXPERIENCE AND PLASTICITY IN THE INJURED ADULT BRAIN

7.1. Plasticity in the injured adult brain When the cortex is damaged there are changes in the remaining cortex that are correlated with functional outcome. For example, when we removed the frontal cortex in adult rats, we found an initial drop in dendritic arborization in proximal cortical regions such as parietal cortex. This atrophy slowly resolved and 4 months later there was a significant increase in dendritic morphology, which was correlated with partial restitution of function (e.g.

151

(39,40,77)). This morphological plasticity may be related to lesion size or locus, however, as we have also found that large unilateral devascularizing lesions including portions of motor, parietal and anterior visual cortex lead to a permanent atrophy of remaining cortical neurons (41). These animals showed little restitution of function, which would be consistent with the neuronal atrophy. Two additional factors may be important in understanding neuronal changes after cortical injury in adulthood. First, in an innovative series of experiments, Jones and Schallert (28,29) demonstrated that rats with unilateral lesions of the forelimb region showed compensatory changes in behavior (they favored the limb ipsilateral to the lesion) and that these behavioral changes were correlated with dendritic growth in the normal hemisphere. When we replicated these experiments we failed to find the same changes in the normal hemisphere (13,61) but one fundamental difference was in the nature of the brain damage. In our experiments, we surgically removed tissue whereas in the Jones and Schallert experiments, tissue was damaged in such a manner as to lead to slow neuronal death. It now appears that slow death of neurons leads to changes in mRNA in neurons in the intact hemisphere that are different from what is observed after rapid death of neurons (59). Thus, a second factor affecting injury-induced neuronal plasticity is the etiology of the disease process. We have already noted the possibility that changes in glia may correlate with changes in neuron morphology. One consistent effect of cortical injury is a marked proliferation of astrocytes in both the grey and white matter proximal to the injury. For example, when we made lesions to motor cortex, we observed a marked increase in the number of reactive astrocytes in the adjacent motor cortex, as well as more distal ipsilateral cortical regions, but there was no detectable change in contralateral, uninjured hemisphere (11,77). Astrocytes also express other proteins that are only observed after brain injury. For example, in postmortem immunohistochemical studies astrocytes show virtually no reaction to antibodies specific for basic fibroblast growth factor (bFGF) until the brain is injured. Then there is a dramatic expression of bFGF, which peaks 7 days after the injury and declines to basal levels about 2 weeks later. We return to this reaction below. The key point here is that like neurons, astrocytes change after cerebral injury. 7.2. Experience and the injured adult brain 7.2.1. Changes in neurons In view of the plastic nature of the injured brain, it is reasonable to suppose that environmental manipulation might influence both morphological change and functional outcome. Although there have been many studies of the effects of experience on functional outcome after cerebral injury, the results have been inconsistent (for reviews, see (82,97)). One difficulty with these studies is that few have actually measured neuronal morphology but many have focused upon functional outcome with different environmental manipulations. One exception is an experiment in which we made large frontal lesions in rats and then either placed them in our rat condominiums or returned them to their standard laboratory cages (44). After 2 months of recovery and experience, the animals were tested on various

152

behavioral tasks. Our functional results were disappointing as we found only a limited benefit from the special housing in the brain-injured animals. This result made sense, however, when we analyzed the neuronal morphology of the animals. We found that the enriched experience interacted with the endogenous lesion-induced changes in neuronal morphology. Thus, the frontal lesions stimulated an increase in dendritic arborization in parietal cortex but did not affect visual cortex. Enrichment had no additional effect on the parietal neurons but stimulated growth in the occipital neurons. It, therefore, appears that neuronal changes induced by the lesion may place limits upon the environment’s capacity for further neuronal, and subsequently functional, change. Specifically, once having been altered in response to a frontal injury, adjacent parietal neurons may be unable to change further in response to experience. However, it may be that although the total change in dendritic space is similar in the enriched and lab-housed brain-injured animals, there is a qualitative difference in the nature of the changes. Indeed, in view of our observations on the changes in spine density with similar experience at different ages, one might predict differences in this measure to vary with experience after injury. This prediction is particularly germane because we did find a small functional benefit from the enriched experience. Unfortunately, we did not measure spine density in this experiment. (We shall see in Section 7.2.2 that changes in spine density are indeed correlated with functional outcome after enriched experience in animals with lesions in infancy.) There is now a long history of experiments in which damage to the hippocampal formation is associated with deficits in various forms of learning and memory (e.g. (55)). It is, thus, a logical question to ask how injury to the hippocampus might affect experience-dependent changes in the cortex. In principle, it should be easy to answer this question. One need simply make hippocampal lesions and place animals in enriched environments and see what happens. Unfortunately, there is a problem with such a study. Hippocampal lesions produce a variety of behavioral changes so it would be difficult to determine if any reduction in the effects of enrichment on the cortex was due to a direct effect of the hippocampal injury preventing experience-dependent change in the cortex or, alternatively, it was due to a change in the behavior of the animals such that they did not interact with the environment in the same manner as unoperated animals. In order to solve this problem, we made unilateral hippocampal lesions in rats and placed them in enriched environments or standard lab cages (89). It was our expectation that, by using this paradigm, we would control behavioral changes since one would anticipate that both hemispheres in the same animal would have the same experience. Thus, any difference in the effects of experience on the two hemispheres would be a result of the hippocampal lesion on cortical plasticity. Our findings were quite unexpected. First, the intact hemisphere changed as it did in intact rats. That is, there was a growth of dendrites and an increase in spine density. In contrast, however, the hemisphere without a hippocampus showed an increase in dendritic length but a decrease in spine density. That is, this hemisphere behaved like an immature cerebrum. In retrospect, this effect is perfectly sensible: the developing brain has an immature hippocampus which is presumably functioning only at a rudimentary level. Hence,

KOLB ET AL.

injury to the hippocampus mimics, to some extent, the immature state. This result is intriguing and leads us to wonder if lesions to other regions implicated in cortical plasticity, such as the striatum, might produce parallel effects. 7.2.2. Changes in astrocytes and neurotrophins We noted earlier that astrocytes change with experience and with cortical injury. In order to determine if lesion and experience interact, we have measured astrocytic morphology in animals housed in our rat condominiums for a month prior to sustaining unilateral motor cortex lesions (76). As would be anticipated, the animals with enriched rearing showed an increased expression of GFAP but, unexpectedly, there was an interaction between injury, enrichment and astrocyte reactivity. Cortically injured animals with condominium experience had an enhanced

FIG. 11. Summary of the effect of 14 days of enriched housing on the number of GFAP-reactive astrocytes (top) or bFGF-reactive astrocytes (bottom). Rats either received a sham lesion (Control) or a unilateral motor cortex ablation in the forelimb region. The lesion animals were killed 7 or 21 days after the lesion. The lesions and the enrichment both increased the reactivity to both GFAP and bFGF antibodies. Note, however, that control animals showed no reactivity to antibodies for bFGF. (After Rowntree (76).)

AGE, EXPERIENCE AND THE CHANGING BRAIN

153

FIG. 12. Schematic illustration of the effects of infusion of nerve growth factor (NGF) into the lateral ventricle of adult rats. Both the dendritic branching and the spine density were increased in neurons throughout the cerebral cortex. (After Kolb et al. (50).)

expression of GFAP, but even more importantly, these animals also had an enhanced bFGF (basic fibroblast growth factor) reaction even though enriched animals without brain damage did not show this bFGF reaction (Fig. 11). These results suggest that prior experience may affect the nature of the astrocytic response to injury. We can speculate that this is likely to enhance the neuronal plasticity and functional outcome but this has not yet been demonstrated. One of the functions of bFGF appears to be to encourage the brain to produce other proteins, one of which is nerve growth factor (NGF). NGF is known to facilitate recovery from certain types of brain injury. For example, Kolb et al. (41) treated rats, that had sustained large unilateral vascular injuries to the cerebral cortex, with intraventricular infusions of NGF. These animals showed enhanced recovery on motor and cognitive tasks and this recovery was associated with a reversal of atrophy in cortical neurons. Furthermore, and unexpectedly, animals with NGF treatment and no brain injury showed a dramatic growth in dendritic length and spine density in cortical neurons (Fig. 12). This result is intriguing because the animals were trained on various tasks after the NGF infusion. Thus, it is possible that the NGF effects reflected an increased sensitivity of the brain to various experiences. It would be instructive to determine the effects of NGF on condominium experience.

arborization and a decrease in spine density in neurons throughout the cortical mantle (43,49). This result is correlated with a miserable functional outcome and is reminiscent of the marked abnormalities in the brains of retarded children (62) (see Fig. 13). In contrast, when the cortical mantle of rats is damaged in the second week of life, which corresponds to the period of rapid dendritic growth and synaptic formation, there is a generalized enhancement of dendritic arborization and/or spine density throughout the remaining cortex (49,51). This enhanced dendritic response is correlated with dramatic functional recovery. Thus, we see that if the injury in the developing brain leads to increased dendritic space there is a good functional outcome, whereas if the injury leads to a retarded development of dendritic material, there is a poor functional outcome. It was our expectation, therefore, that if we could potentiate dendritic growth in the animals normally showing poor recovery of function, we would enhance functional recovery.

8. EXPERIENCE AND PLASTICITY IN THE INJURED DEVELOPING BRAIN 8.1. Plasticity in injured developing brain One of our consistent findings over the past decade has been that the anatomical sequelae of cortical injury vary with precise developmental stage. In brief, when the brain of rats is damaged in the first few days of life, which corresponds to a time just after neural proliferation is complete but neural migration and differentiation are still ongoing, there is a marked generalized atrophy of dendritic

FIG. 13. Representative examples of dendritic branches from cortical neurons of children (A) and rats (B). A. The left branch is from a child of normal intelligence, whereas the right branch is from a child with mental retardation. B. The left branch is from an adult rat that sustained a frontal lesion at 10 days of age, whereas the right branch is from an adult rat that sustained a frontal lesion at 1 day of age. The latter rat performed miserably on all behavioral tests administered. Thus, the animals with severely compromised cognitive skills had dendritic branches with few spines. (After Purpura (62) and Kolb (39), respectively)

154

KOLB ET AL.

8.2. Experience and the injured infant brain We saw earlier that the young brain responds to experience differently at different ages, and that it also responds to injury differently at different ages. It is, thus, reasonable to presume that there will be complex interactions between the type of experience, the time of experience and the time of injury. We shall consider these interactions by considering the effects of the earliest injuries (days 1–5) before considering the effects of injuries at days 7–10. Because the animal with a cortical lesion in the first days of life is functionally devastated in adulthood, and because it shows atrophy of cortical neurons, we anticipated that such animals would benefit the most from early experience. Animals were given frontal or posterior parietal lesions at 4 days of age, followed by tactile stimulation (stroking) until weaning. They were group-housed in laboratory cages and then tested on various tasks sensitive to frontal or parietal injury (47). The rats with tactile stimulation showed an unexpectedly large attenuation of the behavioral deficits of cerebral injury as a result of this rather brief environmental ‘‘therapy’’. Analysis of the brains showed a reversal of the atrophy of cortical neurons normally associated with such early lesions, and more interestingly, a reversal of the decrease in spine density that is normally associated with the tactile stimulation (Table 2). In other words stroking leads to a decrease in spine density in normal animals but to an increase in spine density in the lesion animals. Thus, it is clear that environmental events can have different effects on the normal and the injured brain. We are left, of course, with the question, ‘‘why?’’ One possibility is that the lesion differentially affects the production of something, such as astrocytes and/or growth factors, and these, in turn, influence the effect of experience on the remaining brain. Another possibility is that behavior has been changed, and the change in behavior alters the experience-dependent effects. Although this latter explanation has some appeal in the older animal, it seems unlikely in the neonate that has so little behavior other than feeding and sleeping. In another series of experiments, we placed animals, who had lesions on postnatal day 4, in our rat condominiums for 3 months, beginning at the time of weaning (42,47). Once again, there was a reversal of the behavioral impairments and this was, again, correlated with a reversal of the dendritic atrophy. As with the tactile stimulation, there was also a reversal of the decreased spine density in the normal animal that was associated with the recovery. We have not yet examined the effects of condominium housing in the adult animal with a perinatal cortical injury, but our

prediction is that we would still get improved functional outcome. In this case, however, we would predict enhanced dendritic growth and increased spine density, which would be parallel to effects of enrichment in the control animals. Testing this prediction is important for it would show that behavioral therapy consistently improves behavioral outcome by using a consistent anatomical change, namely growth of synaptic space, even though in the normal brain, the response to experience appears to vary with age. Although we have lumped together the effects of lesions over the first few days, there is a large behavioral difference: lesions on day 1 are far more debilitating than comparable lesions on day 5. Similarly, we have found that the effect of experience varies depending on whether the lesion was on day 1 or 5: animals with injuries on day 1 show a much greater behavioral response to the experience than do animals with lesions on day 5. Furthermore, although we did not measure dendritic changes, we did find that there was an increase in cortical thickness in the younger operates that was more than twice as large as in the older operates, which is consistent with the behavioral result (42). It, thus, appears that the youngest animal with the largest behavioral impairment is most influenced by the experience. Presumably the experience at least partially reverses the dendritic atrophy that lead to the dismal behavioral outcome; the greater the atrophy, the greater the potential recovery. This result speaks directly to the importance of initiating behavioral therapies in children with cerebral injuries late in gestation or at birth for these are the children with the worst functional outcomes after cerebral injury. Now let us consider the brain of the animal with good functional recovery after injury in the second week of life. Our first experiments looked at animals with frontal lesions that were placed in condominiums at weaning (47). These animals showed a much reduced enhancement of behavioral recovery, presumably, in part, because they were not as impaired to begin with. Indeed, on some behavioral tasks there was really no improvement at all with the behavioral treatment. When we analyzed the brains, we found an unexpected result. Recall that normal animals show an increase in branching and a decrease in spine density when they are placed in the condominiums at weaning. Furthermore, animals with day-7 frontal lesions show no change in branching but an increase in spine density, which is associated with a good behavioral outcome. Enrichment in day-7 animals produced an increase in dendritic branching

TABLE 3 SUMMARY OF THE EFFECTS OF ENRICHMENT BEGINNING AT WEANING

TABLE 2 SUMMARY OF THE EFFECTS OF TACTILE STIMULATION ON BASILAR SPINE DENSITY Group

Non-stroked

Stroked

Control Frontal

7.8 ⫾ .1 7.8 ⫾ .1

7.3 ⫾ .1* 7.4 ⫾ .1*

After Kolb et al. (48). Numbers represent mean spines per 10 m ( ⫾ SE) for layer-III pyramidal cells in parietal cortex. *Differs significantly from nonstroked group.

Group A. Apical branches Control Frontal B. Apical spines Control Frontal

Isolated

Enriched

28.8 ⫾ .8 27.0 ⫾ .8

32.1 ⫾ .7* 30.4 ⫾ .8*

5.5 ⫾ .1 5.8 ⫾ .1

5.2 ⫾ .1* 5.2 ⫾ .1*

After Kolb et al. (47). Numbers represent mean branches per cell or spines per 10 m ( ⫾ SE) for layer-III pyramidal cells in parietal cortex. *Differs significantly from isolated group.

AGE, EXPERIENCE AND THE CHANGING BRAIN

but a decrease in spine density relative to cage-reared animals (Table 3). In other words, an increase in spine density is associated with recovery after day-7 lesions but this effect is reversed with enrichment! It appears that the brain’s response to the injury is altered fundamentally by the enrichment, even though the behavioral compensation is only minimally affected. It is unclear what this may mean to the animal but one prediction is that, unlike the adult braininjured rat who could not easily change its brain with experience (see above), the infant brain-injured rat may still be capable of considerable change. For example, we would predict that there would be greater changes in response to specific experiences (e.g. training to reach or pull strings) in the infant versus adult injured animal. To sum up, the effects of experience on the injured brain are complex and vary with precise age at injury as well as the time of onset of experience. Perhaps the most important message is that the infant with the most miserable functional outcome is especially helped by behavioral therapy. This is an important lesson for treatment of children with perinatal brain injuries. 9.

SEX AND EXPERIENCE-DEPENDENT CHANGES IN THE BRAIN

There is accumulating evidence that the male and female brain differ in their structure, respond differently to environmental events, and respond differently to injury. We consider each aspect in turn. 9.1. Structural differences in the male and female brain The first place to look for structural differences in the brains of males and females is in those structures controlling behaviors such as sexual behavior or maternal behavior in mammals or singing in birds. Sexual behavior is known to be highly dependent upon the hypothalamus and there is now considerable evidence that the preoptic area of rodents and primates is sexually dimorphic (3,16). Similarly, there is substantial literature documenting sex-related differences in structures related to singing in the song bird. Thus, there is now little doubt that sexually dimorphic behaviors are related to sexually dimorphic structure in the brain regions of mammals and birds that are related in some way to reproductive behavior. What is more interesting, however, is the more recent evidence that there are sex-related differences in cortical structures that are not known to be related to the control of behaviors related to mating. For example, morphometric analyses of the male and female brain have shown differences in neuron numbers in different cortical regions (e.g. (71,72)), as well as in dendritic structure (32). Recently, Seymoure and Juraska (78) reported that there may be sex differences in visual perceptual abilities of male and female rats. In particular, although they found no differences in visual acuity, they did find that males had more acute vernier acuity than females, a result that has, at least, circumstantial correlation with the morphological difference in the visual cortex of male and female rats. In our own work, we have found differences in prefrontal and parietal regions as well (e.g. (53,86,87)). The sex differences in the prefrontal regions are intriguing because we find greater dendritic arborizations in the midline prefrontal regions in males and, in contrast, greater dendritic

155

arborizations in the orbital prefrontal regions in females (52). Thus, these sex differences are not evidence of more synapses in one sex than the other, but rather they suggest a fundamental difference in the organization of the prefrontal regions in males and females. Furthermore, we have found sex differences in the effects of lesions to these regions, a result that would be consistent with the anatomical results (e.g. (40)). One explanation for the sex-related differences in the prefrontal areas is that the presence of gonadal hormones during development fundamentally alters the structure of the cells. We have tested this hypothesis directly by looking at the effects of neonatal gonadectomy or treatment with testosterone on cortical morphology and have found that hormone manipulations do indeed alter the neuronal structure in the different cortical regions (e.g. (86,87)). Recently, Jacobs et al. (27) have described a sex difference in their quantitative dendritic analysis of Wernicke’s area, which is a language-related area, in humans. Their finding was that females have more extensive dendritic arbors than males. This result is consistent with the reliable finding that females are superior in certain verbal skills than males (for a review see (55)). Furthermore, in a subsequent study, Jacobs and Scheibel (26) found that this sex difference was present as early as age 9, suggesting that such sex differences emerge within the first decade. These sex differences in cortical architecture in humans are parallel to those reported in other studies showing sex differences in cerebral blood flow and glucose metabolism with females having about at 15% higher level than males (e.g. (1,73)). The importance of hormones to cortical structure is not restricted to development. When Stewart and Kolb (87) ovariectomized or gonadectomized adult rats they found a significant change in cortical structure, especially in the females. Thus, the brains of the ovariectomized rats not only grew heavier but the cortical neurons showed an 25% increase in dendritic arbor and an 10% increase in spine density (Table 4). Males showed only an increase in spine density. This result is surprising as it implies that cortical morphology is hormone-dependent throughout the life of a mammal. The finding of hormone-dependent changes in cortical structure in the adult animal is corroborated by more recent studies of astrocytes. To summarize, there are sex-dependent differences in neuronal architecture in the cortex. These differences are due, in part, to the effects of gonadal hormones during development but, especially in females, there is evidence that circulating hormones alter cortical architecture in

TABLE 4 SUMMARY OF THE EFFECTS OF OVARIECTOMY OR CASTRATION IN ADULTHOOD Group

Branches

Terminal spines

Secondary spines

Male Castrated Female Ovariectomy

30.3 28.0 24.8 31.6

5.3 ⫾ .2 6.0 ⫾ 3 5.7 ⫾ .1 5.9 ⫾ .1

8.6 9.4 9.1 9.9

⫾ .2 ⫾ .1 ⫾ .2 ⫾ .2

After Stewart and Kolb (87). Numbers represent mean branches per cell or spines per 10 m ( ⫾ SE) for apical field of layer-III pyramidal cells in parietal cortex.

156

KOLB ET AL.

adulthood as well. Finally, although it has not been studied in detail, it also seems likely that there are sex-related differences in astrocytic numbers and morphology in the cortex as well. 9.2. Sex, experience, and cortical structure Juraska (e.g. (30–35)) was the first to report that neurons in the occipital cortex of the male rat show significant dendritic growth with 30 days of enriched experience, whereas neurons in the occipital cortex of the female rat do not. In contrast, she found that hippocampal neurons show a greater response to the enriched rearing in the female brain than in the male brain. Although we found these experiments intriguing, they were also puzzling since it seemed reasonable to suppose that experience was altering the cortex of both males and females. We hypothesized that perhaps Juraska’s sex differences were at least partly due to differences in the sensitivity of the cortex to experience. In other words, we wondered what would happen if animals were given more prolonged exposure to the enriched environments. We therefore placed adult male and female rats in condominiums for 5 months. Analysis of the pyramidal cells in layer III of both parietal and occipital cortex revealed an unexpected result: Although the dendritic arborization of both males and females increased with 5 months of living in the condos, the increases were greater in females. Thus, in contrast to Juraska’s studies of visual cortex after 30 days of enrichment, we found a larger increase in dendritic arborization in the cells in female cortex than in male cortex (Table 5). Somewhat surprisingly, however, when we examined the spine density, we found a larger increase in spine density in males than in females (46,58). These results lead us to two conclusions. First, both males and females show experience dependent changes in the pyramidal neurons of the cortex. It may be, however, that males are more sensitive than females and show morphological changes sooner than females. If so, this may account for the apparent difference between the results of Juraska and the results of our experiments. Second, there is a sex difference in the details of the structural changes in cortical neurons: males show a greater increase in spine density, whereas females show a greater increase in dendritic length. The next question we considered was whether age at enrichment would interact with sex. We had shown previously that age at the onset of enriched rearing differentially affected spine density (see Section 6.1) but we had

TABLE 5 SUMMARY OF THE EFFECTS OF 5 MONTHS EXPERIENCE IN THE CONDOMINIUMS Group

Male Female

Isolated

Enriched

Apical

Basilar

Apical

Basilar

29.5 26.3

40.6 38.9

30.7 33.5*

51.4* 53.3*

After Kolb et al. (46). Numbers represent mean branches per cell for layer-III pyramidal cells in parietal cortex. *Significantly different from same-sex isolated group.

not considered the role of sex. We therefore reanalyzed our studies in which juvenile rats were placed in the condominiums or infant rats were given tactile stimulation (46). The results showed that the sex-dependent differences in the effects of enrichment varied with age. Firstly, we found that females showed a greater response to tactile stimulation than males. Hence, females showed a greater decrease in spine density than males. Secondly, we found that both sexes showed an increase in dendritic arborization and a decrease in spine density when they were placed in the condominiums at weaning. Thus, it appears that sex differences in the effect of experience are not only influenced by the length of the enrichment experience but that the details of the environment-responsive changes vary with age. 10.

CONCLUSIONS

One of the most intriguing questions in behavioral neuroscience concerns the manner in which the brain, and especially the neocortex, can modify its structure and ultimately its function throughout one’s lifetime. As the review has suggested, the cortex can be changed dramatically by experience and this change is modulated by various factors. Several basic conclusions can be extracted regarding the nature of the relationship between experience, brain plasticity, and behavior. 1. Experience alters the synaptic organization of the cortex. It is possible to visualize with a light microscope the morphological changes in the cortex that reflect synaptic modification. These structural changes include modification of both neurons and glia. The neuronal changes can be quantified as increases or decreases in the amount of dendritic arborization as well as in the density of dendritic spines. The glial changes can be measured by calculating the number and the size of astrocytes and the density of microglia. 2. Experience can alter different parts of neurons differently. For instance, alterations in dendritic length can occur independently of changes in dendritic spine density. Environmental stimulation during the juvenile and adolescent period produces increased growth of dendritic length, while at the same time producing a decrease in spine density. The increase in dendritic length would act to increase the number of synapses, whereas the decrease in spine density would act to decrease the total number of synapses. Furthermore, different parts of the neuron can change independently of other parts. Thus, there may be changes in distal parts of the dendritic arbor that are independent of changes in proximal parts. 3. Changes in synaptic organization are correlated with changes in behavior. Animals with extensive dendritic growth relative to untreated animals show facilitated performance on many types of behavioral measures, especially measures of cognitive activity. Such changes are visible not only in laboratory animals but also in humans. For example, Jacobs and Scheibel (26) and Jacobs et al. (27) found a relationship between the extent of dendritic arborization in a cortical language area (Wernicke’s area) and the amount of education. The cortical neurons from the brains of deceased people with a university education had more dendritic arbor than those from people with a high-school education who, in turn, had more dendritic material than those with less than high-school education.

AGE, EXPERIENCE AND THE CHANGING BRAIN

157

4. The effects of experience vary both qualitatively and quantitatively with age. For example, ‘‘enriched’’ experiences during the immediate postnatal period have no measurable effect on dendritic length and lead to a decrease in spine density. Enriched experiences during the juvenile and adolescent period produce increases in dendritic length but decreases in spine density. Similar experiences later in life produce increases in both dendritic length and spine density. All of these experiences have behavioral sequelae in adulthood. The age-dependent plastic changes in the cortex are likely very important and reflect the differential sensitivity of the child’s brain to experience during development. What is not yet known is how different experiences at different times in life might interact and to what extent the absence of a particular experience might be compensated for later in life. 5. There are compensatory plastic changes in the brain following brain injury that are similar in kind to those observed when animals learn from experience. These injury-induced changes are also age-dependent and are thought to reflect the age-related recovery observed after brain injury at different times throughout the lifetime. For example, animals with cortical injuries in the first week of life show decreased dendritic arborization that is associated with a poor functional outcome. In contrast, animals with cortical injuries in the second week of life show increased dendritic arborization and a good functional outcome. The similarity between the plastic changes in the brain in response to injury or experience suggests that there may be basic mechanisms of synaptic change in the mammalian cortex that are used in many forms of synaptic plasticity. This is an encouraging possibility, for it allows hope that we will be able to improve recovery from cerebral injury by taking advantage of the innate mechanisms that the brain uses for other forms of plasticity. Furthermore, to offer a more speculative hypothesis, it may also be possible to reverse the effects of impoverished experience by using pharmacological treatments that are effective in improving recovery from brain injury. 6. Injury-induced changes in cortical structure are modified by experience. Furthermore, brains that show the least change in response to a cortical injury appear to be the most responsive to experience. For instance, animals with neonatal brain injuries show a poor functional outcome and an atrophy of cortical neurons but these animals show dramatic functional recovery and marked synaptic growth in response to environmental manipulations. This is encouraging for it suggests that behavioral therapies should be especially helpful in reversing some of the devastating consequences of brain damage in the latter periods of prenatal development in human infants.

7. Damage to the hippocampus, and perhaps other structures, alters the brain’s response to experience. One effect of this modification is that, at least in the case of hippocampal injury, the adult brain responds to experience as though it were an immature brain. This has important implications for understanding not only the function of the hippocampus in processes related to learning and memory, but also in understanding the behavior of the individual with hippocampal injury. 8. Gonadal hormones influence synaptic organization of the cortex throughout the lifetime of an animal. Males and females have brains that are, at least in some regions, fundamentally different. Experience may have different morphological effects on the brains of males and females, and even a single sex may have different morphological changes in response to the same experience at different times in the life history. For females, this influence of gonadal hormones may have especially important implications for differential synaptic plasticity at different times in the menstrual cycle. Furthermore, the sudden loss of ovarian hormones at menopause will have important implications not only for cognitive abilities during the second half of life but also for the ways in which the brain copes not only with normal cell death during aging but also with injury and disease. 9. Neurotrophins, which may be produced either by neurons or glia, influence the plastic changes in the cortex. More importantly in the current context, the production of neurotrophins is influenced by experience. Thus, one reason why behavioral therapies may be effective in changing the brain is that the behavioral changes brought about by specific experiences actually stimulate the brain to produce neurotrophins. This link between behavior, neurotrophins, and plasticity requires further study, especially in the context of designing therapies for restitution of function. 10. There are limits to the amount that a brain can change. We still do not know what determines the limits, nor in most cases, what the limits might be. Nonetheless, it does appear that when the brain changes ‘‘spontaneously’’ after injury there is a reduction in the plasticity of the affected regions. This implies that behavioral therapies should be initiated early in the postinjury period to ensure that the ‘‘spontaneous’’ changes can be influenced in such a way as to maximize functional recovery.

ACKNOWLEDGEMENTS

This research was supported by an NSERC of Canada grant to Bryan Kolb.

REFERENCES 1. Baxter, L.R. Jr; Mazziotta, J.C.; Phelps, M.E.; Selin, C.E.; Guze, G.H.; Fairbanks, L., Cerebral glucose metabolic rates in normal human females versus normal males. Psychol. Res., 21, 237–245 1987. 2. Black, J.E.; Isaacs, K.R.; Anderson, B.J.; Alcantara, A.A.; Greenough, W.T., Learning causes synaptogenesis, whereas motor activity causes angiogenesis, in cerebellar cortex of adult rats. Proc. Natl. Acad. Sci. USA, 87, 5568–5572 1990. 3. Breedlove, S.M. Sexual differentiation of the brain and behavior. In:

Becker, J.B.; Breedlove, S.M.; Crews, D., eds. Behavioral endocrinology. Cambridge, MA: MIT Press; 1992:39–70. 4. Capowski, J. Computer techniques in neuroanatomy. New York: Plenum Press; 1989. 5. Chang, F.-L.F.; Greenough, W.T., Lateralized effects of monocular training on dendritic branching in adult split-brain rats. Brain Res., 232, 283–292 1982. 6. Coopersmith, R.; Leon, M., Enhanced neural response to familiar olfactory cues. Science, 225, 849–851 1984.

158 7. Cuello, A.C., Trophic factor therapy in the adult CNS: remodeling of injured basalo-cortical neurons. Prog. Brain Res., 100, 213–221 1994. 8. Diamond, J. Nerve growth factor and the reinnervation of skin after peripheral nerve lesions. In: Flohr, H., ed. Post-lesion neural plasticity. Berlin: Springer-Verlag; 1988:35–48. 9. Diamond, M.C.; Lindner, B.; Raymond, A., Extensive cortical depth measurements and neuron size increases in the cortex of environmentally enriched rats. J. Comp. Neurol., 131, 357–364 1967. 10. Diamond, M.C.; Dowling, G.A.; Johnson, R.E., Morphologic cerebral cortical asymmetry in male and female rats. Exp. Neurol., 71, 261–268 1981. 11. Field, T.; Schanberg, S.M.; Scafidi, F.; Bauer, C.R.; Vega-Lahr, N.; Garcia, R.; Nystrom, J.; Kuhn, C.M., Tactile/kinesthetic stimulation effects on preterm neonates. Pediatrics, 77, 654–658 1986. 12. Flohr, H. (ed.). Post-lesion neural plasticity. Berlin: Springer-Verlag, 1988. 13. Forgie, M.L.; Gibb, R.; Kolb, B., Unilateral lesions of the forelimb area of rat motor cortex: lack of evidence for use-dependent neural growth in the undamaged hemisphere. Brain Res., 710, 249–259 1996. 14. Gibb, R., Kolb, B. A method for Vibratome sectioning of Golgi-Cox stained whole brain. J. Neurosci. Methods, 1997, submitted. 15. Globus, A.; Rosenzweig, M.R.; Bennett, E.L.; Diamond, M.C., Effects of differential experience on dendritic spine counts in rat cerebral cortex. J. Comp. Physiol. Psychol., 82, 175–181 1973. 16. Gorski, R.A. Sexual differentiation of the endocrine brain and its control. In: Motta, M., ed. Brain endocrinology, 2nd edn. New York: Raven Press; 1991:71–104. 17. Greenough, W.T.; Black, J.E.; Wallace, C.S., Experience and brain development. Child Dev., 58, 539–559 1987. 18. Greenough, W.T.; Chang, F.F. Plasticity of synapse structure and pattern in the cerebral cortex. In: Peters, A.; Jones, E.G., eds. Cerebral cortex, vol. 7. New York: Plenum Press; 1988:391–440. 19. Greenough, W.T.; West, R.W.; de Voogd, T.J., Sub-synaptic plate perforations: changes with age and experience in the rat. Science, 202, 1096–1098 1978. 20. Greenough, W.T.; Juraska, J.M.; Volkmar, F.R., Maze training effects on dendritic branching in occipital cortex of adult rats. Behav. Neural Biol., 26, 287–297 1979. 21. Greenough, W.T.; Larson, J.R.; Withers, G.S., Effects of unilateral and bilateral training in a reaching task on dendritic branching of neurons in the rat motorsensory forelimb cortex. Behav. Neural Biol., 44, 301–314 1985. 22. Harris, K.M.; Kater, S.B., Dendritic spines: cellular specializations imparting both stability and flexibility to synaptic function. Annu. Rev. Psychol., 17, 341–371 1994. 23. Hawrylak, N.; Greenough, W.T., Monocular deprivation alters the morphology of glial fibrillary acidic protein-immunoreactive astrocytes in the rat visual cortex. Brain Res., 683, 187–199 1995. 24. Hebb, D.O., The effects of early experience on problem solving at maturity. Am. Psychol., 2, 737–745 1947. 25. Hebb, D.O. The organization of behavior. New York: Wiley; 1949. 26. Jacobs, B.; Scheibel, A.B., A quantitative dendritic analysis of Wernicke’s area. I. Lifespan changes. J. Comp. Neurol., 237, 83–96 1993. 27. Jacobs, B.; Schall, M.; Scheibel, A.B., A quantitative dendritic analysis of Wernicke’s area. II. Gender hemispheric, and environmental factors. J. Comp. Neurol., 237, 97–111 1993. 28. Jones, T.A.; Schallert, T., Overgrowth and pruning of dendrites in adult rats recovering from neocortical damage. Brain Res., 581, 156– 160 1992. 29. Jones, T.A.; Schallert, T., Use-dependent growth of pyramidal neurons after neocortical damage. J. Neurosci., 14, 2140–2152 1994. 30. Juraska, J.M., Sex differences in dendritic responses to differential experience in the rat visual cortex. Brain Res., 295, 27–34 1984. 31. Juraska, J.M. Sex differences in developmental plasticity of behavior and the brain. In: Greenough, W.T.; Juraska, J.M., eds. Developmental neuropsychobiology. New York: Academic Press; 1986. 32. Juraska, J.M. The structure of the cerebral cortex: effects of gender and the environment. In: Kolb, B.; Tees, R., eds. The cerebral cortex of the rat. Cambridge, MA: MIT Press; 1990:483–506. 33. Juraska, J.M.; Fitch, J.; Henderson, C.; Rivers, N., Sex differences in the dendritic branching of dentate granule cells following differential experience. Brain Res., 333, 73–80 1985. 34. Juraska, J.M.; Fitch, J.M.; Washburne, D.L., The dendritic mor-

KOLB ET AL.

35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53.

54. 55. 56. 57. 58. 59. 60. 61.

phology of pyramidal neurons in the rat hippocampal CA3 area. II. Effects of gender and experience. Brain Res., 79, 115–121 1989. Juraska, J.M.; Kopcik, J.R., Sex and environmental influences on the size and ultrastructure of the rat corpus callosum. Brain Res., 450, 1–8 1988. Kass, J., The organization of neocortex in mammals: implications for theories of brain function. Annu. Rev. Psychol., 38, 129–151 1987. Kolb, B., Functions of the frontal cortex of the rat: a comparative review. Brain Res. Rev., 8, 65–98 1984. Kolb, B., Animal models for human PFC-related disorders. Prog. Brain Res., 85, 501–519 1990. Kolb, B. Brain plasticity and behavior. Hillsdale, NJ: Lawrence Erlbaum; 1995. Kolb, B.; Cioe, J. Sex-related differences in cortical function after medial frontal lesions in rats. Behav. Neuroscience, 110, 1271–1281 1996. Kolb, B., Cote, S., Ribeiro-da-Silva, A., Cuello, A.C. NGF stimulates recovery of function and dendritic growth after unilateral motor cortex lesions in rats. Neuroscience, 76, 1139–1151 1997. Kolb, B.; Elliott, W., Recovery from early cortical damage in rats. II. Effects of experience on anatomy and behavior following frontal lesions at 1 or 5 days of age. Behav. Brain Res., 26, 47–56 1987. Kolb, B.; Gibb, R., Sparing of function after neonatal frontal lesions correlates with increased cortical dendritic branching: a possible mechanism for the Kennard effect. Behav. Brain Res., 43, 51–56 1991. Kolb, B.; Gibb, R., Environmental enrichment and cortical injury: behavioral and anatomical consequences of frontal cortex lesions in rats. Cerebral Cortex, 1, 189–198 1991. Kolb, B.; Gibb, R. Experience and the changing brain. I. Agedependent changes in dendritic growth following enriched rearing at different ages. 1997; in preparation. Kolb, B.; Gibb, R.; Gorny, G.; Forgie, M. Experience and the changing brain. II. Sex-dependent changes in dendritic growth following enriched rearing at different ages. 1997; in preparation. Kolb, B.; Gibb, R.; Gorny, G. Experience and the changing brain. III. Therapeutic effects of enriched rearing after frontal lesions in infancy. 1997; in preparation. Kolb, B., Gibb, R., Gorny, G. Experience and the changing brain. IV. Therapeutic effects of tactile stimulation after cortical lesions in infancy. 1997; in preparation. Kolb, B.; Gibb, R.; van der Kooy, D., Neonatal frontal cortical lesions in rats alter cortical structure and connectivity. Brain Res., 645, 85–97 1994. Kolb, B.; Gorny, G.; Cote, S.; Ribeiro-da-Silva, A.; Cuello, A.C. Nerve growth factor stimulates growth of cortical pyramidal neurons in young adult rats. Brain Res., 751, 289–294 1997. Kolb, B.; Ladowski, R.; Gibb, R.; Gorny, G. Does dendritic growth underlay recovery from neonatal occipital lesions in rats?. Behav. Brain Res., 77, 125–133 1996. Kolb, B.; Stewart, J., Sex-related differences in dendritic branching of cells in the prefrontal cortex of rats. J. Neuroendocrinol., 3, 95–99 1991. Kolb, B.; Stewart, J., Changes in neonatal gonadal hormonal environment prevent behavioral sparing and alter cortical morphogenesis after early frontal cortex lesions in male and female rats. Behav. Neurosci., 109, 285–294 1995. Kolb, B.; Tomie, J.; Ouellette, A. Bilateral and unilateral motor experience alters dendritic length but not spine density in cortical pyramidal neurons. 1997; in preparation. Kolb, B.; Whishaw, I.Q. Fundamentals of Human Neuropsychology. 4th edn. New York: W.H. Freeman; 1996. Kolb, B.; Whishaw, I.Q; van der Kooy, D. Brain development in the neonatally decorticated rat. Brain Res., 397, 315–326 1986. Leon, M., The neurobiology of filial learning. Annu. Rev. Psychol., 43, 377–398 1992. Munoz-Cueto, J.; Carcia-Segura, L.; Ruiz-Marcos, A., Developmental sex differences and effect of ovariectomy on the number of cortical pyramidal cell dendritic spines. Brain Res., 515, 64–68 1990. Napieralski, J.A., Butler, A.K., Chesselte, M.-F. Anatomical and functional evidence for lesion specific sprouting of corticostriatal input in the adult rat. J. Comp. Neurol, 373, 484–497 1996. Pasternak, J.F.; Woolsey, T.A., On the ‘‘selectivity’’ of the Golgi-Cox method. J. Comp. Neurol., 160, 307–312 1987. Prusky, G.; Whishaw, I.Q., Morphology of identified corticospinal cells in the rat following motor cortex injury: absence of usedependent change. Brain Res., 714, 1–8 1996.

AGE, EXPERIENCE AND THE CHANGING BRAIN 62. Purpura, D.P., Dendritic spine ‘‘dysgenesis’’ and mental retardation. Science, 186, 1126–1128 1974. 63. Purves, D. Neural Activity and the Growth of the Brain. Cambridge: Cambridge University Press; 1994. 64. Purves, D.; Hadley, R.D.; Voyvodic, J.T., Dynamic changes in the dendritic geometry of individual neurons visualized over periods of up to three months in the superior cervical ganglion of living mice. J. Neurosci., 6, 1051–1060 1986. 65. Purves, D.; Hume, R.I., The relation of postsynaptic geometry to the number of presynaptic axons that innervate autonomic ganglion cells. J. Neurosci., 1, 441–452 1981. 66. Purves, D.; Lichtman, J.W., Geometrical differences among homologous neurons in mammals. Science, 228, 298–302 1985. 67. Purves, D.; Rubin, E.; Snider, W.D.; Lichtman, J.W., Relations of animal size to convergence, divergences and neuronal number in peripheral sympathetic pathways. J. Neurosci., 6, 158–163 1986. 68. Purves, D.; Voyvodic, J.T., Imaging mammalian nerve cells and their connections over time in living animals. Trends Neurosci., 10, 398– 404 1987. 69. Pylyshyn, Z.W., Computation and cognition: issues in the foundations of cognitive science. Behav. Brain Sci., 3, 11–69 1980. 70. Ramon y Cajal, S. Degeneration and Regeneration of the Nervous System. London: Oxford University Press; 1928. 71. Reid, S.N.M.; Juraska, J.M., Sex differences in the gross size of the rat neocortex. J. Comp. Neurol., 321, 442–447 1992. 72. Reid, S.N.M.; Juraska, J.M., Sex differences in neuron number in the binocular area of the rat visual cortex. J. Comp. Neurol., 321, 448–455 1992. 73. Rodriguez, G.; Waarkentin, S.; Risberg, J.; Rosandini, G., Sex differences in regional cerebral blood flow. J. Cereb. Blood Flow Metabol., 8, 783–789 1988. 74. Rosenzweig, M.R.; Krech, D.; Bennett, E.L.; Diamond, M., Effects of environmental complexity and training on brain chemistry and anatomy: a replication and extension. J. Comp. Physiol. Psychol., 55, 429–437 1962. 75. Rosenzweig, M.R.; Bennett, E.L. Experiential influences on brain anatomy and brain chemistry in rodents. In: Gottlieb, G., ed. Studies on the Development of Behavior and the Nervous System. New York: Academic Press; 1978:289–387. 76. Rowntree, S. Basic fibroblast growth factor in the injured brain. Unpublished Master’s thesis, University of Lethbridge, Lethbridge, Alta., Canada; 1995. 77. Rowntree, S.; Kolb, B. Antibodies to bFGF block functional recovery and dendritic compensation after motor cortex lesions. Eur. J. Neurosci. 1997, in press. 78. Seymoure, P.; Juraska, J.M. Vernier and grating acuity in adult hooded rats: the influence of sex. Behav. Neurosci., 1997; in press. 79. Schanberg, S.M.; Field, T.M., Sensory deprivation stress and supplemental stimulation in the rat pup and preterm human neonate. Child Dev., 58, 1431–1447 1987. 80. Sirevaag, A.M.; Greenough, W.T., Differential rearing effects on rat visual cortex synapses. III. Neuronal and glial nuclei, boutons, dendrites, and capillaries. Brain Res., 424, 320–332 1987.

159 81. Sirevaag, A.M.; Greenough, W.T., Plasticity of GFAP-immunoreactive astrocyte size and number in visual cortex of rats reared in complex environments. Brain Res., 540, 273–278 1991. 82. Shulkin, J. Preoperative events: their effects on behavior following brain damage. Hillsdale, NJ: Erlbaum; 1989. 83. Skinner, B.F. The behavior of organisms. New York: AppletonCentury-Crofts; 1938. 84. Solkoff, N.; Matuszak, D., Tactile stimulation and behavioral development among low-birthweight infants. Child Psychiatr. Hum. Dev., 6, 33–37 1975. 85. Steward, O. Synapse replacement on cortical neurons following denervation. In: Peters, A.; Jones, E.G., eds. Cerebral cortex, vol. 9. New York: Plenum Press; 1991:81–131. 86. Stewart, J.; Kolb, B., The effects of neonatal gonadectomy and prenatal stress on cortical thickness and asymmetry in rats. Behav. Neural Biol., 49, 344–360 1988. 87. Stewart, J.; Kolb, B., Dendritic branching in cortical pyramidal cells in response to ovariectomy in adult female rats: suppression by neonatal exposure to testosterone. Brain Res., 654, 149–154 1994. 88. Sullivan, R.M.; Leon, M., Early olfactory learning induces an enhanced olfactory bulb response in rats. Dev. Brain Res., 27, 278– 282 1986. 89. Sutherland, R.J.; Kolb, B.; Gibb, R., Experience-induced neocortical dendritic growth is reduced by hippocampal formation damage. Soc. Neurosci. Abstr., 19, 362 1993. 90. Turner, A.M.; Greenough, W.T. Synapses per neuron and synaptic dimensions in occipital cortex of rats reared in complex, social, or isolation housing. Acta Sterol. 2 (Suppl. 1):239-244; 1983. 91. Turner, A.M.; Greenough, W.T., Differential rearing effects on rat visual cortex synapses. I. Synaptic and neuronal density and synapses per neuron. Brain Res., 329, 195–203 1985. 92. Vokmar, R.F.; Greenough, W.T., Rearing complexity affects branching of dendrites in visual cortex of the rat. Science, 176, 1445–1447 1972. 93. West, R.W.; Greenough, W.T., Effect of environmental complexity on cortical synapses of rats. Behav. Biol., 7, 279–284 1972. 94. Whishaw, I.Q. The decorticate rat. In: Kolb, B; Tees, R.C., eds. Cerebral cortex of the rat. Cambridge, MA: MIT Press; 1990:239– 268. 95. Whishaw, I.Q.; Pellis, S.M.; Gorny, B.P.; Pellis, V.C., The impairments in reaching and the movements of compensation in rats with motor cortex lesions: an endpoint, videorecording, and movement notation analysis. Behav. Brain Res., 42, 77–91 1991. 96. Whishaw, I.Q.; Nonneman, A.J.; Kolb, B., Environmental changes can improve grooming, swimming and eating in decorticate rats. J. Comp. Physiol. Psychol., 95, 792–804 1981. 97. Will, B., Kelche, C. Environmental approaches to recovery of function from brain damage: a review of animal studies (1981 to 1991). In: Rose, F.D.; Johnson, D.A., eds. Recovery from brain damage: reflections and directions. New York: Plenum Press; 1992:79–104. 98. Woo, C.C.; Leon, M., Sensitive period for neural and behavioral response development to learned odors. Dev. Brain Res., 36, 309–313 1987.