Book - Brain and behavioural development 7

From Embryology
Revision as of 08:56, 2 July 2018 by Z8600021 (talk | contribs) (Chapter Seven - Determinate and Plastic Principles in Neuropsychological Development)

Chapter Seven - Determinate and Plastic Principles in Neuropsychological Development

Denis M. Parker


To those interested in the relationship between brain mechanisms and behaviour, study of the outcome of damage to the central nervous system currently provides the most useful information concerning the structural basis of cognition and action. Observation of the pattern of behavioural loss and the extent to which recovery is possible following specific brain injury enables differing models of brain organization to be specifically tested. In fact, this question of the pattern of loss and the extent of recovery lies at the heart of a controversy, between the advocates of functional localization and those who proposed a diffuse physical basis for cognitive functions, which began during the nineteenth century. Some investigators stressed the return of almost complete function following a transient period of loss (Flourens, 1824), or stressed the re- emergence of functions at a reduced level while denying that behavioural effects contingent on the damage were specifically related to the region destroyed (Goltz, 1892). These views were amplified during the present century by the experimental work of Lashley (1929) who argued that, excluding the primary sensory and motor areas of the cortex, the association areas contributed in a unified way to the performance of any complex skill—the well-known principle of Mass Action. The degree of functional loss that could be detected following brain damage was assumed to be determined by the extent, rather than the location, of damaged tissue. The results of Lashley’s experiments, together with his theoretical exposition of them, supported the views of Goldstein (1939) who



regarded cognitive skills as a function of the entire organism. He saw the symptoms following brain damage as manifestations of a complex adaptive syndrome in which higher intellectual abilities, in particular the ability to abstract, were likely to be depressed. This general intellectual loss was likely to be seen in the patients’ failure on a variety of cognitive tasks, but particularly in linguistic skills since these demanded, more than other capabilities, the integrity of the entire brain. Holding views such as these, there was no problem in accounting for the return of functions after cerebral injury since the intact areas of the brain would continue to function in their usual way following an initial period of shock, although there would of course be a loss of efficiency. Factors such as the age of the subject at the time of injury were seen to play a part, and indeed one of Lashley’s students (Tsang, 1937) had provided evidence that proportionately equivalent lesions in adult and juvenile rats produced less impairment in the younger animals.

In contrast to models of the brain based on assumptions of global processing and diffuse representation, from the middle nineteenth century onwards evidence began to accumulate rapidly that restricted brain lesions often produced highly specific disorders in which the patients’ remaining spectrum of skills were relatively unaffected. Broca (1861) produced two cases in which the patients’ verbal expression was drastically impaired but whose comprehension of language appeared to be relatively good. Wernicke (1874) produced evidence that patients’ ability to comprehend and utter meaningful linguistic statements could be grossly impaired yet their pronunciation and rate of verbal output could remain unimpaired. Other descriptions followed of patients with restricted and often bizarre symptom complexes. Dejerine’s description (1892) of Mr. C is perhaps one of the classic cases. This patient, who had right visual field blindness, could nevertheless recognize objects and continue his everyday life. Despite apparently normal vision in his left visual field and the ability to speak and write normally, he was unable to read by sight even the sentences he had written himself a short while before. He was however able to ‘read’ by feeling the shapes of cut out letters. Other syndromes described included those of disordered motor planning (Leipmann, 1908), the inability to recognize objects (Lissauer, 1890), and the inability to recognize melody (Henschen, 1926). The list could be continued, but the importance of these cases was that they convinced a substantial number of investigators that the brain could not be organized in a diffuse equipotential manner but, on the contrary, must contain relatively specialized processors even for such uniquely human processes as the comprehension and generation of language.

Modern neuropsychology has moved away from the exclusive study of such rare and extraordinary cases to the systematic examination of groups of patients in which the behavioural effects of differences in known cerebral pathology can be investigated. The more extreme forms of cerebral localization have been



rejected, but the acceptance of specialized processing within and between the two hemispheres of the brain has gained wide acceptance (see Walsh, 1978, and Hecaen and Albert, 1978). This current position regarding the distribution of functional systems within and between the cerebral hemispheres in the adult’s central nervous system is of considerable importance to those interested in neuropsychological development. The relative paucity of information concerning the specific effects of brain damage during development necessitates the use of the effects of adult brain damage as a yardstick against which the outcome of early brain damage may be set. By contrasting the effects of early brain injury with those which occur following damage to the mature brain, it may be possible to ascertain whether the organizational pattern typical of the adult brain is present also in early development or whether it emerges gradually.

It may seem odd that this question should be asked at all, given the fact that, superficially at least, the anatomical structure of the brain is broadly similar in the neonate and the adult, even to the extent of the presence of the adult pattern of anatomical asymmetry (Wada et al, 1975). However, evidence has been presented from time to time which argues that certain crucial functional differences are present. Animal studies have indicated that effects of motor cortex damage in infancy are less deleterious than equivalent damage at maturity since the younger animals escape the flaccid paralysis and spasticity of their elders and retain their postural and locomotor capabilities (Kennard, 1940). Even in humans the claim has been made that children may escape some of the effects of cerebellar injury suffered by adults (Geschwind, 1972). In the area of language it has been known for some time that children who become aphasic following massive left hemisphere injury usually recover speech rapidly (Guttman, 1942), a phenomenon which is not unknown but is certainly rare in the adult (Dejerine and Andre-Thomas, 1912). In such cases it is possible to resort to concepts like plasticity in order to explain the greater resilience of the immature brain, but in some cases the reorganization following injury appears to be so expeditious that it is difficult to believe that other areas of the brain not previously involved have acquired behavioural functions so rapidly (Kennard, 1940; Geschwind, 1972). Even if the idea of greater neural plasticity is accepted, the question may still be asked as to how great this capacity for compensation is? When children become aphasic following unilateral brain damage and language eventually resides in the contralateral hemisphere, does it cope as adequately as the damaged hemisphere would have done? The existence of greater plasticity in the developing brain does not necessarily mean that disparate structures are functionally equivalent. This inference would be valid only if different structures were shown to attain the same degree of functional sophistication.

These two issues concerning the establishment of the adult pattern of neural organization and the extent to which it can be modified following early brain damage will be explored in the following pages.


The context within which damage occurs during development

It is clear, that, as a child grows from birth, the various indices of maturation (behavioural, neurophysiological and morphological) tend to move together towards levels accepted as indicating greater maturity. The range and complexity of spatial, linguistic and social skills increases. In association with them electrophysiological measures of brain activity move from showing a slow and irregular pattern at birth to faster rhythms and widening of the bandwidth characteristics with increasing age. Neuroanatomical changes are also evident both at a gross level, where changes in surface area and the fissural pattern of the brain are seen (Turner, 1948, 1950), and at the microscopic level, where the size of neurones and the complexity of the dendritic pattern increases (Marshall, 1968). When investigators are concerned exclusively with behavioural development, electrophysiological and anatomical factors may provide an interesting but non-essential background. However, where one is concerned with the outcome of brain injury, which occurs early rather than late in the developmental sequence, then the broad neurobiological context in which the injury occurs assumes increasing prominence. A child who suffers brain injury has not just lived for a shorter time, and consequently experienced less, than an adult. The injury has occurred within a system which has a complex and uncompleted developmental plan which is rapidly unfolding, rather than as in the adult where injury occurs within a system where the developmental sequence may have reached a plateau of several decades’ duration. With this in mind some relevant data which outline the changing status of cerebral structures during development will be examined.

Electrophysiological and neuroanatomical maturation

Electrophysiological measures of cerebral development in man indicate a progressive increase in mean frequency content of the EEG from birth to maturity. Whilst in the neonate this is characterized by a labile pattern with periods of almost total electrical silence and an unclearly differentiated sleep pattern, the adult shows a pattern of continuous activity with a frequency spectrum that is at least superficially related to the subject’s state of alertness and with a well-differentiated pattern during the various stages of sleep (see Marshall, 1968, and Milner, 1976, for reviews, and Thompson, this volume). The adult pattern, with the occipital alpha frequency centred in the region of 10 Hz, and the presence of clear beta activity (14-30 Hz) during arousal, is attained only gradually (Henry, 1944). The most typical frequency band shifts from the delta region (1-3 Hz) at age 1 year and below, through the theta band (4-7 Hz) between 2 and 5 years, to the alpha band (8-13 Hz) between 6 years and adulthood. Beta frequencies become increasingly common in late childhood, and



are an increasingly common feature of the adult EEG. Measurement of evoked potentials reveals that peak latencies are long in infancy and decrease as maturation proceeds, whilst in general, amplitude tends to decline after an initial rise. This amplitude trend is also characteristic of the development of the gross EEG (Shagass, 1972). Whilst the behavioural significance of the EEG in man is currently an extremely active research area, the status of the research findings is not always unequivocal. The existence during infancy and childhood of a differential density of waveforms in the left and right hemispheres (Walter, 1950) and the presence of lateral differences in auditory evoked potentials (Molfese, 1977) cannot be taken as evidence of functional asymmetry within the cortex at these ages, since they may indicate changes in the thalamus and corpus striatum, which are simply mirrored in an as yet incompletely differentiated and immature cerebral mantle. The EEG data do, however, allow us to see the gradual emergence of intracerebral rhythms whose mean cycle time becomes shorter with age, and whose responsiveness to external stimulation becomes crisper. Furthermore, this developmental trend continues throughout the period from birth to sexual maturity and may actually show reversal in old age (Shagass, 1972).

The impression of a long-term developmental cycle which emerges from examination of human electrophysiological development perhaps finds even stronger support when neuroanatomical development is examined. It is apparent that although the full complement of neurones is probably present in the central nervous system at birth, the brain continues to grow in overall size and the characteristic sulcal and gyral pattern of the adult emerges gradually over the first 6 years, and perhaps even later in the case of the frontal lobe (Turner, 1948, 1950). When the internal differentiation of the brain is considered a complex and protracted pattern of development unfolds (Yakovlev and Lecours, 1967; Lecours, 1975). Using the density of myelin, the lipid sheath surrounding the axons of neurones, as a criterion of maturation these workers have mapped the development of CNS pathways and structures and used the termination of the myelogenetic cycle as an index of when the system reaches final functional maturity.

It is apparent from this research that brain development does not proceed uniformly in all central nervous subsystems. Sensory fibre tracts and associated nuclei myelinate before the cortical intra- and interhemispheric communication systems. Different sensory systems however may show a diversity of time courses. The optic radiations subserving the visual cortex show a short cycle of myelination closely following the myelination of the optic tract, and is virtually complete by 4 months of age. The acoustic radiation subserving the auditory cortex does not show complete development until the 4th year, in marked contrast to the prethalamic auditory system, which is mature by the 4th postnatal month. The intracortical and interhemispheric association fibres,


which allow communication between different regions and whose disruption leads to many of the bizarre neuropsychological disconnection syndromes (Geschwind, 1965), have prolonged maturational cycles extending into the second decade of life. Examination of the cortex itself indicates that some regions implicated in linguistic competence (inferior parietal lobule) show a slow onset of myelination and that this process may also continue beyond the second decade of life. It should be emphasized that although the study of the myelogenetic cycles of maturation shows a complex extended pattern of development from birth to maturity, a caveat is necessary. Neural conduction can occur in fibres before they become myelinated (Ulett et al , 1944) and indeed myelination may be aided by neural activity (Langworthy, 1933). However, myelin contributes greatly to the efficiency and speed of neural conduction, which underlies the complex analysis and planning characteristics of human cognition and action. The importance of myelin may be appreciated by considering the devastating behavioural effects of demyelination disease in man such as occurs in multiple sclerosis. The myelogenetic analysis emphasizes the manner in which the brain becomes structurally mature. Phylogenetically older structures, in general, mature earlier than those of recent phyletic origin. There are exceptions to this, for example the reticular formation, whose developmental cycle again extends into the second decade of life, partly no doubt because its final operational capability is not required until cortical systems are completely mature.

Both the electrophysiological and neuroanatomical data indicate that the human central nervous system develops over an extended period. Particularly fascinating are the observations that the higher auditory systems (geniculo- temporal pathways) exhibit a developmental cycle that is more than ten times longer than visual structures at the same level of the CNS. It is probable that this reflects the necessity of incorporating into the growing brain the species specific demand of learning the complexities of language via the acoustic mode and allows for the building of the culture-specific phoneme system which is eventually used. However, it is apparent that a considerable part of the developing child’s life is characterized by the presence of cortical structures which are functionally (by electrophysiological and anatomical criteria) immature. The status that these immature structures have in subserving the growing child’s cognitive repertoire is not straightforward. It is possible to envisage the mapping of cognitive growth within the cerebral structures which subserve cognition in the adult in at least two ways. The first view would regard all those structures which are involved in a given functional system in the adult as also being involved in infancy, although the immaturity of the system considered as a whole would limit its capabilities. The limitations seen for instance in perceptual and motor systems in the neonate would be ascribed to the immaturity of the cortical system, which was nevertheless functional, rather



than to the fact that lower but morphologically more mature structures are controlling the behaviour. The alternative view would regard neurobehavioural development as a process beginning with relatively primitive analysis and response control systems residing in phylogenetically older structures which mature earlier in the developmental sequence. As development proceeds, phylogenetically newer structures, which attain maturity later in the developmental sequence, will capture and modulate the activity of the older systems as well as adding more complex control processes. This process will not have a uniform time course across all systems. In the case of the visual system this transition may occur relatively early in development at two to three months of age since the cortical network in this case has a short developmental cycle. In the case of audition the whole process may be more protracted with cortical control becoming ascendant in the period of 2-4 years, whilst in the complex action systems controlled by the frontal lobe, the processes may take even longer, given the differential growth of the region beyond 6 years of age (Turner, 1948, 1950). Evidence that it is this latter view which is the more reasonable description of neurobehavioural development comes from a number of observations that will be subsequently outlined, but it must be said in advance that the case is by no means proven.

Neuropsychological evidence concerning functional maturation

When damage occurs in the mature brain the effects are usually immediate; the degree of functional loss, whether it be sensory, motor or cognitive, is greatest immediately after the lesion and usually shows some remission over time. In the immature brain in a number of instances the pattern is almost the reverse of this. In the case of hemiplegia, sustained as a consequence of prenatal or perinatal injury, the symptoms of the disorder frequently emerge gradually over a prolonged period. Abnormalities in the use of the hand and arm may emerge at between 4 and 6 months, whilst differences in the behaviour of the legs may not become apparent until 10 months of age or more. Deficits in the use of the lower limbs may only become apparent when the child begins to walk, a milestone that is frequently attained in the normal age range (Lyon, 1961). Finally, athetoid movements do not make their appearance until much later, typically between three and four years of age (Lenneberg, 1968). Thus the developmental pattern is seen to move through a series of stages: for example, the grasp reflex is present initially in both the affected and the normal hand, but it persists on the affected side and usually the hand becomes clenched into a tight fist. Individual finger movements do not appear on the hemiplegic side, although movement of the thumbs may be possible. As the child grows, then, the deficit becomes more severe. However, it is noticeable that these children escape some of the more drastic consequences of hemiplegia resulting from cerebral injury at maturity.


The severity of paralysis in the upper limb is less and the lower limb is capable of greater use than after comparable damage in adults. That the residual function on the affected side is not attributable to surviving undamaged tissue in the diseased hemisphere is shown by the fact that removing the diseased hemisphere later in development does not cause further impairment and may frequently result in improvement (Cairns and Davidson, 1951).

Two features then are apparent in these cases. The first is the greater capacity of the young brain to compensate and escape some of the consequences of equivalent injury in adulthood—its plastic propensity. The second feature is cogently described by Lenneberg (1968): ‘one may say that the child with a perinatal cerebral injury only gradually grows into his symptoms’. This emergence of deficit with age is compatible with the view that certain systems are pre-programmed to appear at certain stages in development and injury to them will only become apparent when they fail to appear. Similar affects in the motor sphere were observed by Kennard (1940) in macaque monkeys where ablation of the motor cortex in infancy resulted in surprisingly little immediate effect, but precluded the development of fine manipulatory skills, and this, together with signs of dyskenesia, became more apparent as the animals grew older. It appears that the motor cortex only begins to exert its influence gradually and its loss or malfunction may not become apparent until a later stage of development.

Further support for the notion that behaviour early in development may be mediated exclusively by subcortical structures, with cortical processors becoming involved later, has been provided by Goldman (1976). It has been apparent for some time that tasks which require a monkey to remember the location of a stimulus over a brief interval of time are drastically affected by damage to the dorsolateral prefrontal cortex (Chow and Hutt, 1953). These are known as the delayed-response and delayed-alternation tasks. When the dorsolateral frontal cortex is removed within the first 2 months of life however, the operated animals perform as well on this task as unoperated controls as long as testing is carried out within the first year of life (Harlow et al , 1970). If these monkeys are followed into the second year of life they show evidence of increasing impairment of delayed response tasks. Lesions in subcortical structures which are functionally connected with the dorsolateral prefrontal cortex (dorsomedial nucleus of the thalamus and the head of the caudate nucleus), indicate that the monkeys operated on as juveniles show the same pattern of deficit as adults, i.e. failure on delayed alternation tasks (Goldman, 1974). These results emphasize that not only may the effects of brain damage fail to appear early in development and only reveal themselves as the animal grows, they also show that structures which are required for the adequate performance of a task in adulthood (the dorsolateral prefrontal cortex), are not necessary for the performance of the same task when the animal is an infant or a juvenile.

These experiments contribute a salutary warning to those who would assume



that, because a cortical structure is implicated in a functional system in adulthood it must of necessity be involved in performance of those same functions at an earlier developmental stage. In infant and juvenile macaques subcortical systems are capable of mediating responses which in adults require the participation of the cortex. Furthermore, evidence is available which indicates that the time course over which the effects of cortical lesions become apparent during development differs for different regions. While monkeys aged 2\ months with orbital frontal cortex lesions are equivalent to unoperated controls in the performance of object reversal learning, deficits in performance became apparent by the time the animals are one year old (Goldman, 1974). Thus, while the effects of dorsolateral frontal cortex damage sustained in infancy on delayed alternation tasks do not become apparent until after one year of age, the effects of orbitofrontal damage in infancy on object reversal learning become apparent before one year of age. In the case of both delayed alternation and object reversal learning early in development subcortical structures are capable of mediating the behaviour successfully. With respect to these studies it should be pointed out that macaque monkeys attain sexual maturity between 24 and 30 months of age, so the effects of early brain damage may take a considerable proportion of the developmental cycle before they become apparent.

The evidence cited so far has argued for a model of development in which the complete maturational cycle is prolonged, but within which different systems may attain functional maturity at widely divergent times. Initially, behaviour may be controlled by subcortical systems, and depending on the time-course of the maturational cycle in the higher reaches (cortical) of each system so control will pass to the phylogenetically newer and more adaptive system. Bronson (1974) has argued, on the basis of changes in the pattern of visual behaviour in infants, that in the case of the visual system a transfer of control passes from the superior colliculus to the striate cortex, from the ‘ambient’ to the ‘focal’ system (Trevarthen, 1968), during the 2nd and 3rd postnatal month. Such a view is certainly compatible with the rapid postnatal development of the geniculo- striate system (Yakovlev and Lecours, 1967). In the case of motor function the longer maturational cycle of the medullary pyramids (up to 12 months), which carry the axons of the Betz cells of the motor cortex to the spinal cord and are required for control of individual fingers and skilled sequences (Lawrence and Hopkins, 1972), leads one to expect that transition of control in this case may take place over a more protracted period. The emergence of comparative motor deficit between affected and normal side, in cases in infantile hemiplegia, between 4 and 12 months of age (Lyon, 1961) appears to reflect this longer cycle of development. In man, however, the long-term emergence of a deficit in previously established skills after early brain injury akin to that reported by Goldman (1974, 1976) for the macaque monkey, has not as yet been reported.


Nevertheless, evidence is available which indicates that, on some perceptual tasks, brain-damaged children may show either a consistent difference from controls over a wide age range (Cobrinik, 1959), or a progressive change, which either diverges from or converges to control values between the ages of 5 and 15 years depending on the particular measure being considered (Teuber and Rudel, 1962). Thus, slowly emerging effects of early brain damage in man are not unknown, but long-term effects, with an intervening ‘silent’ interval, have not, to the present author’s knowledge, been described.

Interactive effects and their interpretation

Generalized effects on intelligence that result from presumed structural incapacity (Hebb, 1942) or following localized injury (Thompson, 1978) have indicated that while early injury may produce generalized depression of full- scale IQ, late injury produces a more specific pattern of deficit, with less depression of full-scale IQ. Research findings of this nature can be seen as support either for a model of brain development in which neural structures only gradually attain their mature state, or alternatively, as support for the view that reorganizational capacities are at work during development which allow savings on specific skills but at the cost of overall depression in intellectual attainment. However, there is no doubt that both views are valid and have independent evidence to support them. The developing brain is more vulnerable than the mature brain and this can be seen in the long term generalized deleterious consequences of nutritional deprivation during the brain-growth spurt, as measured by both neurochemical and anatomical criteria (Dobbing, 1968), and by behavioural criteria (Chase, 1973). However, it is also clear that substantial left hemisphere damage, which usually leads to long-term severe dysphasia in adulthood, results in only transitory dysphasic symptoms when the damage occurs early in childhood (Hecaen, 1976), and there is evidence of savings in visuo-spatial skills following early right hemisphere dysfunction, which is not apparent with equivalent damage at maturity (Kohn and Dennis, 1974).

These findings provide unequivocal evidence of the reorganizational and plastic capacities of the developing brain. However, it is also evident that the usual predetermined pattern of left hemisphere specialization for language and right hemisphere specialization for visuo-spatial skills set limits to this plastic capacity. Where one hemisphere has sustained early injury, the language skills exhibited by a remaining and intact right hemisphere are less proficient than those exhibited by a remaining intact left hemisphere (Dennis and Kohn, 1975). It also appears that visuospatial abilities are better developed in an intact right hemisphere than when the left hemisphere alone remains fully functional (Kohn and Dennis, 1974). Thus, while reorganization is possible following early brain injury, this may be more limited than has been previously thought. Indeed, given



the anatomical asymmetry present in the neonatal brain (Wada et al, 1975) it would be surprising if this structural specialization were not detectable at some stage in the behavioural effects of unilateral brain injury. The results of unilateral brain injury in early development nicely juxtapose two major tendencies which emerge from neurobehavioural research: on the one hand, the trend towards carrying through the construction of a preprogrammed system over a protracted time period and, on the other, an adaptive plastic capacity which allows the partial redirection of functional systems should damage occur.

The complexity of the processes involved in the analysis of sensory information (Hubei and Wiesel, 1963; Werner and Whitsell, 1973) that has emerged from single neurone recording makes it clear why a great deal of preprogramming must be involved in the construction of complex adaptive neural systems. It is less easy to describe the mechanisms which allow complex brain systems to exhibit the degree of plasticity that they evidently do in early development. The explanation may reside in the fact that younger individuals are less susceptible to transneural degeneration effects and also show higher levels of biosynthetic activity in brain tissue, which may allow damaged networks greater restitutional capacity (see Goldman and Lewis, 1978, for review). It is also evident that behavioural recovery, seen after early injury in some cases, depends on the ability of young animals to profit from experience in a way not available to the adult. Monkeys sustaining brain injury in infancy show greater recovery the earlier training experience is given in development, although the ability to profit from this experience depends on the nature of the task and the site of the lesion in the brain (Goldman and Lewis, 1978). The results of this study gave some indication that it was the nonspecific stimulation effects of the training which were important in recovery, rather than specific carry-over from common features of the task.

The fact that experience per se may be an important factor in promoting recovery is of considerable importance, since there has been suspicion about the value of training programmes in the recovery from brain damage (Byers and McLean, 1962). Given the intricacy of the neurobiological factors involved in early brain injury it may be too easy to forget about the cognitive dimension. During development a child moves from a relatively primitive analysis and interaction with the external world to a stage where his conceptual structures are complex and enable sophisticated analysis and prediction of the environment. The cognitive capacities evident at maturity have been built on simpler ones gathered progressively during childhood. Thus, impairment of systems which gather and utilize information as a consequence of cerebral damage could result in diminished intellectual achievement, not through injury to critical higher-level systems, but because essential ‘feeder’ mechanisms have malfunctioned. An adult with a subcortical lesion which destroys the left auditory radiation and the callosal input from the contralateral hemisphere may show a normal audiogram


but be incapable of interpreting spoken language (Gloning et al ., 1963). Nevertheless, speech, reading and writing may be normal since their associated cortical circuitry is undamaged and language has been previously established. In a child, bilateral disruption of the auditory radiations by a lesion damaging the lips of the sylvian fissures and the insulae can preclude the normal development of language, despite evidence of hearing in the normal range (Landau et al , 1960). Even children who become peripherally deaf after acquiring speech may not only show arrested language development but lose previously acquired linguistic skills (Bay, 1975). This ‘cognitive starvation’ effect, which is easy to comprehend in the instances cited, may also operate within association areas of the cortex and may underlie the tendency of early brain pathology to produce more global intellectual depression effects than are seen with lesions in adolescence and adulthood.

The data briefly reviewed in this section emphasize that the effects of brain injury in the developing nervous system are particularly complex. When injury occurs, it is within a system whose functional capacities are still unfolding, and the effects of injury are sometimes not immediately apparent. The plastic capacity of the system may also mask the extent of any physical injury and lead to a false assumption that effects have been transitory. There is also the added complication that, whilst injury may not have damaged structures critical for the attainment of certain cognitive skills, because critical ‘feeder’ systems have been impaired, these cognitive mechanisms may not have the experiental basis upon which to build.

The consequences of early brain damage

Global and specific processing

In 1942 Hebb drew attention to the different patterns seen following brain damage in both mature and immature nervous systems. Adults, he found, could be considered to be either aphasic or non-aphasic types (Hebb, 1942). The aphasic type showed obvious evidence of deterioration on verbal tasks and some also showed evidence of impairment on such non-verbal tasks as detecting absurd errors in pictures and block manipulation performance. However, some of the aphasic patients showed evidence of almost normal performance on nonverbal tasks, and Hebb remarked on the wide disparity of abilities seen on particular tests in individual aphasic patients. This conclusion of Hebb’s (that non-verbal skills may be retained to a remarkable extent in some cases of aphasia) is reinforced by more recent research on this topic (Zangwill, 1964; Kertesz and McCabe, 1975). In the non-aphasic type of brain injury on the other hand many verbal tasks could be adequately completed, but there was usually severe impairment on maze learning, block manipulation and picture absurdity tasks, as well as impairment on some verbal tasks, e.g. the defining of abstract



words, or naming opposites. The two types were, however, sufficiently different for Hebb to state that following adult brain injury a reasonably specific pattern of deficit often emerges—some skills showing deterioration and others being relatively intact. In contrast, a group of children with what he termed ‘exogenous’ brain injury showed no evidence of a dual pattern, that is the ‘aphasic’ and ‘non-aphasic’ types did not occur as a consequence of early brain injury. The group as a whole showed depression of verbal IQ, but since he thought it unlikely that every case of brain injury in infancy involved damage to the language areas it must be that ‘low verbal test scores are produced by early lesions outside the speech areas’ (Hebb, 1942, p. 286). He went on to argue that the more global pattern of intellectual depression seen after early brain injury occurs as a result of the differing demands being made on the adult and the child after cerebral damage. The adult has merely to make use of skills which have already been acquired, whereas the child has still to assimilate a range of skills. Since a greater cognitive demand is made during the acquisition of a skill than by the performance of one already acquired, the growing child is at a greater loss than the adult when an equivalent amount of brain tissue has been lost in both. Hebb went further and argued—following Lashley (1929)—that some degree of equipotentiality must exist in the cortex and that areas outside the classical language areas must be involved in the development, but not the maintenance, of linguistic skills once they have been mastered.

Two major hypotheses then emerge from Hebb’s work (1942). The first is that early, rather than late, brain damage has a more global depressive effect upon intellectual development. The second hypothesis is that the developing nervous system is characterized by a greater degree of equipotentiality than that of the adult, since the attainment of normal adult performance on a range of specific skills seems to depend on the integrity of whole cerebrum. The first hypothesis has, in general, received support from subsequent research. Bryan and Brown (1957) found that there is a strong relation between the age of injury and mean IQ, so that those with an injury present at birth averaged a score of 62, those injured in infancy averaged 66 while those with injuries occurring between 3 and 10 years and between 10 and 20 years averaged 71 and 85 respectively. Thompson (1978) reported that in 282 subjects who sustained localized cerebral injury in childhood, there was a linear relationship between age of injury and full-scale IQ with those injured before 5 scoring 97 and those injured above 15 years scoring 106.5. It should be noted, however, that whilst McFie (1961a) found a rise in mean IQ between those injured in the age bands 1-4 and 5-9 years from 88.8 to 106.0, he found a fall in IQ with those sustaining injury between 10 and 15 years (82.7). However, on balance the findings would seem to support Hebb’s initial contention.

The second hypothesis emerging from Hebb’s (1942) study, that the developing nervous system is characterized by a greater degree of equi-


potentiality than that of the adult, is rather more contentious since it is more difficult to test than it might appear at first sight. It has already been pointed out that general depression of IQ cannot be used as evidence for a type of mass action operating during development, since it may also argue for an interdependence of separate capabilities being required for the construction of more complex schemata. It is also apparent that brain-damaged children show widely differing patterns of impairment, which would be difficult to comprehend if there were a tendency for the brain to act uniformly in the acquisition of cognitive skills (Strauss and Lehtinen, 1968). There is also the added difficulty that IQ tests may be rather insensitive to specific patterns of disability produced by brain injury, both in children (Boll and Reitan, 1972) and adults (Walsh, 1978), a factor which has resulted in the construction of specialized test batteries.

However, instead of asking whether the general depressive effect of early brain damage on IQ is due to a greater degree of global processing in the immature CNS, it might be more fruitful to consider whether a similar pattern of impairment emerges on specific skills after similar damage in the child and the adult. McFie (1961a), in an investigation of the effects of localized post-infantile cerebral lesions in children, found that there was a tendency for Wechsler verbal scores to be lower following left hemisphere injury and performance scores to be lower following damage to the right hemisphere. He also noted a similarity in the pattern of impairment shown on the Memory for Designs component of the Terman-Merrill scale (1937) between children and adults when comparing the effects of frontal, temporal and parietal injury. He reported that the greatest deficit is to be found in both groups following right parietal damage. Fedio and Mirsky (1969) examined the pattern of impairment exhibited by children with either unilateral temporal lobe or with centrencephalic epilepsy on a test battery designed to measure performance on both verbal and non-verbal tasks, and a task of sustained attention. The children, who had a history of illness dating from early school years, showed similar impairment profiles to those of adults with similar pathology. Those with left temporal epileptiform foci required a greater number of trials to learn lists of ten words and showed greater loss after a 5- minute interval than those with right temporal or centrencephalic pathology. Those with right temporal pathology showed greatest impairment on the recall of the order of random shapes and on production of the Rey-Osterrieth figure. The centrencephalic group showed the greatest deficit on a task requiring sustained attention. Annet et al (1961) also found a similar pattern of verbal and spatial difficulties in children classified on the basis of lateralized EEG abnormalities. These results would suggest that children show impairments of the same type as those found in adults with similar pathology. It may be objected however, that in these cases the damage is characteristic of juvenile rather than infant brain damage, and, if adult cortical specialization appears gradually, then patterns of specific loss will also begin to appear, producing the observed



similarity to the adult type by middle childhood. However, evidence does exist which would suggest that there is hemisphere specialization, even when damage occurs perinatally or in infancy.

Damage to the left, but not the right, hemisphere before the end of the first year of life results in impairment in the rate at which combinations of words (elementary syntax), but not single words, are learned (Bishop, 1967). Furthermore, children who have had either a left or right hemisphere removed as a consequence of damage sustained during the first year of life show differential effects depending on which hemisphere is involved. Those with left hemisphere removal show greater difficulty in the comprehension of syntax than right hemi- decorticates (Dennis and Kohn, 1975). In particular, difficulties as shown by a greater number of errors are apparent in comprehension of the passive negative, e.g. ‘the girl is not pushed by the boy’, as opposed to the active affirmative, ‘the boy pushes the girl’. In these experiments comprehension was assessed by having the child choose a picture which depicted the sentence. The greater difficulty in the comprehension of passive sentences was also apparent in longer response latencies. Children with early right hemisphere damage followed by hemi- spherectomy show difficulties on spatial tasks (Kohn and Dennis, 1974). The types of spatial task on which they show relative deficit are those which continue to show improvements in normal subjects through the teens, e.g. the WISC and Porteus mazes, and map reading tasks, which require a subject either to state the direction to be taken or to follow a route through markers placed on the floor. Early maturing skills, such as tactile form matching and visual closure, were unaffected in contrast to right hemisphere injury at maturity which severely depresses these abilities. Further analyses of three cases where hemispherectomy antedated the beginnings of speech were presented by Dennis and Whitaker (1976) and two cases where hemispherectomy for unilateral pathology was carried out at ages 3 and 4 years respectively were reported by Day and Ulatowska (1979). In these cases the pattern of deficit is similar to those previously reported and supports the view that the hemispheres are differentially involved in different aspects of cognition. Specialization would appear to be an early-established characteristic of the child’s brain although its potential for reorganization complicates the issue (see p. 225).

Given that differences in the patterning of intelligence subtest scores only really become obvious during adolescence and beyond (Thompson, 1978), whereas specific deficits are detectable by specially devised tests in younger children, IQ tests appear to be insensitive instruments on which to base theories concerning brain development. One obvious factor which will lessen the sensitivity of IQ tests to brain damage early in development is the capacity to transfer the development of linguistic and spatial skills to the contralateral hemisphere should damage occur. This plastic capacity should not however be confused with ideas about mass action or equipotentiality. These latter ideas


imply the existence of a type of diffuse processing network to which the association cortex within each hemisphere contributes uniformly (Lashley, 1929). The transfer of functional capability from one hemisphere to the other, however, is to be understood in a rather different context. Each cerebral hemisphere contains anatomical structures which are essentially duplicates of those found in the other, with the difference that certain cytoarchitectonic areas may be relatively larger or smaller. In some instances, particularly in the language area, these cytoarchitectonic differences in size may be as great as 700% in favour of the left hemisphere (Galaburda et al ., 1978). However, each specialized cortical area has a pattern of connections to the rest of the brain which, in essence, is a lateral reversal of those found in the other hemisphere. The ability of these ‘duplicates’ to assume some of the functional capacities previously assumed by their cytoarchitectonic counterparts is perhaps not too surprising. Indeed, the puzzling feature is that in a substantial majority of the adult population this ability is lost. The assumption that somehow the areas within each hemisphere act as a kind of equipotential unit, is not supported by the available evidence. If such a diffusely organized system were operative in the left hemisphere during development, then it would be reasonable to assume that unilateral damage, resulting in suboptimal processing capacity, would be sufficient reason to transfer linguistic processing to the remaining intact hemisphere. Milner (1973) has provided evidence, based on language lateralization, tested by the Wada (intracarotid injection of sodium amytal) technique, that only when early injury invades those areas shown to be important for linguistic processing in the adult, will language transfer to the right hemisphere. These results suggest that certain key regions, not the global processing capacity or total operational mode of one hemisphere, promote the establishment of language within that hemisphere.

In considering the question of global or specific processing in development, the evidence on balance suggests a specific processing configuration not dissimilar to that found in the adult brain. It has also been argued that one of the reasons why early brain damage often has a markedly depressive effect on intellectual growth may be that the substrate of complex cognitive processes may require the integrity of fundamental systems in order to attain their full potential. One line of evidence which supports this is the finding that right hemisphere damage, which occurs before one year of age, may have more depressive effects on cognitive growth than damage to the left hemisphere at an equivalent age, or damage to either hemisphere after one year of age (McFie, 1961 b; Woods, 1980). This greater deleterious effect of early right hemisphere pathology on both verbal and performance IQ scores argues for the participation of the right hemisphere in certain fundamental processes that may underlie both linguistic and spatial competence. Given the specialization of the right hemisphere for the acquisition of spatial skills it appears possible that a


certain elementary sensori-motor coordinate system may normally be established by the right hemisphere, and in some way aid the differentiation of more complex skills which do not, at first sight, appear directly connected to it. The importance of elementary sensori-motor expedience in evolving more complex cognitive operations has been stressed by many theorists, and by Piaget (1979) in particular. However, regardless of whether or not the presence of this correspondence between theories of cognitive development and neuropsychological research finding is accepted, the results obtained by McFie (1971a, b) and Woods (1980) emphasize the importance of the presence of a particular structure at a particular stage during development rather than supporting the view that different structures are functionally equivalent.

Aphasia in children

The study of language disorders in children is important to a number of problems in developmental neuropsychology. It has provided evidence concerning the extent of functional specialization in the cerebral hemispheres during maturation; data concerning the reorganizational or plastic capacity of the developing brain; and a third, equally important, question, evidence as to the manner in which language becomes established in cortical structures. This question is, in fact, separate from those of hemispheric specialization and plasticity. Here we are concerned with the similarity between the aphasic symptoms of the child and the adult. The greater the similarity of the syndromes the more likely it is that the adult structural pattern has become established, even if subsequently the recovery of the child is more complete because duplication of function in the contralateral hemisphere is still possible. In the present section each of these three topics will be examined, but the characteristics of childhood aphasia will be examined first.

The clinical picture found in childhood aphasia was described by Guttman (1942), who noted that despite claims that the syndrome was rare, he had found it not an unusual accompaniment of head injury or intracranial pathology. In contrast to the adult, where complaints with speech difficulties, failures to name objects and paraphrasia are common, the aphasic child is usually apathetic and morose with such extreme poverty of speech that it approaches mutism. Absence of spontaneous speech, lack of willingness to speak, and a hesitant dysarthric telegrammatic-style speech are frequently noted, these symptoms being more common in the younger child. In contrast to the extreme poverty of speech production, comprehension of simple instructions is evident so that parts of the body or objects can be appropriately indicated when a request is made. When prompted to speak difficulties in the manipulation of lips and tongue may be apparent together with failure to produce sound. As recovery progresses, initially the child will speak single words when prompted. It will then move to


sparse spontaneous utterances before speech moves into a stage when impoverished but spontaneous conversation occurs with persistent dysarthria. This pattern, which bears the imprint of an almost exclusively motor disorder, occurred in all cases below 10 years of age and occurred regardless of the location of the lesion within a hemisphere. Where damage occurred after the tenth year in some cases speech showed a lack of spontaneity but symptoms more characteristic of the adult pattern with impaired auditory comprehension, syntactical and paraphrasic errors, together with difficulties of naming were found. Dysarthria may or may not accompany these symptoms.

This picture of aphasia in childhood has been supported by subsequent research in which the onset of the language loss is sudden following external injury or internal pathology. In Guttman’s cases the five instances of injury which produced aphasia in which the symptom was exclusively one of speech production difficulty were aged 8 years or below, and the two instances in which speech output was not affected but other aphasic symptoms were present were over 10 years of age. The series of cases reported by Hecaen (1976) show that in two instances where the disorder was exclusively one of speech production the children were aged 6, and in the remaining case 3j years. In children aged 7 years and more, comprehension, naming and paraphrasic disorders were more likely to occur. Both Guttman (1942) and Hecaen (1976) stress the fact that recovery may be extremely rapid, marked improvement sometimes being noted in as little as 6 weeks. In children aged over 10 years however, the time course of the disability may sometimes be prolonged. In middle to late childhood the aphasic symptoms begin to mimic the adult pattern, whereas in early childhood the disorder appears to be purely expressive. In adults expressive aphasia is often produced by lesions located in the anterior region of the hemisphere and difficulties in comprehension occur more frequently with temporal and inferior parietal damage (Geschwind, 1970). However, in children below 10 years, disorders of expression appear to occur regardless of the location of the lesion (Guttman, 1942). A second distinctive feature of the aphasic syndrome in young children is the absence of cases of jargon aphasia, where the speech output is rapid with relatively normal articulation but contains many circumlocutions and paraphrasias (Geschwind, 1972). Woods and Teuber (1978) claim to have one documented case of jargon aphasia in a 5-year-old child, but the clinical description is unlike the adult form. The child produced a stream of meaningless sounds, but when recognizable words were uttered they were most often the names of objects. The child also showed evidence of a purely apraxic disturbance, e.g. sticking out his tongue when asked to blow out a light. To the present author, the picture is too dissimilar to the adult form to be classed as an instance of jargon aphasia, the only feature in common being the high rate of vocal, as opposed to verbal, output.

The cases of language loss in childhood so far described were instances where



loss was abrupt, following external or internal injury, and where injury was usually confined to one cerebral hemisphere. A rather different type of childhood aphasia occurs when the language loss is associated with either the onset or development of bilateral epileptiform abnormalities (Landau and Kleffner, 1957; Worster-Drought, 1971; Gascon et al , 1973). In these cases, the loss of language is associated with difficulties in understanding speech, which, in some cases, may evolve over a matter of days or weeks. The child shows lack of response to speech, which may be mistaken for peripheral deafness. Audiometric testing reveals either mild or moderate hearing loss, but this loss is insufficient to account for the comprehension disorder and, in any case, hearing usually shows progressive improvement after an initial depression. In some cases auditory evoked potentials to pure tones may be normal, but evoked potentials to speech show abnormalities (Gascon et al ., 1973). Loss of language is gradual and persistent, and while in some cases recovery may occur over a period of years (Landau and Kleffner, 1957), in other cases it appears to be permanent (Worster- Drought, 1971). In some cases loss of speech may be almost total and auditory comprehension limited to less than a dozen words. Despite gross impairment in the development of language, frequently these children do not show impairment on non-verbal tasks in intelligence tests. Of the 14 cases described by Worster- Drought (1971), performance IQ ranged from 96 to 140, with only one case falling below 100. This remains true despite the fact that, in many cases, the onset of pathology is at less than 5 years of age. These cases of bilateral abnormality are in contrast to cases where a unilateral lesion produces aphasia, from which the child subsequently recovers yet shows a low overall IQ (Hecaen, 1976).

When damage to a single hemisphere produces aphasia the child usually recovers language, and in the young child this recovery is usually better than when damage occurs above 10 years (Lenneberg, 1967). This has often been seen as evidence that the two cerebral hemispheres are initially equipotential as far as the development of language is concerned. Further, it has sometimes been claimed that both cerebral hemispheres are involved initially in language development with lateralization increasing with age (see Dennis and Whitaker, 1977 for a review). It has already been noted that as far as attainment on certain language tests is concerned the two hemispheres are not equivalent. The view that the right hemisphere is involved in language acquisition in the infant and young child comes from reports of the high incidence of speech disturbances following right hemisphere damage. The incidence of language disorders with lesions of the left and right hemispheres described by different researchers varies widely. In the case of the left hemisphere, damage has been estimated to produce language disorder with an incidence varying from 25% (Ingram, 1964) to over 90% (Dunsdon, 1952). In the case of the right hemisphere the estimated incidence has varied from less than 1 % (Ingram, 1964) to nearly 38 % (Dunsdon, 1952). Only one investigator has claimed an equal frequency of language dis-


order following either left or right cerebral damage (Basser, 1962). The discrepancies seem too large to attribute to statistical sampling fluctuations. One of the problems encountered in this area is the definition of what constitutes an aphasic language disturbance. Language difficulties are associated with depressed general intelligence (Mein, 1960) so that severe brain damage which produces severe retardation may produce language disturbance indirectly. There is also the problem of whether speech disturbance should be considered an aphasic disturbance (Ingram, 1965). It is already apparent that the syndrome of aphasia • in children may vary from almost total mutism to a clinical picture similar to that of the adult with comprehension disturbance and naming impairment. That the type of impairment can vary not just with age of the child but also be related to the damaged hemisphere can be seen by examining the series of Hecaen (1976). Of 6 cases of right hemisphere damage, only the two youngest (6 and 3^ years) showed any disturbance and this was articulatory. Bishop (1967) has reported that in cases of infantile hemiplegia, articulatory disturbances are equally likely following damage to either hemisphere, but that left hemisphere damage additionally delays the acquisition of word combinations rather than single words.

The possibility of a different pattern of impairment following left and right hemisphere injury is not the only factor which complicates the issue. Woods and Teuber (1978) have pointed out that there is a tendency for investigators since 1940 to report a lower incidence of aphasia following right hemisphere injury than earlier workers. They attribute this to the fact that in older investigations aphasias and hemiplegias were frequently complications of systemic infectious illnesses such as scarlet fever, bacterial pneumonia and diphtheria, which can produce not only focal lesions but also diffuse bilateral encephalopathy. Undoubtedly the frequent reliance on hemiplegia alone as the sign indicating exclusive damage to one hemisphere is likely to result in the inclusion of cases where a less extensive pathology is also present in the hemisphere that is assumed to be intact. Bearing these facts in mind, it would obviously be hazardous to speculate concerning the true incidence of language disturbance following right hemisphere pathology. For the moment it is sufficient to say that the incidence of aphasia following right hemisphere damage may be considerably less than previously thought, perhaps as little as 5 % in those who were previously right-handed (Woods and Teuber, 1978).

A finding that has already been mentioned several times is that concerning the capacity of the right hemisphere to acquire language following early left hemisphere injury. There is little doubt that the capacity to transfer language to the right hemisphere is a real factor in the recovery from aphasia in children. However, it cannot be assumed that in every case of childhood aphasia recovery of language is due to transfer to the contralateral hemisphere. Milner (1974) noted, on the basis of the Wada test, that in adults who were left-handed but had



sustained early left hemisphere damage, language was present in the left hemisphere in 30 % and bilaterally present in 16 % of cases. Thus, in 46 % of cases who had left hemisphere injury, the left hemisphere was still involved in language to some degree. Whether or not language transferred depended on whether certain critical areas were damaged. In cases where left hemispherectomy is performed, following widespread unilateral damage, it is clear that the presence of linguistic competence is dependent on the remaining hemisphere (Dennis and Whitaker, 1977). When such language transfer does occur, while verbal IQ may not be significantly depressed relative to performance IQ, it should be remembered that such tests do not directly sample knowledge of language structure. Where tests are designed to evaluate grammatical comprehension then deficits appear (Dennis and Kohn, 1975; Teuber, 1975; Dennis and Whitaker, 1976; Day and Ulatowska, 1979). However, with these reservations in mind, children exposed to left hemispherectomy do show an adequate degree of language competence in relation to their overall IQ and it has frequently been remarked that it would be an incredible improvement if each adult aphasic could recover the same level of language competence (Geschwind, 1972).

The duration of such plasticity in the developing brain has been the subject of disagreement. Lenneberg (1967) believed that the period of plasticity in regard to language mechanisms lasted until puberty. Krashen (1973) has challenged this view mainly on the basis that right hemisphere damage above the age of 5 does not often produce aphasia whereas below this age it frequently does. However it should be understood that the issue of the degree to which both hemispheres are involved in language acquisition early in life (and evidence has already been cited that right hemisphere aphasia may be quite different in form from left hemisphere aphasia in young children) is quite a different one from the question of whether interhemispheric transfer is possible. Children between 5 and 10 years do show good recovery from aphasia and it would be surprising indeed if language could have survived in the left hemisphere given the extent and severity of the damage in some instances, e.g. right hemiplegia and hemianopsia (Hecaen, 1976). On the balance the evidence would appear to favour a period of plasticity extending to at least 10 years of age. There is even some indication that a period of reduced plasticity may extend far beyond this age although whether it involves inter-hemispheric transfer or improved within-hemisphere recovery is another question. Teuber (1975) noted that an analysis of 167 cases of brain injury sustained during the Korean campaign showed that the population who were under 22 at the time of injury showed better recovery of language than those who were 23 years and over. It may be premature then to try to set rigid cut-off points for recovery.

The evidence presented here suggests that aphasia in children is not one syndrome but several. In children of 6 years and below mutism and dysarthria appear as the main symptoms with comprehension being relatively well


preserved. Furthermore this pattern appears to occur regardless of whether the lesion is in the left or right hemisphere and also appears to be insensitive to the precise location of the lesion within a hemisphere (whether it is frontal, temporal or parietal). Above 6 years, symptoms which are regarded by many as truly aphasic (comprehension and naming disorders) appear. The symptoms appear to occur largely following left rather than right hemisphere lesions. Between the ages of 6 and 14, jargon aphasia in its adult form is infrequent although the extended circumlocutions that are one of the characteristics of aphasia do occur (Guttman, 1942). The rapidity of the recovery process in some cases and, in very young children, the preservation of comprehension, makes it extremely unlikely that language has been totally relearned by the right hemisphere (Geschwind, 1972). This factor has suggested to some investigators that the right hemisphere must be involved in linguistic processing at an early developmental stage and in fact retains some capacity for comprehension even in the adult after cerebral differentiation (Kinsbourne, 1975).

It is possible then that during the early stages of language learning both hemispheres acquire comprehension and share control of the speech mechanism. This may be necessary in the initial stages, because fine bilateral control of the speech mechanism is required since suitable motor synergisms for a culture- specific phoneme system are not yet well established in subcortical structures. The consequence of this arrangement is that a lesion to either hemisphere can disrupt speech production but comprehension is relatively unaffected because the structural basis of language as opposed to speech does not require a bilateral component. However, the establishment of subcortical synergisms for the execution of the basic components of speech production together with the presence of structurally more specialized language mechanisms in the temporoparietal region of the left hemisphere normally leads to left hemisphere capture of the speech output mechanism. This process is probably a gradual one, but as it proceeds there is less functional demand for right hemisphere processing of language and there may even be active inhibition of its linguistic processing by the left hemisphere. Eventually this isolation of the right hemisphere may lead to structural changes at the synaptic level so that the re-establishment of control is no longer possible. To the extent that this isolation process is incomplete transfer of control is still possible. Thus in young children (under 5 years) the loss of the left hemisphere will show itself in only transitory speech output disturbances since this process of capture is just beginning and both hemispheres are still involved. Even at this age however the linguistic superiority of the left hemisphere is already apparent (see Young, Chapter 6, this volume, for a review of the psychophysical literature), and it is in fact this superiority that will allow the eventual suppression of the right hemisphere. In children between the ages of six and ten years the speech mechanism is probably under the control of the left hemisphere but right hemisphere control mechanisms have not yet functionally



atrophied. Left hemisphere damage at this age can produce speech disturbances solely and/or truly aphasic disturbances which are transitory, while damage to the right hemisphere only rarely affects these mechanisms. With the passage of time however, the capability of the right hemisphere wanes through disuse and in the majority of cases only a token linguistic capacity remains. Even if in later life the possibility of direct inhibition by the left hemisphere is removed, by severing the corpus callosum, the right hemisphere has residual linguistic comprehension but remains mute (Gazzaniga and Sperry, 1967). This is probably because the process of gaining control of the speech mechanisms involves the regulation of neuromuscular synergisms at subcortical levels, and these remain under left hemisphere control. It should be noted that the control of the vocal apparatus by the left hemisphere may be specific to its use in the context of spoken language. Where lesions of the left hemisphere produce expressive (Broca’s type) aphasia the ability to use the voice in the context of singing including the fluent production of words may be well preserved (Yamadori et al ., 1977). Evidence for a motor capture account of left hemisphere language dominance can also be found in studies of adult aphasics (cf. Kinsbourne, 1975).

The plasticity of the developing brain

It is usually accepted that the younger the individual when the brain sustains injury, the greater the resilience and the greater the capacity for functional restitution. Against this one must set the view that the developing brain is particularly vulnerable and long-term effects emerge if normal development is impaired. These two views may be partially reconciled by proposing that following early brain damage, specific skills may be spared but at a cost that will be seen in the overall lowered cognitive capacity of the brain (Teuber, 1975). Thus language or visuo-spatial skills may be spared following left or right hemisphere injury respectively, but intellectual achievement as measured by IQ tests or by school performance will show depression. Evidence already cited concerning the specific effects of early brain damage makes it clear that the consequences are not just seen in a uniformly lowered total processing capacity but depend on the site of injury. Language achievement is specifically lowered following left hemisphere injury and spatial skills depressed specifically following right hemisphere injury.

The evidence in favour of a greater degree of plasticity comes from a number of sources. Some animal species show spared sensory capacity following cortical lesions in infancy (Schneider, 1969) while in other species age at time of injury does not appear to affect the magnitude of the deficit (Doty, 1973). Even where pattern vision is spared following early lesions of the striate cortex, the animals may still take longer to pretrain before formal testing can commence (Schneider,


1970). In man there is evidence of age-related sparing of sensory function. Rudel et al (1974) noted minimal impairment in brain damaged children on somaesthetic thresholds but these children were still impaired on tactile object recognition. Elementary motor function in children may also show greater savings following early massive unilateral injury (Cairns and Davidson, 1951), but such abilities as are preserved, are rudimentary. In man following early unilateral damage that is extensive enough to destroy large areas of the striate cortex, the visual field defects are similar to those produced in adults with similar pathology (Paine, 1960). In this case it might be expected that savings would be possible given the existence of a second, phylogenetically older, visual structure in the midbrain. Where lesions are more restricted, however, savings on visual (in terms of shrinkage of the size of scotoma), somatosensory and motor functions are age-related and show relatively better recovery even when damage occurs early in the third decade of life as compared to later (Teuber, 1975). Without doubt however, the most outstanding examples of functional recovery are those which occur in the areas of language and spatial skills in man following early injury.

The explanation of the functional recovery that does occur following early brain damage is not straightforward. As discussed above, part of the restitu- tional capacity may lie in mechanisms that enable individual neurones to withstand injury so the functional extent of a lesion may be less than in the mature system. It may also lie in neural regeneration per se, although Schneider (1979) has provided evidence that such anomalous regeneration, when it occurs, may actually result in greater behavioural deficit. In the case of the somatosensory and motor systems, while the greater volume of neural circuitry is concerned with analysis and control of the contralateral side of the body, ipsilateral pathways do exist. In the case of the motor system there is even evidence that hypertrophy of ipsilateral pathways occurs after early hemi- spherectomy (Hicks and D’Amato, 1970). These ipsilateral pathways may assume greater functional importance in the case of unilateral brain damage and sustain the limited behavioural savings that occur. The continued development of linguistic and spatial skills after early brain damage are however of a different order. It has been suggested that the survival of these skills in one hemisphere is due to the fact that the necessary processors exist initially in each hemisphere but that during development one hemisphere suppresses the influence of the other. This suppression of the influence of the contralateral hemisphere may be a necessary prerequisite for the development of higher cognitive skills, since processing space may be at a premium. When both language and spatial skills are acquired by only one hemisphere (following early hemispherectomy) neither skill reaches its full potential (Teuber, 1975). Where the corpus callosum is absent during development, and normal interhemispheric communication is consequently impossible, a rather bizarre pattern of cognitive development is



seen. In such cases of callosal agenesis it appears that either performance IQ or verbal IQ becomes pre-eminent despite the existence of functional capacity in two hemispheres (Dennis, 1977). Extreme discrepancies do not in these cases appear to be predictable from age of the subject at the time of testing, sex, type of agenesis, handedness or specific neurological signs. It appears rather that a mechanism which enables a normal balance of cognitive skills to occur is absent. In the normal individual then the existence of two intact hemispheres may not be sufficient for normal cognitive growth. Some additional mechanism which ensures that unnecessary duplication does not occur and enables an efficient use of available processing capacity seems to be necessary. The consequence of the presence of such a mechanism during development is that usually specific skills become established predominantly in one hemisphere or the other and once they are so established there is little opportunity to recapitulate the process. The failure of the adult brain to fully re-establish linguistic or spatial skills following damage is a consequence of the presence of a mechanism (whose effector path is the corpus callosum) that enables a balanced and complete cognitive growth to occur.


Consideration of the evidence concerning the outcome of brain injury in the developing nervous system leads to the inescapable conclusion that age at the time of injury is a critical factor in determining both the initial syndrome and the pattern of adaptation that follows. Certain abilities such as early maturing spatial skills, language and various elementary sensory and motor functions may show relative recovery, the extent being determined by the location and size of the lesion. Other abilities, particularly those involving fine motor coordination, late maturing spatial skills as well as the overall level of intellectual attainment, may show more profound defects. These latter effects may be partly ascribed to a lower overall processing capacity, but in the case of general intellectual attainment, more profound effects are seen following damage to the right hemisphere before (compared with after) one year of age, a factor which argues that volume of tissue damaged is not the sole consideration. Differing patterns of language disturbances are also seen in children depending on the child’s age at the time pathology develops. These considerations, which emphasize that damage is occurring within a system whose state is continually changing, present particular difficulties for those with interests in the outcome of early brain damage. As Teuber and Rudel (1962) point out: ‘Whether we are working with infrahuman forms or with children, we must define (1) those aspects of behaviour in which the effects of early injury appear only with a delay, as development progresses; (2) those other aspects of performance in which there will be impairment at all ages; and finally, (3) those aspects of performance


where there is an immediate effect which, however, disappears as development proceeds’. The interpretation of research which finds age-dependent differences is often far from clear. Reports that the patterning of IQ subtests is insensitive to the location of injury in young children may be a genuine indication of an age- dependent difference, despite the fact that such tests were not specifically designed to assess brain damage. However, since it is found that in adults that IQ tests are only indicative of lesion laterality in the acute but not the chronic phase (Fitzhugh et al ., 1962), the failure to find specific indication of lesion location in children with infantile injury may be attributable to the interval between injury and testing rather than an age-related difference in functional organization.

The emergence of specific functional deficit following early brain injury stresses the importance of critical structures for cognitive achievement. Just how extensively the predesignated structures of the nervous system determine the detailed characteristics of cognitive development remains to be seen. It is only relatively recently that explorations of the specific patterns of loss, as measured by specially developed tests, have begun. It is also apparent that much of the research has been concerned with hemispheric asymmetry where the structural similarity of the hemispheres means that the capacity for interhemispheric reorganization may lead to an over-valuation of the plastic capacities of the developing brain. Cases where bilateral loss of a structure is involved give one much less confidence in the restitutional capacity of the young brain. Bilateral frontal lobe damage in childhood appears to have effects at least as serious as equivalent damage in adults (Russell, 1959; Ackerly, 1964) although the paucity of research in this area and the anecdotal nature of some of the findings make conclusions tentative. There is a scarcity of unequivocal evidence that areas of the brain not normally involved in the development or performance of specific functions may assume those functions when other areas are damaged. Goldman and Lewis’s (1978) demonstration that the dorsolateral prefrontal cortex in the macaque may assume some of the functional capability of the orbitofrontal cortex, provided damage occurs early in development, remains one of the few clear demonstrations of such effects in the CNS of primates following bilateral lesions. However, the fact that the normal limits of plasticity and the extent to which they may be influenced by specific experience remain undetermined, gives developmental neuropsychologists particular problems in understanding the precise nature of brain-behaviour interrelations. Nevertheless, the fact that plasticity is a real phenomenon and may be influenced by specific experience (Goldman and Lewis, 1978) gives some hope that understanding of its nature may lead to more effective remediation regimes designed to capitalize on its characteristics.


I am grateful to H. D. Ellis and E. A. Salzen for advice and discussion during preparation of this manuscript.




Ackerly, S. S. (1964) ‘A case of paranatal bilateral frontal lobe defect observed for thirty years’, in The Frontal Granular Cortex and Behavior (eds. J. M. Warren and K. Akert), McGraw-Hill, London, 192-218.

Annet, M., Lee, D. and Oimsted, C. (1961) Intellectual disabilities in relation to lateralised features of the EEG. Little Club Clin. Devel. Med., 4, 86-112.

Basser, L. S. (1962) Hemiplegia of early onset and the faculty of speech with special references to the effects of hemispherectomy. Brain, 85, 427-459.

Bay, E. (1975) ‘Ontogeny of stable speech areas in the human brain’, in Foundations of Language Development, Yol. 2 (eds. E. H. Lenneberg and E. Lenneberg), Academic Press, London, 21-29.

Bishop, N. (1967) ‘Speech in the hemiplegic child’, in Proc. 8th Medical and Educational Conference of the Australian Cerebral Palsy Association, Tooronga Press, Melbourne, 141-153.

Boll, T. J. and Reitan, R. M. (1972) Comparative ability interrelationships in normal and brain injured children. J. Clin. Psychol, 28, 152-156.

Broca, P. (1861) Remarques sur le siege de la faculte du langage articule suivies d’une observation d’aphemie (pert de la parole). Bull. Soc. Anat. Paris, 3, 330-357.

Bronson, G. (1974) The postnatal growth of visual capacity. Child Devel, 45, 873-890.

Bryan, G. E. and Brown, M. H. (1957) A method for differential diagnosis of brain damage in adolescents. J. Nerv. Ment. Dis., 125, 69-72.

Byers, R. K. and McLean, W. T. (1962) Etiology and course of certain hemiplegias with aphasia in childhood. Pediatrics, 29, 376-383.

Cairns, H. and Davidson, M. A. (1951) Hemispherectomy in the treatment of infantile hemiplegia. Lancet, ii, 411-415.

Chase, H. P. (1973) The effects of intrauterine and postnatal undernutrition on normal brain development. Ann. New York Acad. Sci., 205, 231-244.

Chow, K. L. and Hutt, P. (1953) The association cortex of Macaca mulatta: a review of recent contributions to its anatomy and functions. Brain, 76, 625-677.

Cobrinik, L. (1959) The performance of brain-injured children on hidden figure tasks. Amer. J. Psychol, 72, 566-571.

Day, P. S. and Ulatowska, H. (1979) Perceptual, cognitive and linguistic development after early hemispherectomy: two case studies. Brain and Lang., 1, 17-33.

Dejerine, J. (1892) Contribution a l’etude anatomopathologique et clinique des differentes varieties de cecite verbale. C. R. Soc. Biol, 4, 61-90.

Dejerine, J. and Andre-Thomas, J. (1912) Contribution a l’etude de l’aphasie chez les gauchers. Rev. Neurol, 24, 213-226.

Dennis, M. (1977) ‘Cerebral dominance in three forms of early brain disorder’, in Topics in Child Neurology (eds. M. E. Blaw, I. Rapin and M. Kinsbourne), Spectrum Publications, London, 189-212.

Dennis, M. and Kohn, B. (1975) Comprehension of syntax in infantile hemiplegics after cerebral hemidecortication: left hemisphere superiority. Brain and Lang., 2 , 472-482.

Dennis, M. and Whitaker, H. A. (1976) Language acquisition following hemidecortication: linguistic superiority of the left over the right hemisphere. Brain and Lang., 3, 404-433.

Dennis, M. and Whitaker, H. A. (1977) ‘Hemispheric equipotentiality and language acquisition’, in Language Development and Neurological Theory (eds. S. J. Segalowitz and F. A. Gruber), Academic Press, London, 93-106.

Dobbing, J. (1968) ‘Vulnerable periods in developing brain’, in Applied Neuro chemistry (eds. A. N. Davidson and J. Dobbing), Blackwell, Oxford, 287-316.

Doty, R. W. (1973) ‘Ablation of visual areas in the central nervous system’, in Handbook of Sensory Physiology, Vol. Ill, 3B, Central Processing of Visual Information (ed. R. Jung), Springer Verlag, Berlin, 438-541.

Dunsdon, M. I. (1952) The Educability of Cerebral Palsied Children. Newnes, London.

Fedio, F. R. and Mirsky, A. F. (1969) Selective intellectual deficits in children with temporal lobe or centrencephalic epilepsy. Neuropsychologia, 1 , 287-300.

Fitzhugh, K. B., Fitzhugh, L. C. and Reitan, R. M. (1962) Wechsler-Bellevue comparisons in groups of‘chronic’ and ‘current’ lateralized and diffuse brain lesions. J. Consult. Psychol, 26, 306-310.


Flourens, P. (1824) Recherches Experiment ales sur les Proprietes et les Fonctions du Systeme Nerveux dans les Animaux Vertebres. Crevot, Paris.

Galaburda, A. M., Le May, M., Kemper, T. L. and Geschwind, N. (1978) Right-left asymmetries in the brain. Science, 199, 852-856.

Gascon, G., Victor, D., Lombrosso, C. T. and Goodglass, H. (1973) Language disorder, convulsive disorder and electroencephalographic abnormalities. Arch. Neurol, 28, 156-162.

Gazzaniga, M. S. and Sperry, R. W. (1967) Language after section of the cerebral commissures. Brain, 90, 131-148.

Geschwind, N. (1965) Disconnexion syndromes in animals and man. Brain , 88, 237-294 and 585 644.

Geschwind, N. (1970) The organisation of language and the brain. Science, 170, 940-944.

Geschwind, N. (1972) Disorders of higher cortical functions in children. Clin. Proc. Child. Hosp., 28, 261-272.

Gloning, I., Gloning, K. and Hoff, H. (1963) ‘Aphasia—a clinical syndrome’, in Problems of Dynamic Neurology (ed. L. Halpern), Jerusalem Post Press, Jerusalem, 63-70.

Goldman, P. S. (1974) ‘An alternative to developmental plasticity: heterology of CNS structures in infants and adults’, in Plasticity and Recovery of Function in the Central Nervous System (eds. D. G. Stein, J. J. Rosen and N. Butters), Academic Press, New York, 149-174.

Goldman, P. S. (1976) ‘Maturation of the mammalian nervous system and the ontogeny of behavior’, in Advances in the Study of Behavior, Vol. 7 (eds. J. S. Rosenblatt, R. A. Hinde, E. Shaw and C. Beer), Academic Press, London, 1-90.

Goldman, P. S. and Lewis, M. E. (1978) ‘Developmental biology of brain damage and experience’, in Neuronal Plasticity (ed. C. W. Cotman), Raven Press, New York, 291-310.

Goldstein, K. (1939) The Organism. American Book Publishers, New York.

Goltz, F. (1892) Uber die Verrichtungen des Grosshirns. Pflugers Arch. Ges. Physiol., 51, 570- 614.

Guttman, E. (1942) Aphasia in children. Brain, 65, 205-219.

Harlow, H. F., Thompson, C. I., Blomquist, A. J. and Schiltz, K. A. (1970) Learning in rhesus monkeys after varying amounts of prefrontal lobe destruction during infancy and adolescence. Brain Res., 18, 343-353.

Hebb, D. O. (1942) The effects of early and late brain injury upon test scores and the nature of normal adult intelligence. Proc. Amer. Phil. Soc., 85, 275-292.

Hecaen, H. (1976) Acquired aphasia in children and the ontogenesis of hemispheric functional specialisation. Brain and Lang., 3, 114-134.

Hecaen, H. and Albert, M. L. (1978) Human Neuropsychology. John Wiley and Son, New York.

Henry, C. E. (1944) Electroencephalograms of normal children. Mon. Soc. Res. Child. Devel, 9, Serial No. 39.

Henschen, S. E. (1926) On the function of the right hemisphere of the brain in relation to the left hemisphere in speech, music and calculation. Brain, 49, 110-123.

Hicks, S. F. and D’Amato, C. S. (1970) Motor-sensory and visual behavior after hemispherectomy in newborn and mature rats. Exp. Neurol, 29, 416-438.

Hubei, D. H. and Wiesel, T. N. (1963) Receptive fields of cells in the striate cortex of very young, visually inexperienced kittens. J. Neurophysiol., 26, 994-1002.

Ingram, T. T. S. (1964) Pediatric Aspects of Cerebral Palsy. E. and S. Livingstone, Edinburgh.

Ingram, T. T. S. (1965) Specific retardation of speech development. Speech Pathol Ther., 8, 3-11.

Kennard, M. A. (1940) Relation of age to motor impairment in man and in subhuman primates. Arch. Neurol. Psychiat., 44 , 377-397.

Kertesz, A. and McCabe, P. (1975) Intelligence and aphasia: performance of aphasics on Raven’s coloured progressive matrices (RCPM). Brain and Lang., 2, 387-395.

Kinsbourne, M. (1975) ‘Minor hemisphere language and cerebral maturation’, in Foundations of Language Development, Vol. 2 (eds. E. H. Lenneberg and E. Lenneberg), Academic Press, London, 107-116.

Kohn, B. and Dennis, M. (1974) Selective impairment of visuo-spatial abilities in infantile hemiplegics after right cerebral hemidecortication. Neuropsychologia, 12, 505-512.

Krashen, S. (1973) Lateralisation, language learning and the critical period. Lang. Learning, 23, 63-74.



Landau, W. M. and Kleffner, F. R. (1957) Syndrome of acquired aphasia with convulsive disorder in children. Neurol., 7, 523-530.

Landau, W. M., Goldstein, R. and Kleffner, F. R. (1960) Congenital aphasia: a clinicopathologic study. Neurol, 10, 905-921.

Langworthy, O. R. (1933) Development of behavior patterns and myelinisation of the nervous system in the human foetus and infant. Carnegie Instit. Pub. No. 139, Washington.

Lashley, K. S. (1929) Brain Mechanisms and Intelligence: A Quantitative Study. Univ. Chicago Press, Chicago.

Lawrence, D. G. and Hopkins, D. A. (1972) Developmental aspects of pyramidal motor control in the rhesus monkey. Brain Res., 40, 117-118.

Lecours, A. R. (1975) ‘Myelogenetic correlates of the development of speech and language’, in Foundations of Language Development, Vol. 1 (eds. E. H. Lenneberg and E. Lenneberg), Academic Press, New York, 121-135.

Lenneberg, E. H. (1967) Biological Foundations of Language. John Wiley and Sons, New York.

Lenneberg, E. H. (1968) The effect of age on the outcome of central nervous system disease in children’, in The Neuropsychology of Development (ed. R. L. Isaacson), John Wiley and Sons London, 147- 170.

Liepmann, H. (1908) Drei Aufsatze aus dem Apraxiegebeit. Karger, Berlin.

Lissauer, H. (1890) Ein Fall von Seelenblindheit nebst einen Beitrag zur Theorie derselben. Arch. F. Psychiat, 21, 222-270.

Lyon, G. (1961) First signs and mode of onset of congenital hemiplegia. Little Club Clin. Dev. Med., 4 , 33-38.

Marshall, W. A. (1968) Development of the Brain. Oliver and Boyd, Edinburgh.

McFie, J. (1961a) Intellectual impairment in children with localized post-infantile cerebral lesions. J. Neurol. Neurosurg. Psychiat., 24, 361-365.

McFie, J. (1961b) The effects of hemispherectomy on intellectual functioning in cases of infantile hemiplegia. J. Neurol. Neurosurg. Psychiat., 24, 240-249.

Mein, R. (1960) A study of the oral vocabularies of severely subnormal patients. J. Ment. Def. Res., 4, 130-143.

Milner, B. (1973) ‘Hemispheric specialisation: scope and limits’, in The Neurosciences: Third Study Program (eds. F. O. Schmitt and F. G. Worden), M.I.T. Press, Cambridge, 75-89.

Milner, B. (1974) ‘Sparing of language function after early unilateral brain damage’, in Functional Recovery After Lesions of the Nervous System (eds. E. Eidelberg and D. G. Stein), Neurosci. Res. Prog. Bull., 12, 213-217.

Milner, E. (1976) ‘CNS maturation and language acquisition’, in Studies in Neurolinguistics, Vol. 2 (eds. H. Whitaker and H. A. Whitaker), Academic Press, London, 31-102.

Molfese, D. L. (1977) ‘Infant cerebral asymmetry’, in Language Development and Neurological Theory (eds. S. J. Segalowitz and F. A. Gruber), Academic Press, London, 21—35.

Paine, R. (1960) Disturbances of sensation in cerebral palsy. Little Club Clin. Dev. Med., 2, 105- 109.

Piaget, J. (1979) ‘Correspondences and transformations’, in The Impact of Piagetian Theory (ed. F. B. Murray), University Park Press, Baltimore, 17-27.

Rudel, R. G., Teuber, H.-L. and Twitched, T. E. (1974) Levels of impairment of sensori-motor functions in children with early brain damage. Neuropsychologia, 12, 95-108.

Russell, W. R. (1959) Brain, Memory and Learning: A Neurologist's View. Clarendon Press, Oxford.

Schneider, G. E. (1969) Two visual systems. Science, 163, 895-902.

Schneider, G. E. (1970) Mechanisms of functional recovery following lesions of visual cortex or superior colliculus in neonate and adult hamster. Brain Behav. EvoL, 3, 295-323.

Schneider, G. E. (1979) Is it really necessary to have your brain lesions early? A revision of the ‘Kennard Principle’. Neuropsychologia, 17, 557-583.

Shagass, C. (1972) ‘Electrical activity of the brain’, in Handbook of Psychophysiology (eds. N. S. Greenfield and R. A. Sternback), Holt, Rinehart and Winston Inc., London, 263-328.

Strauss, A. A. and Lehtinen, L. E. (1950) Psychopathology and the Education of the Brain Injured Child. Grune and Stratton, New York.

Terman, L. M. and Merrill, M. A. (1937) Measuring Intelligence. Harrap, London.

Teuber, H.-L. (1975) ‘Recovery of function after brain injury in man’, in Outcome of Severe Damage to


the Central Nervous System (Ciba Foundation Symposium No. 34), Elsevier, Amsterdam, 159-186.

Teuber, H.-L. and Rudel, R. G. (1962) Behavior after cerebral lesions in children and adults. Dev. Med. Child Neurol ., 4, 3-20.

Thompson, J. (1978) ‘Cognitive effects of cortical lesions’, in Psychology Survey No. 1 (ed. B. M. Foss), Allen and Unwin, London, 86-98.

Trevarthen, C. B. (1968) Two mechanisms of vision in primates. Psychol. Forsch., 31, 299-337.

Tsang, Y.-C. (1937) Maze learning in rats hemidecorticated in infancy. J. Comp. Psychol., 24, 221-254.

Turner, O. A. (1948) Growth and development of cerebral cortical pattern in man. Arch. Neurol. Psychiat., 59, 1-12.

Turner, O. A. (1950) Postnatal growth of the cortical surface area. Arch. Neurol. Psychiat., 64, 378-384.

Ulett, G., Dow, R. S. and Landsell, O. (1944) The inception of conductivity in the corpus callosum and the corticoponto-cerebellar pathway of young rabbits with reference to myelinisation. J. Comp. Neurol., 80, 1-10.

Wada, J. A., Clark, R. and Hamm, A. (1975) Cerebral hemispheric asymmetry in humans: cortical speech zones in 100 adult and 100 infant brains. Arch. Neurol, 32, 239-246.

Walsh, K. W. (1978) Neuropsychology: A Clinical Approach. Churchill Livingstone, Edinburgh.

Walter, W. G. (1950) ‘Normal rhythms—their development, distribution and significance’, in Electroencephalography (eds. D. Hill and G. Parr), Macdonald and Co., London, 203-227.

Werner, G. and Whitsell, B. L. (1973) ‘Functional organisation of the somatosensory cortex’, in Somatosensory Systems, Handbook of Sensory Physiology, Vol. 2 (ed. A Iggo), Springer Verlag, New York, 621-700.

Wernicke, C. (1874) Der Aphasische Symptomenkomplex. Cohn and Weigert, Breslau, Poland.

Woods, B. T. (1980) The restricted effects of right-hemisphere lesions after age one; Wechsler test data. Neuropsychologia, 18, 65-70.

Woods, B. T. and Teuber, H.-L. (1978) Changing patterns of childhood aphasia. Ann. Neurol, 3, 273-280.

Worster-Drought, C. (1971) An unusual form of acquired aphasia in children. Dev. Med. Child Neurol, 13, 563-571.

Yakovlev, P. I. and Lecours, A. R. (1967) ‘The myelogenetic cycles of regional maturation of the brain’, in Regional Development of the Brain in Early Life (ed. A. Minkowski), Blackwell, Oxford, 3-70.

Yamadori, A., Osumi, Y., Masuhara, S. and Okubo, M. (1977) Preservation of singing in Broca’s aphasia. J. Neurol. Neurosurg. Psychiat., 40, 221-224.

Zangwill, O. L. (1964) ‘Intelligence in aphasics’, in Disorders of Language (eds. A. V. S. de Reuck and M. O’Connor), Ciba Symposium, Little Brown and Co., Boston, 261-284.

Chapter Eight - Sex Differences in Brain Development: Process and Effects

Miranda Hughes


Identifying the neural mechanisms which underlie particular behavioural and cognitive functions has become a fundamental aspect of psychological research, and in recent years considerable progress has been made in understanding the way in which both pre- and postnatal hormones can affect brain differentiation. The notion that prenatal hormones which are differentially produced by males and females may have irrevocable effects on the brain as well as on physical morphology is politically provocative; nonetheless, improving our knowledge of the neural substrates of behaviour ought also to facilitate our understanding of how postnatal environment exerts its influence. Thus, to find sex differences in brain differentiation, and to link these to sex differences in cognitive ability and behaviour, does not necessarily imply biological determinism; rather, it enhances our understanding of the raw materials which educational and cultural pressures may mould in a variety of ways.

This chapter discusses the way in which pre- and postnatal hormones affect brain differentiation, and it is argued that the long-term effects of the early hormone environment may predispose any individual to certain ‘masculine’ or ‘feminine’ type behaviours. However, different aspects of our behavioural repertoire are certainly under different degrees of hormonal influence and human behaviour is not clearly sexually differentiated. As Money (1977a) put it so cogently ‘... the only irreducible sex differences are that women menstruate, gestate and lactate, and men impregnate ... most sexually dimorphic behaviour



as we know it is the product of cultural history and not of some eternal verity programmed by non-cultural biology. 5 (pp. 32, 33).

Following Pfeiffer’s (1936) innovative and now classic work, the precise role of prenatal hormones in the development of the hypothalamic mechanisms which control hormone release at puberty, and which are responsible for the development of sexually differentiated physical characteristics, is well established (Harris, 1964, 1970). The well documented cases of children exposed to abnormal levels of particular steroids in utero (Money and Ehrhardt, 1972), and the work of Dorner (1979) on human homosexuality, have subsequently raised a number of questions concerning the effect of hormones on a wide range of behaviour. The line of reasoning seems to be that if (a) some neural mechanisms (e.g. for gonadotropin release at puberty) are determined by the role of prenatal hormones, and (b) foetuses exposed to abnormal levels of types of particular hormones behave in specific and atypical ways, then it follows that (c) just as the prenatal hormonal environment has ‘wired-up 5 the brain in such a way as to determine the expression of certain endocrine functions, so too can it predispose an organism to specific behaviour patterns. A closer examination of the three stages of this argument should facilitate the development of a conceptual framework within which to extend our understanding of the variety of expression in human abilities and behaviour.

Prenatal sex differences in development

Distinctively male or female development begins at around the seventh week after conception when the initially bi-potential embryonic gonad differentiates to form either a testis (in the case of a male) or an ovary (in the case of a female). This differentiation of the gonads is determined by the genetic sex of the zygote (46XY in the male; 46XX in the female); where there is no second sex chromosome as in Turner’s syndrome (45X) the gonads are undifferentiated at birth, although germinal follicles may have been present in the early foetal stages (Scott, 1978). Jost (1979) tentatively suggests that there may be a specific membrane protein controlled by a locus on the Y chromosome (the H-Y histocompatibility antigen) which is responsible for the differentiation of testes, and whose individuals who do not produce this antigen will form an ovary. Once a testis has been formed the release of a substance (probably a foetal protein) known as Mullerian inhibiting substance (MIS) induces the regression of the Mullerian ducts, and secretion of androgenic hormones enables the development of the male reproductive tracts and genitalia. In the absence of testicular hormones female development occurs; ‘ mammals and birds body sex shows a basic developmental trend corresponding to that of the homozygous sex. Characteristics of the heterozygous sex have to be actively imposed by the secretions of the corresponding gonads 5 (Jost, 1979, p. 8). Thus, the appearance


of Turner’s syndrome infants is unequivocally female, and that of Klinefelter’s syndrome infants (47XXY and 48XXXY) is unequivocally male.

Sexual differentiation does not, however, always proceed entirely smoothly, and Scott (1978) has provided a useful classification of some ‘intersex’ conditions. He suggests that there are four basic processes which may distort normal sexual differentiation: (i) chromosomal intersex, in which extra or missing sex chromosomes affect development; (ii) gonadal intersex, in which the gonadal tissue is at variance with the chromosomal constitution of the individual; (iii) partial masculinization of chromosomal and gonadal females, due either to a disorder of adrenal functioning or the exogenous administration of steroid hormones to the mother; and (iv) incomplete masculinization of chromosomal and gonadal males which may occur either because an individual is insensitive to the androgen being produced by the testes or because there is some failure in androgen production. These medical conditions have often been described as ‘nature’s experiments’ because they shed light on the various ways in which hormones affect development.

In Turner’s syndrome, the missing chromosome may be either an X or a Y: evidence for this comes from a report by Leujeune (1964) of monozygotic twins, one of whom was a normal male of 46XY karyotype, the other of whom was born with 45X karyotype (and therefore a female phenotype). Turner’s syndrome females are typically short in stature, and require oestrogen therapy at puberty to effect normal breast development. There may be a range of other physical stigmata present (e.g. shield chest, neck webbing, low-set ears), but general intelligence is not significantly affected (Money and Ehrhardt, 1972). The streak gonads are often entirely non-functional, but they may contain some ova in which case pregnancy is possible.

There are approximately two cases per thousand of males with a 47XXY karyotype, and a similar number with 47XYY karyotype. The former often have small testes and prostates and diminished body and facial hair; they may also show some breast development at puberty, and are frequently infertile. Males with 47XYY karyotype have a tendency to be taller than average, but otherwise display no specific physical abnormalities. It has been estimated that both 47XXY and 47XYY males are over-represented in mental or penal institutions during late adolescence or adulthood. Differences between these groups in deviant behaviour are not significant; however, their crimes are more likely to be sex or property offences than those of their delinquent peers (Meyer-Bahlburg, 1974).

Cases of gonadal intersex (hermaphroditism) show widely varying arrangements of gonadal tissue and genitalia. Scott (1978) suggests that there may be an interchange of genetic material between the X and Y chromosome before the first meiotic division in the primary spermatocyte, which could lead to widely varying sexual differentiation according to the cells in which the Y chromosomal


material is active. Such cases clearly do not provide a homogeneous subject sample, but are nonetheless interesting individually.

The partial masculinization of females and the incomplete masculinization of males illustrate clearly the role of steroids in the development of sex-related physical characteristics. The former of these conditions is usually due to congenital adrenal hyperplasia (CAH) which occurs as a result of an enzymatic deficiency in the adrenal steroid metabolic pathways. The most common form of CAH is 21-hydroxylase deficiency which results in a build-up of 17-hydroxypro- gesterone, the metabolic derivatives of which have a virilizing influence on the female foetus. Male infants appear normal at birth (although puberty may be accelerated by as much as ten years if the condition is not diagnosed and treated), but female infants have masculinized external genitals. The female internal organs are normal, and with appropriate medical treatment (including surgery to feminize the genitals) these girls may menstruate at puberty and eventually bear children.

Masculinization of females may also arise from the influence of steroid hormones administered to the mother during pregnancy to prevent miscarriage. Ehrhardt and Money (1967) report ten such cases, and Scott (1978) describes an individual case following the administration of norethisterone to the mother. As in the CAH cases, masculinization is apparently restricted to the external genitalia, and can be corrected surgically.

There are three possible defects of the androgenization process which can give rise to the incomplete masculinization of the male: defective androgen production, defective Mullerian regression, and androgen insensitivity. The last of these is the best documented and is often known as ‘testicular feminization’. In this condition infants with a normal male 46XY chromosome complement are born with a female phenotype. Since Mullerian regression has occurred normally, there is a short blind vagina, but the external genitalia are unequivocally female. At puberty, there is spontaneous breast development, although pubic hair tends to be scant. The condition occurs despite normal steroid output from the testes, when the receptor cells fail to respond to the androgens which are present.

Individuals with complete androgen insensitivity may be quite oblivious to their condition until puberty, when they seek medical advice for amenorrhea. However, there are incomplete forms of androgen insensitivity: for example, where there is a failure to convert testosterone to the more potent 5a- dihydrotestosterone there may be incomplete masculinization of the external genitalia. Imperato-McGinley et al (1974) described some such cases in which the affected infants are given a female assignment at birth, but at puberty have to undergo a gender re-assignment because ‘anabolic events at puberty, in particular the increase in muscle mass, the growth of the phallus and scrotum, and the voice change, appear to be mediated by testosterone’ (p. 1214).



All of these ‘intersex’ conditions have considerable interest for psychology, in that they provide an opportunity to examine the possible behavioural effects of the prenatal hormones. It is certainly true that the prenatal hormonal environment affects cell differentiation in the brain; the speculations which require critical examination are those concerning the behavioural implications of such hormonal effects.

Hormonal action

The role of prenatal hormones in the development of the internal sex organs and genitalia is clearly established. If these hormones exert equally critical influences on brain differentiation, one would expect to find different patterns of neural networks in male and female brains. It is instructive therefore to examine the mechanism whereby hormones exert their influence so that sex differential developmental processes can be appropriately evaluated, and any anomalies of normal development can be interpreted.

The steroid hormones include the male sex hormones (androgens), the female sex hormones (oestrogens and progestins), and the hormones secreted by the adrenal glands (corticosteroids). Structurally, they resemble one another quite closely but differ radically in function. Their common core structure consists of four interconnected carbon rings. The pattern of bonding and the different side groups affect the overall shape of each molecule, and it is these subtle differences in shape which enable the hormones to attach themselves to specific target cells.

Hormones act directly on genetic mechanisms, so that when gene action is blocked (for example, by the action of certain antibiotics) hormones become powerless to exert their characteristic effect. A single hormone can activate an entire set of functionally related but otherwise quite separate genes, and hormonal specificity is dependent on the functional integrity of the target cells as much as on the hormone itself. The cytoplasm of target cells contains specific intracellular receptor proteins which accumulate and retain the hormone (this in contrast to the receptor mechanism of say, amines, for which the receptor site lies in the cell membrane). The steroid hormones then give rise to an increase in RNA synthesis, and can also effect the synthesis of a new variety of messenger RNA; these RNA molecules direct the formation of new protein molecules in the cytoplasm of the cell which enable the target cell to make its functional responses to the hormone.

During development the presence of male hormones will (in general) have a masculinizing effect on a genetic female. However, in experiments on rats it was found that whilst testosterone increased the amount of RNA produced in the liver cells of both males and females, in the female not only was there an increased amount of RNA, but a new type of RNA was being produced; this finding does suggest that even when male and female developing embryos are


exposed to similar hormonal environments, the consequences need not necessarily be identical (Davidson, 1965). There has been some attempt to discover whether sex differences in brain differentiation are mediated by sex differences in cytoplasmic receptors. Data from Maurer (1973) and from Whalen (1974) show that there was selective cytoplasmic binding of oestrogen in the anterior hypothalamic-preoptic area (of rats), in the median eminence, but not in the cortex; however, the sex differences were not striking \ .. it seems unlikely that the small difference in nuclear retention that we found can account for the large differences existing between males and females in their behavioural responses to oestrogen’ (Whalen, 1974, p. 278).

Sex differences in brain differentiation

Pfeiffer (1936) was the first to establish that sex differences in the reproductive endocrinology of rats were determined by the hormone environment at a specific stage of development. He demonstrated that if a male rat is castrated within 3 days after birth and is subsequently (in adulthood) given ovarian grafts, he will respond to endogenous hormones with a surge of luteinizing hormone (LH) which is sufficient to produce corpora lutea in the ovarian graft. When the ovaries of newborn females were replaced with testes, many of these females failed to show any sign of oestrous cycles when they became adult, but entered a state of constant vaginal oestrus. However, female rats, which were ovari- ectomized at birth and subsequently had received ovarian implants, showed normal oestrous cycles and formation of corpora lutea. Male rats in which the testes were transplanted into the neck region at birth, and which received ovarian implants as adults, showed no capacity to form corpora lutea in the ovarian grafts. Pfeiffer concluded (erroneously) that the pituitary gland becomes sexually differentiated; subsequent experiments (see Harris, 1964, 1970) made it clear that in fact permanent control by the hypothalamus over the pituitary was established by the presence or absence of testosterone in a critical neonatal period. In the absence of testosterone a pattern of cyclic release of follicle stimulating hormone (FSH) and LH by the pituitary was established; when testosterone was present, release of hormones was tonic. Reznikov (1978) states that the critical periods for the sexual differentiation of the brain centres which regulate gonadotropin release \ .. occur in rabbits during the period of 19-23 days, and in guinea pigs at 36-38 days of pre-natal life, in rats, mice and hamsters, in the course of the first five days after birth. In the case of humans, the most probable period of sexual differentiation is considered to be the second trimester of pregnancy. It should be emphasised that experimental influences exerted outside the “critical” period are incapable of moderating the sex- specifying parameters of differentiation of the brain’ (p. 127).

Barraclough and Gorski (1961) demonstrated that cyclic gonadotropin



release in female rats is regulated from a specific centre in the pre-optic hypothalamic region, whereas tonic gonadotropin response is regulated from the hypothalamic ventromedial arcuate region. Bari Kolata (1979) reviews the recent evidence that (in rats) it is the aromatization of testosterone to oestrogen which is crucial in the sex differences which occur during brain differentiation: when testosterone reaches the brain cells of newborn male rats it is converted to oestrogen and dihydrotestosterone but newborn female rats’ brains are protected from the effects of endogenous oestrogen by a-fetoprotein (a protein made by the fetal liver) which binds oestrogen and thus prevents it from reaching the developing brain. However, animals whose critical period for brain differentiation ends before birth (such as humans) have a-fetoproteins which do not bind oestrogens, and it is not yet clear what mechanisms might protect those animals’ brains from the effects of oestrogen.

Behavioural effects of sex differences in brain differentiation (i) Sexual behaviour

The effects of pre- and perinatal hormones on the sexual behaviour of infrahuman species are reviewed carefully by Hoyenga and Hoyenga (1980), and the interested reader is referred to their text for a detailed list of primary sources. The evidence that early hormones are critical in determining sexual behaviour is unequivocal: neonatal castration of male rats (i.e. deandrogenization) increases all types of female sexual behaviours; and the prenatal androgenization of female rats increases the incidence of mounting and decreases the incidence of lordosis (the female sexual response consisting of concave arching of the back with simultaneous raising of the head and hind-quarters). Comparable evidence is available from primate studies. However, the perinatal administration of androgen to a female rat does not entirely masculinize her complete repertoire of sexual behaviour, any more than the castration of a male entirely suppresses all male-type responses.

Whalen (1974) proposed an orthogonal model of sexual differentiation in which he suggested that ‘during development hormones can defeminize without masculinizing and masculinize without defeminizing, and that hormones can defeminize one behavioural system (e.g. mating) while masculinizing another system’ (p. 469). This conception is not really satisfactory, for if one considers any specific aspect of sexual behaviour (such as lordosis) it is difficult to see how ‘masculinization’ does not also imply ‘defeminization’; however, it does try to deal with the data which indicates that lordosis in the female is not necessarily inhibited by perinatal administration of testosterone, even though she also exhibits increased incidence of mounting. In the same article Whalen raises some important criticisms of the naivety of the behavioural analysis which has often been employed in studies of sexual behaviour, and similar criticism is reiterated


by Beach (1979). Responses such as lordosis can be only partially completed, and neonatally androgenized females do exhibit weak or partial lordosis responses with moderate frequency. Similarly, mounting is not always accompanied by intromission and ejaculation. A fmer-grained categorization of the behavioural units which comprise ‘sexual behaviour’, and due attention to controlling for the stimulus conditions in which it occurs, might facilitate our understanding of its general structure, and thus enhance our knowledge concerning the differential effects of various hormones. Beach suggests that both male and female brains have the appropriate neural substrates for homotypical and heterotypical sexual behaviour, and that sexual differentiation of the brain serves to alter the probability of a particular response being elicited in a given set of stimulus conditions. Thus, demasculinization does not eradicate the possibility of a male type response, it simply reduces its probability of occurrence. Figure 8.1 shows the critical period during which sexual differentiation of the brain occurs in rats. The degree to which the behaviour of the female rat is masculinized is dependent both on dosage and on timing of testosterone administration.

Figure 8.1 The effects of perinatal testosterone or castration on neonatal rats.



There is some interesting evidence from Dorner’s laboratories (Dorner, 1977, 1979) that human sexual behaviour may be affected by the prenatal hormonal environment. ‘An androgen deficiency in genetic males during a critical period of brain organization gives rise to predominantly female differentiation of the brain. This androgen deficiency in early life can be largely compensated by increased hypophyseal gonadotropin secretion in later life. Thus, the predominantly female-differentiated brain is post-pubertally activated by an approximately normal androgen level, leading to homosexual behaviour’ (Dorner, 1979, p. 87). The evidence from which this conclusion is derived comes partly from an experiment in which adult males were given an intravenous oestrogen injection: in homosexual males there was a subsequent rise in LH values above initial levels (a response which would be normal in females), whereas in bisexual and heterosexual males no such rise was detected. Goy and McEwen (1980) express some discomfiture with these data, and in particular point to evidence of time- dependent partial dissociation between the differentiation periods of central nervous centres regulating gonadotropin secretion and those responsible for sexual behaviour. However, Dorner (1977, 1979) clearly believes that the evidence of a relationship between prenatal hormones and adult sexual behaviour is now sufficiently strong to contra-indicate the prescribing of any androgenic or anti-androgenic substances to pregnant women, and recent data on females with CAH may tentatively support this view. In contrast to earlier findings which suggested that CAH females were no different from normal controls in their heterosexual interests and behaviour (Ehrhardt et al ., 1968a; Ehrhardt et a/., 19686) a more recent investigation by Money and Schwartz (1977) has suggested that early treated CAH females may be delayed in establishing dating and romantic interests. In addition, they found that in their sexual fantasies CAH females showed an increased rate of awareness of bisexuality relative to controls (although this did not necessarily reflect actual experience). It is plausible that these more recent data reflect a less prescriptive social climate than that which prevailed during the early 1960s when the original data were presumably collected, and one can only conclude that the nature of the biological, cognitive and social factors which regulate human sexual behaviour are by no means well established. This area remains wide open to debate.

(ii) Non-sexual behaviour

The effects of pre- and perinatal hormones on animals are not restricted to endocrinology and sexual behaviour. Levine (1966) cites evidence which demonstrates that female rats who have been injected as neonates with testosterone show male-type behavioural responses in an open field; and that female rhesus monkeys injected with testosterone in utero show levels of rough and tumble play which are approximately equivalent to those of normal male monkeys. Goy (1968, 1970) reports that initiation of play and pursuit play are


greater in neonatally androgenized female monkeys than in normal females, and a number of workers have reported effects of neonatal hormones on activity (Gray et al ., 1975; Stewart et al ., 1975), exploration (Quadagno et al ., 1972; Gummow, 1975), and learning (Beatty and Beatty, 1970; Dawson, 1972; Dawson et al ., 1973). Quadagno et al. (1977) have reviewed the extensive literature on the effects of perinatal hormones on non-sexual behaviours with particular reference to energy expenditure, maternalism and learning, and they are able to conclude that the effects of early hormones on the behaviour of infrahuman species are well established.

McEwen (1976) and Goy and McEwen (1980) describe the experimental data which have led to the identification of specific neural pathways that are established by the influence of sex hormones and are sexually differentiated, and which underlie sex differences in behaviour. The work of Raisman and Field (1973) represented an important breakthrough in this field: they found that adult female rats have more dendritic spine connections in the preoptic area than males, but that males castrated within 12 hours of birth have spine connections equivalent *o those of the female. They demonstrated that those animals which show frequerd lordosis have different patterns of synaptic connectivity than animals with a limited capacity for lordosis. Various other studies have also shown that the brain of a male rat deprived of androgen and the female exposed to androgen will take on heterotypical characteristics: for example, the size of the cell nuclei in the preoptic area is positively correlated with the degree of lemaleness’ in the rat’s sexual behaviour (Dorner and Staudt, 1968, 1969); both serotonin levels (Ladosky and Gaziri, 1971) and RNA metabolism (Clayton et al ., 1970) are also affected. Litteria and Thorner (1974) and Phillips and Deol (1973) report sex differences in the cerebellum and septum which can be reversed by the presence or absence of androgens. However, even if these differences do indeed underlie the observed differences in behaviour (as seems plausible) and we assume that similar mechanisms of differentiation occur in humans, it is nonetheless unlikely that human behaviour would be so strongly determined by neural networks (particularly in the face of conflicting socialization).

Hormonal anomalies in human development

The data from the above animal studies provide sufficient evidence for the assertion that hormones are critical in determining patterns of brain differentiation, and suggest that pre- and perinatal hormones may also exert long-term effects on behaviour patterns. It is instructive then, to consider the effects of early hormones on human behaviour insofar as this can be achieved within the limitations of ethical considerations (see Reinisch and Gandelman (1978) for an interesting discussion of these issues). It has already been noted that prenatal hormones affect the development of sex-typical physical characteristics, and



individuals with anomalous genital development at birth, or who present with related problems at puberty (e.g. amenorrhea in patients with testicular feminization), have been studied by psychologists interested in the possible effects on hormones on behaviour.

Two clinical syndromes can be regarded as close counterparts of experimental anti-androgenization (or demasculinization) in animals: Turner’s syndrome and testicular feminization due to androgen insensitivity. In Turner’s syndrome the missing chromosome may be either an X or a Y, and if a few androgen-secreting cells remain in the gonadal streak tissue there may be a mild degree of labial fusion and an enlarged clitoris. Some individuals have a 45X/45XY mosaic karyotype: they have testes, but these are not properly formed and are at high risk for cancer (Money, 1911b). Thus, deandrogenization in Turner’s syndrome is due to a failure of the gonads to manufacture androgens; in contrast, other testicular feminization syndromes are a result of the failure of the target organ cells to take up and utilize the androgens which are secreted from testes in foetuses with the normal 46XY karyotype.

The behaviour of girls and women with either Turner’s syndrome or testicular feminization is unequivocally feminine. In the case of Turner’s syndrome there seems even to be a tendency of extreme conformity to female sex stereotypes: they are known to fight less, to be less athletic and to be more interested in personal adornment than control comparisons (Money and Ehrhardt, 1972); and Theilgaard (1972) reported that women with Turner’s syndrome preferred to wear very feminine-style clothing and jewellery. All but one of the 15 girls in the group studied by Money and Ehrhardt (1972) had played exclusively with dolls, and most of them expressed a very strong interest in maternalistic activities associated with child care. In their anticipation and imagery of romance and motherhood, Turner’s syndrome females were found to be no different from their control comparisons. From these data, one may infer that differentiation of a feminine gender role is not dependent on the presence of prenatal gonadal hormones, nor does it require the presence of a second X chromosome. Indeed, Money and Ehrhardt are prepared to assert that ‘a feminine gender identity can differentiate very effectively without any help from prenatal gonadal hormones that might influence the brain and perhaps, in fact, all the more effectively in their absence’ (p. 108).

Babies born with the testicular feminization syndrome look like absolutely normal females, although these females tend to be of above average height (Money, Ehrhardt and Masica (1968) quote a mean height of 5 feet 1\ inches for their sample of ten patients). Diagnosis of their condition normally follows referral for primary amenorrhea so data regarding behaviour in early childhood are necessarily based on retrospective report (which may be influenced by knowledge of their condition). Even with this caveat in mind the data reported by Money et al (1968) and Money and Ehrhardt (1972) do seem to provide


strong evidence for the unequivocal differentiation of female gender role in these patients. They reported playing primarily with dolls in early childhood and having dreams and fantasies which reflected the normal sex-role stereotypes of marriage and motherhood. With one exception these women rated themselves as fully content with the female role, and at adolescence they conformed with the normal patterns of heterosexual behaviour. Most of them expressed positive enjoyment in adopting ‘feminine’ styles of dress and personal adornment. ‘Babies with the androgen insensitivity syndrome who are consistently reared as girls have no uncertainties about themselves as girls, women, wives, sexual partners, and mothers by adoption ... they grow to be womanly in their behaviour, in their erotic mental imagery, and in their self-perception, even when they know the medical terminology of their diagnosis’ (Money, 1977a, p. 262).

Reifenstein’s syndrome resembles that of complete androgen insensitivity except that there is partial masculinization of the genitalia during foetal life and the neonate is thus sometimes classified as a male. At puberty the development of secondary sex characteristics nevertheless proceeds as described above. According to Money and Ogunro (1974) those infants assigned as males did not show any preference in childhood for female-type activities (doll play etc.) and made concerted efforts to compensate for their relative inferiority in athletic pursuits. At puberty, their breasts had to be surgically removed; in adulthood their physiognomy is beardless and unvirilized, and because of their extremely small, surgically repaired genitalia they may encounter some difficulty establishing a sex life (none reported homosexual preference). On the whole, gender identity conforms with socialization and there seems to be no evidence from these cases of any biologically based behavioural imperative for feminization. These cases may reflect the experimentally induced ‘demasculinization’ without accompanying ‘feminization’; and as far as we can tell from these few cases the social environment is a paramount factor in influencing preferred activity and gender identity.

The form of male pseudohermaphroditism described by Imperato-McGinley et al (1974) results from a 5a-reductase deficiency which leads to incomplete differentiation of the external genitalia at birth, and thus a female sex assignment is often made. At puberty, however, differentiation of male characteristics occurs and sex re-assignment is necessary. A recent report by Savage et al (1980) confirms the rather surprising finding that this gender-role transition is made relatively easily and they conclude ‘... that exposure of the brain to androgens during foetal life and thereafter appears to have had more effect on determining gender identity than the pre-pubertal sex of rearing’ (p. 404).

In the light of this conclusion it is interesting to consider the effects of the masculinization of a female foetus. These have been documented in two clinical syndromes: progestin-induced hermaphroditism (PIH) and the adrenogenital syndrome (CAH). PIH occurred following the administration of synthetic



progestins to pregnant mothers with histories of miscarriage; these steroids were devised as substitutes for the pregnancy hormone, progesterone, but because their chemical structure was similar to androgen, they exerted an unexpected masculinizing effect on a female foetus (Walker and Money, 1972). Once this effect was discovered (in the early 1950s) the use of these hormones was discontinued; however, the subsequent development of girls born with PIH has been studied (Ehrhardt and Money, 1967; Money and Ehrhardt, 1972). If the external genitalia were surgically feminized shortly after birth, no further surgical or hormonal treatment was required; this is in contrast to girls born with CAH who require constant maintenance on cortisone to prevent continuing postnatal masculinization and accelerated pubertal development. Table 8.1 summarizes some of the data obtained on the reported behaviour of these cases. Basically, there is little difference between that of the PIH and CAH girls, but both these groups differ significantly from control comparisons on measures of tomboyism, athletic skills and preference for boys’ toys (e.g. cars, guns etc.). Perhaps, as a result of these interests, it is not surprising that these girls also prefer male playmates.

The accuracy of assessment of behaviour in these cases is difficult to evaluate and Ehrhardt and Baker (1974) are clearly aware of this when they discuss, in some detail, exactly how the interviews with patients and their parents were conducted. It is important to be aware'that no observations were made of the

Table 8.1 Behavioural effects of prenatal exposure to androgens*

Childhood behaviour




above average

above average

Athletic interests and skills

above average

above average

Preference for male playmates

above average

above average

Preference for ‘functional’ clothing

above average

above average

Preference for toy cars, guns etc. over dolls

above average

above average

Anticipation of future

Priority of career over marriage

above average

above average

Heterosexual romanticism



Anticipation of pregnancy


Less frequently reported than controls

Dissatisfaction with female role



Sexual behaviour

Childhood-shared genital play/copulation play



Adolescent boyfriend and dating



Bisexual/homosexual fantasy

(data not available)

above average

Bisexual/homosexual behaviour


within normal range

  • Data adapted from Ehrhardt (1977); Ehrhardt and Baker (1974); Epstein and Money (1968);

Ehrhardt and Money (1967); Money and Ehrhardt (1972); Money and Schwartz (1977).


children and that reliability was assessed purely in terms of the concordance between the mother’s and child’s reports. Even so, these data do seem to reflect a tendency for increased activity in females who have been exposed to abnormally high levels of androgen in utero ; and compatible with these tomboyish interests, these girls also seem less interested than control comparisons in personal adornment and maternal behaviours. Their gender identity is nonetheless entirely female (although 35% of them said they would not mind being a boy).

It appears then, that the effect of prenatal androgens on gender identity cannot be as imperative as Imperato-McGinley et al. (1974) and Savage et al (1980) suggest; it is more likely that the activational effects of circulating male hormones at adolescence are crucial to the satisfactory transition to the male gender role for these male pseudo-hermaphrodites. However, the surmise that the behavioural development of CAH and PIH females is in some way analogous to that of prenatally androgenized monkeys (Goy, 1968) is certainly supported by the available data. Furthermore, it is interesting to note that whilst the excess of androgens may be contributing to a masculinizing effect on some behaviours it does not have a global ‘defeminizing’ effect. Indeed, a sample of late-treated CAH patients described by Ehrhardt, Evers and Money (1968) conform strongly to female sex stereotypes in their careers and/or marriages. In fact, the influence of prenatal androgen exposure is probably limited to a specific effect which in some way creates a predilection for physical energy expenditure; associated preferences for functional clothing and male playmates may be no more than a reflection of this basic trait. This conclusion is confirmed to some extent by the finding that males with CAH are no different from a comparison group of unaffected male siblings except that they are more frequently (80 % of CAH males: 20% sibs) reported to engage in intense energy expenditure (Ehrhardt and Baker, 1975).

In two studies (Zussman et al. (1975) cited in Goy and McEwen, 1980; Ehrhardt et al, 1975) which considered the effects of prenatal progesterone on childhood behaviour (not the androgenic progestins which caused PIH), subjects were found to exhibit lower energy levels and a tendency to prefer ‘female type’ clothing styles. They suggest that non-androgenic progestins may actively counteract androgen effects in utero in both males and females.

During childhood, then, the major behavioural effect of prenatal androgenic hormones is on activity level: when the foetus has been exposed to androgen, he/she will subsequently display a predilection for high levels of physical energy expenditure (and these effects appear to be dose-related). These results are consonant with the findings on the effects of androgens in rodents and primates (Quadagno et al , 1977), and they do suggest that these hormones have an organizing effect on brain differentiation which will usually be sexually dimorphic.




Reinisch (1977) argues that prenatal exposure to (non-androgenic) progestin also has long-term effects on personality, and her data confirm the earlier suggestions of Ehrhardt and Money (1967) that progestin-exposed subjects show high levels of self-assertive independence and self-reliance. Twenty-six subjects, whose mothers had been administered a minimum dosage of 40 mg progestin for at least four weeks during the first trimester of pregnancy, were tested on age-appropriate Cattell Personality Questionnaires. They exhibited high scores on individualistic, self-assured and self-sufficient factors relative to sibling controls. In contrast, subjects exposed to high oestrogen levels in utero were found to be more group-dependent and group-oriented than a sibling control group.

An investigation by Yalom, Green and Fisk (1973) also attempted to evaluate the long-term effects of prenatal oestrogens on personality. Because diabetic women produce lowered levels of oestrogen and progesterone during pregnancy they are sometimes prescribed supplemental doses of these hormones; Yalom et al. studied the male children of diabetic mothers who had received high oestrogen doses, and compared them to a control group of children with normal mothers and a group of children of untreated diabetic mothers. At the age of 6 the boys who had been exposed to the highest levels of oestrogen were rated by their teachers as being less assertive and less athletic than their male peers. By the age of 16 a whole range of behaviours seemed to be related (albeit weakly) to the level of oestrogen exposure: athletic coordination, competitiveness, assertiveness, aggression, and global measures of ‘masculinity’. The children of diabetic mothers who had not received oestrogen supplements were consistently more masculinized than the control group of sons of normal mothers, and the children of mothers who had received supplemental oestrogen were the least masculine. It is possible that some of these effects may be due to differing levels of activational hormones in these boys since the development of the testes and output of testicular hormones are likely to have been affected (Zondek and Zondek, 1974).

Cognitive ability

In an exhaustive review of psychological sex differences, Maccoby and Jacklin (1974) concluded that males show superior visuo-spatial and mathematical abilities relative to females. Females though, are better at some verbal skills: they are more fluent, they are better readers and spellers, and their speech is more comprehensible than that of males (Harris, 1977). The extent to which these differences reflect underlying differences in neural organization has been a matter of considerable debate (Archer, 1976) since the influence of differential


socialization in the development of sex-typed abilities is difficult to evaluate. Males and females show similar rates of early babbling (Moss, 1967; Lewis, 1972), but by six months of age girls receive more physical, visual and vocal contact with their mothers (Goldberg and Lewis, 1969; Messer and Lewis, 1972). Infant boys are encouraged more than girls to explore and to be independent of their mothers (Baumrind and Black, 1967; Hoffman, 1972). McGuinness (1976) argues convincingly that sex differences in cognitive abilities may develop from fundamental differences in auditory and visual acuity—from an early age females show lower auditory thresholds and superior pitch discrimination compared to males, and the sex difference increases with higher frequencies and with age (McGuinness, 1972); males have superior foveal vision, greater sensitivity to light and longer photopic persistence.

The aspect of spatial ability in which males most consistently excel is the capacity to rotate mentally three-dimensional images, or to redefine visual images into new planes; males thus perform better on mathematical problems which require spatial visualization (Fennema and Sherman, 1977; Petersen, 1979) and which involve the ability to ‘break set’ and restructure (Garai and Scheinfeld, 1968; Hutt, 1972a, b). Until adolescence, the majority of studies show no sex differences in quantitative skills, but males move ahead after this point and show consistently superior performance (Maccoby and Jacklin, 1974).

If these sex differences in cognitive abilities are subserved by the neural organizing effects of androgens in utero, a sample of females exposed prenatally to androgen would be expected to show a male pattern of abilities. Similarly, if enhanced oestrogen levels affect the neural substrates of verbal behaviour then males exposed to supplemental oestrogen in utero would show a female pattern of abilities. In fact, neither of these hypotheses is substantiated by the available data.

Ehrhardt and Money (1967) report identical mean verbal and performance IQ scores for a PIH sample of ten females (mean verbal IQ = 125, s.d. = 11.4; mean performance IQ = 125, s.d. = 12.5). Although Perlman (1971) (cited in Reinisch et al., 1979, and in Baker and Ehrhardt, 1974) found that CAH girls performed significantly lower than their matched controls on Verbal and Comprehension sub-tests of the Wechsler IQ scale, they also scored lower on Block Design. However, the scores of CAH girls on the Healy Pictorial Completion Test were comparable to those of CAH and normal boys; Perlman suggests that this result may reflect the higher activity levels of the CAH girls which would have made them more familiar with the kinds of situations depicted on the test. Baker and Ehrhardt (1974) report no statistically significant difference on perceptual or verbal factors between AGS patients and sibling control comparisons, although the trends were in the expected direction (i.e. CAH females performed slightly less well on the verbal sub-tests of the WISC than their unaffected female siblings, but slightly better on the perceptual sub-tests).



Curiously, patients exposed to prenatal androgen do seem to have above average IQ scores, but close examination of the relevant data reveal this finding to be due to factors other than the androgenic influence. Baker and Ehrhardt (1974) tentatively suggest that the recessive genetic trait for CAH may somehow be linked to another trait which favours postnatal intellectual development, and this notion is supported by the finding that the IQ levels of CAH patients do not differ significantly from those of their parents and siblings which are also higher than normal. The elevated IQ of the PIH group (Ehrhardt and Money, 1967) can be ascribed to social class factors among the parents: six of the nine families involved in this study had at least one parent who was a college graduate. Thus, there is no substantial evidence to link prenatal androgens with enhanced IQ scores.

Dalton (1968, 1976) suggested that prenatal progesterone (not of the androgenic type) increased intellectual achievement, but these data were not replicated in a study reported by Reinisch and Karow (1977) and have been discredited on statistical and theoretical grounds (Lynch et al ., 1978; Lynch and Mychalkiw, 1978).

The only study to consider the effects of prenatal oestrogen on cognitive ability is that of Yalom et al. (1973). These (male) subjects were administered the Embedded Figures Test to evaluate their spatial ability: those boys who had been exposed to supplemental oestrogen in utero showed slightly inferior performance relative to the two comparison groups, but this result did not reach statistical significance.

Other hormonally anomalous clinical conditions in no way implicate the role of prenatal hormones in determining the future patterns of intellectual abilities. Patients with testicular feminization show the typical female pattern of lower spatial than verbal ability: in the study reported by Masica et al. (1969) a sample of fifteen cases had a mean Wechsler verbal score of 111.8 and a mean performance score of 102.3. Since their exposure both to hormones and socialization is equivalent to that of genetic females, one can conclude from these data simply that superior male visuo-spatial abilities are not genetically determined from a locus on the Y chromosomes. For some time it was thought that spatial ability was partly determined by a locus on an X-linked gene (O’Connor, 1943; Stafford, 1961), but recent data indicate that the pattern of spatial abilities within familial groups is better explained by a model of an autosomal dominant gene which has reduced penetrance in females (Fain, 1976, cited in Vandenberg and Kuse, 1979). Whether this mechanism might influence brain differentiation must be purely speculative, and there is, as yet, no evidence to this effect.

Turner’s syndrome females have IQ scores within the normal range (Money, 1964; Shaffer, 1962), but also tend to show specific deficiencies in spatial ability. Shaffer (1962) quotes a mean verbal IQ of 106, but a mean performance IQ of 88.


Alexander, Ehrhardt and Money (1966) showed that Turner’s syndrome females experienced great difficulty on a visual memory test which requires the reproduction of angulated shapes, and Theilgaard (1972) reported that they performed badly on an embedded figures task.

It is reasonable to speculate from these data that androgens play some role in facilitating spatial ability. Since Turner’s syndrome females produce no androgens, and testicular feminized patients are insensitive to their effects, spatial ability is thus slightly impaired. In the oestrogen-exposed patients, the testes may have been producing less androgen than normal (Zondek and Zondek, 1974). The data from the CAH patients indicate that it is not the prenatal hormonal environment which is crucial, so the effect of androgens on spatial ability appears to be activational rather than organizing. This conclusion is supported by data from Petersen (1979) which indicate that females with androgynous somatic characteristics have better spatial ability than their more ‘feminine’ peers.

Similarly, the effect of oestrogens on verbal ability may also be an activational one. Dawson (1972) reports a study of West African males feminized by kwashiorkor-induced endocrine dysfunction. In severe cases of kwashiorkor the liver becomes unable to inactivate the normal amount of oestrogen which the male produces, and Dawson found that males with this condition had ‘significantly lower spatial ability and a more feminine field-dependent cognitive style than controls. In addition these subjects had significantly lower numerical and higher verbal ability compared to normal males’ (p. 24). Presumably though, these males had had equivalent gestational experiences to their controls and so the prenatal hormonal environment is not implicated in these results.

Sex differences in postnatal brain development

The human brain is not fully mature until around sixteen years of age. The main ‘growth spurt’ of the human brain begins during the last trimester of pregnancy and continues into the second year of life. During this period there is an increase in the number of glial cells (from which myelin is derived) and hypertrophy of all cells, specifically in the form of increased axonal terminal and dendritic branching (i.e. interneuronal connections). Most cortical areas are fully myelinated by the child’s third year but myelination of the reticular formation the cerebral commissure, and the intracortical association areas may continue into the second and third decades of life (Marshall, 1968).

Recent evidence suggests that sex differences in brain development are partly reflected in sex differences in hemispheric specialization (Hutt, 1979a; McGlone, 1980). For example, Witelson and Pallie (1973) in a study of infants up to 3 months old, reported that the increased size of the left (relative to the right)



temporal planum (the posterior surface of the temporal lobe, including part of Wernicke’s area which subserves language) was significant in neonate females but not in males (although a significant difference was found for slightly older (20-90 days) males). Buffery and Gray (1972) cite evidence that in four-year-old girls the degree of myelination in the temporal planum is greater than that for four-year-old boys and they suggest that this may account for the female precocity in language development.

Witelson (1976) describes an experiment which suggests that in boys, the right hemisphere is specialized for spatial processing from as early as six years of age, whereas females show evidence of bilateral representation. She suggested that this specialization might subserve superior spatial skills in males. Levy (1969) postulated that bilateral representation of language in females could interfere with the development of spatial processing abilities in the right hemisphere— thus the cerebral organization which is presumed to give females an advantage in language development and verbal abilities may serve also to impede their development of spatial skills.

Waber (1976) has argued that lateralization is a function of maturation rate rather than sex. On the whole, girls mature faster than boys and generally display a greater tendency towards bilateral representation of skills. However, late-maturing adolescents of either sex are more likely to be strongly lateralized than their early-maturing peers, and are also more likely to show evidence of superior spatial skills.

The evidence for the existence of anatomical substrates which would underlie lateralization processes is both limited and confusing. For example, Wada et al (1975) did not replicate Witelson and Pallie’s (1973) findings on sex differences in cerebral asymmetry in infants: they report that both male and female infants tend to have a larger temporal planum in the left hemisphere than in the right, yet adult females are more likely than males to show the reverse pattern of asymmetry. It is possible that the anatomical asymmetry reported for adult females is a reflection of the greater plasticity of localization of function in females than in males. In a recent report (Hughes et al , 1980), females performed faster on a task which had both a verbal and a visuo-spatial component, whereas both sexes performed at the same speed on the verbal task alone. The authors interpret this finding as reflecting the ability of females to process both aspects of the task in one hemisphere; in males additional time is needed to complete the combined task because information has to be transferred between hemispheres. However, this sort of speculation awaits support or rebuttal from further anatomical evidence.

The ontogeny of hemispheric specialization and lateralization is simply not yet adequately charted. It is not known whether (or how) the environment might modify lateralization and thus we cannot know whether the data of Wada et al. (1975) from adult females are the result of endogenous, hormonally-mediated,


changes or a reflection of educational experience. Tomlinson-Keasey and Kelly (1979) report that lack of early hemispheric specialization is predictive of better reading skills, and that right hemisphere specialization is positively associated with mathematical skills —data which confirm stereotypic achievements (i.e. females tend to be less lateralized and are better readers, males show a greater tendency towards lateralization of spatial skills in the right hemisphere and are better at mathematics). However, the nature of these relationships needs to be carefully explored.

Witelson (1977) argues that the functional neural substrates for the lateralization of particular abilities may show a plasticity during development which is lost in adulthood, and that this plasticity may reflect a susceptibility to environmental influences. However, experimental support for such an idea is still thin. It may be that lateralization predisposes cognitive strategies and atten- tional biases rather than specific skills. The female precocity in language development leads to a preferential use of language as a processing mode and consequent inferior performance in visuo-spatial skills (McGlone and Kertesz, 1973). Bryden (1979) offers a review of experimental data which serve as a useful reminder that sex differences in cerebral organization are not clearly defined: the degree of overlap between the sexes is often substantial, and seems to vary as a function of the experimental paradigm.

How different are sex differences?

It is easy to fall into the habit of discussing sex differences in ability and behaviour as though these represented absolute differences between two quite distinct populations. A sex difference in mean scores on a particular ability tends to deflect our attention from the within-group variances which indicate how much the groups overlap. Even when there is a statistically significant difference between the mean scores of males and females on a test the majority of both sexes may score within the same range.

The interpretation of a report of sex differences will depend on whether one is concerned with socio-political/practical issues (in which case differences are often trivial and meaningless) or with scientific/theoretic issues (in which case small but consistent differences may yield important insights). Thus, consistent reports of sex differences in verbal and visuo-spatial skills have raised interesting hypotheses both about hemispheric specialization, and the role of prenatal hormones; they do not, of course, provide any justification for boys to do badly when studying modern languages or for girls to abandon mathematics education at the earliest possible opportunity.

It may be that sex differences in certain skills are a result of long-term evolutionary pressures. For example, Hutt (1972 b) argued that athletic and visuo-spatial skills in males maximize hunting success and thus increase the



probability of survival, whereas the socially communicative abilities and superior manual dexterity of the females have evolved as an adaptive consequence of her predominantly nurturant role in caring for dependent infants. Yet characteristics with a presumed evolutionary adaptive basis are not fixed for every individual: cultural pressures will influence the expression of abilities, and ‘typical’ sex differences are simply not found in some cultures. For example, cross-cultural studies of field-independence (presumed to be related to visuo- spatial skills) reveal no sex differences in Eskimo and Zambian cultures (MacArthur, 1967; Siann, 1972). McGuinness (1976) argues that \ .. the fact that boys do learn to read and write fluently, suggests that though initial processes may be guided by certain sensory differences, there is no reason to assume that these differences must remain. Parents insist that boys learn to speak, read and write but no such insistence induces the females to learn about spatial- mechanical relationships’ (p. 144).

In our own culture then, there is an attempt to educate males in heterotypical skills whereas the converse is not true for females. Even in homes where parents believe that they do not discriminate between male and female children Rheingold and Cook (1975) found that \ .. the rooms of boys contained more animal furnishings, more educational art materials, more spatio-temporal toys, more sports equipment and more toy animals. The rooms of girls contained more dolls, more floral furnishings, and more “ruffles’” (p. 461). The agents of socialization are evidently insidious, and may tend to exaggerate sex differences in proclivities for particular forms of behaviour. Even if we accept that evolutionary pressures have resulted in sex differences in neural organization which may differentially predispose males and females to specific abilities, we must stand this against our knowledge that the plasticity of the human brain will probably enable us to modify the expression of those abilities. This in turn implies a responsibility of educators and caretakers to provide an appropriate range of educational opportunities and role exemplars for their male and female charges.

At present, not only are females less likely to be given the wide range of toys that males have, their role is also under-represented by the literature and television media: \ .. females were under-represented in the titles, central roles, pictures, and stories of every sample of books we examined ... Even when women can be found in books, they often play insignificant roles, remaining both inconspicuous and nameless’ (Weitzman et al ., 1972). Sternglanz and Serbin (1974) made a study of T.Y. programmes with high popularity ratings, and found that half of these programmes did not portray any female roles: of those that did, the authors comment ‘female children are taught that almost the only way to be a successful human being if you are a female is through the use of magic’ (p. 714). Exposed to these kinds of socialization pressures it comes as little surprise that females tend to be diffident about their own ability and are


particularly unwilling to tackle those skills which they perceive as falling within the male domain (Hutt, 1979b; Byrne, 1978).

Socialization experiences in our culture thus tend to exaggerate a dichotomy of roles and abilities between males and females. Historically, this socialization has acted to repress the female more than the male (her rights to be educated and to vote have, after all, been won only comparatively recently), but there is little doubt that a deliberate educational policy could serve to increase the range of both male and female behaviour. Goy and McEwen (1980, pp. 60-61) present some interesting evidence that female attachment to an infant may be innate (its expression being in part activated by elevated hormonal levels during pregnancy and birth), whereas male attachment is socially learned. This in no way implies that males are unable adequately to perform parenting behaviours, but its expression may be subserved by different neural mechanisms. There is no reason to believe that the expression of sex typical intellectual abilities is any less modifiable.

Towards a model of human sex differences

Waddington’s (1957) notion of‘canalization’ in an epigenetic landscape provides a useful conceptualization for understanding differences in the degree of sexual dimorphism in behaviour. Waddington suggested that, for all members of a species, a set of target physical characteristics (eyes, arms, legs etc.) is defined by the genotype and, despite underlying genetic variability, genetic processes operate together to ensure that these targets are achieved. He depicts the development pathways of the phenotype as a ball rolling through a set of valleys (the epigenetic landscape); the valleys can vary in steepness and thus vary the opportunity of the phenotype to deviate from a given course—-the steepness of the valley reflects the degree of canalization. At certain critical points in development, when the phenotype is undergoing rapid change, it is susceptible to certain environmental or genetically induced stresses. For example, the embryological development of the arms takes place around 38-48 days (postmenstrual); this development is strongly canalized (i.e. all normal humans have arms) but a teratogen, such as thalidomide, taken by the mother during this critical period will inhibit this normal phenotypic development and the foetus will eventually be born either with no arms at all or severe under-development (see Fishbein, 1976, pp. 46-47). The development of some other physical characteristics may also be affected by thalidomide during this period, but in general each character has its own critical period.

As we noted earlier, the basic developmental trend of the body’s sexual characteristics is in a direction corresponding to that of the homozygous sex (i.e. the female). This trend is canalized to develop a female foetus from the zygote which is formed at conception. However, if the embryonic gonad differentiates to



form a testis, then phenotypic development is deflected from the female pathway when the testis begins to secrete MIS and androgens. Differentiation of neural networks occurs in the same manner: there will be critical periods when the presence or absence of biochemical agents (usually hormones) will affect the development of RNA which is specific to particular structures. Behaviours which depend on these specific neural anlagen for their expression will subsequently be affected. The specificity of hormonal effects is well illustrated by some data presented by Short (1979): certain aspects of male-type sexual behaviour were exhibited by ewes which were androgenized late during gestation (days 50-100, or 70-120) and had essentially female external genitalia; androgenization during early gestation (days 30-80) resulted in complete masculinization of the external genitalia which was not accompanied by male type sexual behaviour. Masculinization of urination behaviour could be effected by androgenization at a relatively late period of gestation after it was no longer possible to masculinize sexual behaviour. The positive feedback effect of oestrogen on LH, which is normally exhibited only as a female characteristic, was sometimes abolished by androgenization but ‘gave no clue whatsoever to the type of sexual behaviour to expect from the animal’ (p. 258). This example illustrates clearly that the degree of masculinization of behaviour cannot be inferred from physical characteristics. The developmental pathways for specific behaviour patterns are also independent of one another, so the masculinization of sexual behaviour does not necessarily imply masculinization of activity levels.

If we return to the image of males and females rolling through their (sometimes overlapping) epigenetic landscapes, it is possible to visualize the way in which different levels of canalization will result in different degrees of sexual dimorphism in the eventual expression of behaviour. If the pathway for a particular neural substrate is very steep, it will be difficult to deflect the phenotype from its developmental path—thus most genetic females will manifest the appropriate ‘female’ behaviour pattern and most males will not. If the pathways are gentle then the phenotypes may be spread more thinly, and a linear male-female dimension may be evident in the subsequent behavioural pattern. The neural mechanism which mediates gonadotropin release is clearly strongly canalized, other neural substrates in humans are less strongly canalized and therefore enable the expression of greater variability in behaviour and skills.

The critical periods for development of the neural networks which underlie particular behaviours may vary in length; they may overlap in time, but they are independent of one another. Thus, in the female rat, by the judicious administration of neonatal testosterone, it is possible to decrease the incidence of lordosis but not to increase the incidence of mounting. Armstrong’s attempt (cited in Jost, 1974) to relate sexual orientation to body type is therefore quite misconceived: there is no reason to believe the homosexuals will have heterotypical body characteristics.


There are three ways in which hormones can act on the brain to produce sex- differentiated effects. The prenatal hormones organize neural networks in distinctively male and female patterns; they also have a critical role in the development of physical characteristics. Postnatally, the output of gonadotropic hormones can activate these neural networks (for example, in the control of the menstrual cycle). Alternatively, sex-related hormones may have independent effects: an example of this is the yawning behaviour of rhesus monkeys which is normally displayed more frequently by males than by females, but which can be increased in the female by the administration of exogenous testosterone (Goy and McEwen, 1980).

In humans, evolution has operated to permit a high degree of behavioural phenotypic plasticity, which would in turn imply weak canalization of the neural mechanisms which subserve particular behaviours and abilities. (This may account for the conflicting findings in anatomical studies of hemispheric asymmetry, supra). This phenotypic plasticity enables individuals with very different genotypes to exhibit similar or identical behaviour. In the expression of human behaviour and ability then, phenotypic plasticity and not biological canalization may produce conformity of behaviour within a same-sex group. Evidence for sexually differentiated canalization of a particular behaviour requires that its manifestation be virtually universal and not restricted to a single cultural group: the only behaviours which fulfil this stringent requirement are indeed ‘menstruation, gestation and lactation’ in females and ‘impregnation’ by males. The weaker canalization of sex-related brain differentiation in humans relative to infra-human species would also lead us to expect less sexual dimorphism of behaviour in humans; this should always be borne in mind when extrapolating from animal to human studies. Rodent studies have been crucial in extending our knowledge of brain differentiation—they do not necessarily tell us much about human behaviour.

The identification of the brain areas and mechanisms which subserve particular behaviours or the articulation of specific cognitive skills in humans is by no means straightforward. Whilst it has been possible in infra-human species to identify the critical period of development for the expression of certain behaviours (vide Short, 1979) this has not been possible in humans and probably (for ethical reasons) never will be. Nor is it entirely clear how circulating hormones affect human behaviour: studies of the menstrual cycle produce conflicting evidence (Hutt et al ., 1980), and studies of sexual behaviour (e.g. Bancroft and Stakkebaek, 1979) or cognitive ability (Peterson, 1979) have not yielded definitive conclusions. We also lack evidence on the way in which educational experience affects brain development. Thus, information which is vital to a definitive model of the effects of brain differentiation on psychological sex differences is not available. Nonetheless, the existing evidence leaves no doubt that brains of males and females differ as a function both of the prenatal



environment and subsequent maturational effects. These differences may well underlie the predilection for males and females to act in particular ways, but they cannot be seen as constituting a biological imperative.


De facto sex differences in ability and behaviour are not as dimorphic as sex differences in physical characteristics. The major influences on sex differential brain development are the sex steroids (androgens, oestrogens and progestins), but their effect on behaviour is attenuated by the weak canalization of the human behavioural repertoire. Sexual differentiation of the brain may create predilections for particular behaviour and specific cognitive strategies, but it does not constitute a biological imperative for psychological sex differences.


I would like to thank Dr. J. E. Blundell and Professor J. Scott for their helpful comment and criticism during the preparation of this manuscript. I am also grateful to Derrick Pritchatt who translated the article by Reznikov (1978).


Alexander, D., Ehrhardt, A. A. and Money, J. (1966) Defective figure-drawing, geometric and human, in Turner’s syndrome. J. Nerv. Ment. Dis., 142, 161-167.

Archer, J. (1976) ‘Biological explanations of psychological sex differences’, in Exploring Sex Differences (eds. B. Lloyd and J. Archer), Academic Press, London, 241-266.

Baker, S. W. and Ehrhardt, A. A. (1974) ‘Prenatal androgen, intelligence and cognitive sex differences’, in Sex Differences in Behaviour (eds. R. C. Friedman, R. M. Richart and R. L. Vande Wielde), John Wiley, New York, 53-76.

Bancroft, H. and Skakkebaek, N. E. (1979) ‘Androgen and human sexual behaviour’, in Ciba Foundation Symposium, 62, Sex, Hormones and Behaviour, Elsevier/N. Elolland, 209-226.

Bari-Kolata, G. (1979) Sex hormones and brain development. Science, 205, 985-987.

Barraclough, C. A. and Gorski, R. A. (1961) Evidence that the hypothalamus is responsible for androgen-induced sterility in the female rat. Endocrinol, 68, (1), 68-79.

Baumrind, D. and Black, A. E. (1967) Socialization practices associated with dimensions of competence in pre-school boys and girls. Child Devel, 48, 187-194.

Beach, F. A. (1979) ‘Animal models for human sexuality’, in Ciba Foundation Symposium, 62, Sex, Hormones and Behaviour, Elsevier/N. Holland, 113-143.

Beatty, W. W. and Beatty, P. A. (1970) Effects of neonatal testosterone on the acquisition of an active avoidance response in genotypically female rats. Psychon. Sci., 19, 315-316.

Bryden, M. P. (1979) ‘Sex-related differences in cerebral organization’, in Sex-related Differences in Cognitive Functioning-Developmental Issues (eds. M. A. Wittig and A. C. Petersen), Academic Press, New York, 121-143.

Buffery, A. and Gray, J. (1972) ‘Sex differences in the development of spatial and linguistic skills’, in Gender Differences — 77 leir Ontogeny and Significance (eds. C. Ounsted and D. Taylor), Williams & Wilkins, Baltimore.

Byrne, E. M. (1978) Woman and Education. Tavistock, London.

Clayton, R. B., Kogura, J. and Kraemer, H. C. (1970) Sexual differentiation of the brain: effects of testosterone on brain RNA metabolism in newborn female rats. Nature, 226, 810-812.


Dalton, K. (1968) Ante-natal progesterone and intelligence. Brit. J. Psychiatr., 114, 1377-1382.

Dalton, K. (1976) Prenatal progesterone and educational attainments. Brit. J. Psychiatr., 129, 438-442.

Davidson, E. H. (1965) ‘Hormones and genes’, in Hormones and Reproductive Behaviour (eds. R. Silver and II. H. Feder), W. H. Freeman and Co., 1979, 84-93.

Dawson, J. L. M. (1972) Effects of sex hormones on cognitive styles in rats and men. Behav. Genet., 2, 21-42.

Dawson, J. L. M., Cheung, Y. M. and Lau, R. T. S. (1973) Effects of neonatal sex hormones on sex- based cognitive abilities in the white rat. Psychologia, 16, 17-24.

Dorner, G. (1977) ‘Sex-hormone-dependent brain differentiation and reproduction’, in Handbook of Sexology (eds. J. Money and H. Musaph), Elsevier/N. Holland.

Dorner, G. (1979) ‘Hormones and sexual differentiation of the brain’, in Ciba Foundation Symposium, 62, Sex, Hormones and Behaviour, Elsevier/N. Holland, 102-112.

Dorner, G. and Staudt, J. (1968) Structural changes in the preoptic anterior hypothalamic area of the male rat following neonatal castration and androgen substitution. Neuroendocrinol, 3, 136-140.

Dorner, G. and Staudt, J. (1969) Structural changes in the hypothalamic ventromedial nucleus of the male rat following neonatal castration and androgen treatment. Neuroendocrinol, 5,103-106.

Ehrhardt, A. A. (1977) ‘Prenatal androgenization and human psychosexual behaviour’, in Handbook of Sexology (eds. J. Money and H. Musaph), Elsevier/N. Holland.

Ehrhardt, A. A. and Baker, S. W. (1975) ‘Fetal androgens, human CNS differentiation, and behaviour sex differences’, in Sex Differences in Behaviour (eds. R. C. Friedman, R. M. Richart and R. L. Vande Wiele), John Wiley, New York, London, 33-51.

Ehrhardt, A. A., Epstein, R. and Money, J. (1968) Fetal androgens and female gender identity in the early-treated adrenogenital syndrome. Johns Hopkins Med. J., 122,160-167.

Ehrhardt, A. A., Evers, K. and Money, J. (1968) Influence of androgen and some aspects of sexually dimorphic behaviour in women with the late-treated adrenogenital syndrome. Johns Hopkins Med.J., 123,115-122.

Ehrhardt, A. A., Grisanti, G. C. and Meyer-Bahlburg, H. F. L. (1977) Pre-natal exposure to medroxyprogesterone acetate (MPA) in girls. Psychoneuroendocrinol, 2, 391-398.

Ehrhardt, A. and Money, J. (1967) Progestin-induced hermaphroditism: IQ and psychosexual identity in a study of ten girls. J. Sex Res., 3, 83-100.

Fain, P. R. (1976) Major gene analysis: an alternative approach to the study of the genetics of human behaviour. Unpublished doctoral dissertation, University of Colorado.

Fennema, E. and Sherman, J. (1977) Sex related differences in mathematics achievements, spatial visualization and affective factors. Amer. Educ. Res. J., 14, 51-71.

Fishbein, H. D. (1976) Evolution, Development and Children's Learning. Goodyear Publishing Co., California.

Garai, J. E. and Scheinfeld, A. (1968) Sex differences in mental and behavioural traits. Genet. Psychol. Monogr., 77,169-299.

Goldberg, S. and Lewis, M. (1969) Play behaviour in the year-old infant: early sex differences. Child Devel, 40,21-31.

Goy, R. W. (1968) ‘Organising effects of androgen on the behaviour of rhesus monkeys’, in Endocrinology and Human Behaviour (ed. R. P. Michael), Oxford University Press, 12-31.

Goy, R. W. (1970) ‘Early hormonal influences on the development of sexual and sex-related behaviour’, in The Neurosciences: second study programme (ed. F. O. Schmitt), Rockefeller University Press, New York, 199-207.

Goy, R. W. and McEwen, B. S. (1980) Sexual Differentiation of the Brain. MIT Press, Cambridge, Mass.

Gray, J. A., Drewett, R. F. and Lalljee, B. (1975) Effects of neonatal castration and testosterone injection on adult open-field behaviour in rats with atypical sex differences in defecation. Animal Behav., 23, 773-778.

Gummow, L. J. (1975) Postnatal androgenization influences social behaviour of adult rats tested in standard male and female sexual paradigms. Behav. Biol, 13, 385-399.

Harris, G. W. (1964) Sex hormones, brain development and brain function. Endocrinol, 75, 627-648.

Harris, G. W. (1970) Hormonal differentiation of the developing central nervous system with respect to patterns of endocrine function. Phil. Trans. Roy. Soc. Lond., 259,165-177.



Harris, L. J. (1977) ‘Sex differences in the growth and use of language’, in Women: a Psychological Perspective (eds. E. Donelson and J. E. Gullahorn), John Wiley, New York, 79-94.

Hoffman, L. W. (1972) Early childhood experiences and women’s achievement motives. J. Social Issues, 23,129-155.

Hoyenga, K. B. and Hoyenga, K. T. (1979) The Question of Sex Differences. Little, Brown and Co., Boston/Toronto.

Hughes, M., Wilson-Derose, M. and Kiely, B. (1980) Can alphabet recall be part of a visuo-spatial task? Brit. J. Psychol. In press.

Hutt, C. (1972 a) ‘Neuroendocrinological, behavioural and intellectual aspects of sexual differentiation in human development’, in Gender Differences: Their Ontogeny and Significance (eds. C. Ounsted and D. Taylor), Churchill Livingstone, Edinburgh.

Hutt, C. (19726) Males and Females. Penguin, Harmondsworth.

Hutt, C. (1979a) Cerebral asymmetry and hemispheric specialization: some implications of sex differences. Int. J. Behav. Devel, 2, 73-86.

Hutt, C. (19796) Why do girls underachieve? Trends in Education, 4, 24-28.

Hutt, S. J., Frank, G., Mychalkiw, W. and Hughes, M. (1980) Perceptual-motor performance during the menstrual cycle. Hormones and Behaviour, 14, 116-125.

Imperato-McGinley, J., Guerrero, L., Gautier, T. and Peterson, R. E. (1974) Steroid 5a-reductase deficiency in man: an inherited form of pseudohermaphroditism. Science, 186, 1213-1215.

Jolly, C. J. (1970) The seed-eaters: a new model of hominid differentiation based on a baboon analogy. Man, 5, 5-26.

Jost, A. D. (1971) ‘Development of sexual characteristics’, in Human Reproduction (from the Science Journal), Paladin, London, 94-108.

Jost, A. D. (1979) ‘Basic sexual trends in the development of vertebrates’, in Sex, Hormones and Behaviour, Ciba Foundation Symposium, 62, Elsevier/N. Holland, 5-18.

Ladosky, W. and Gaziri, L. C. J. (1970) Brain serotonin and sexual differentiation of the nervous system. Neuroendocrinol., 6, 168-174.

Leujeune, J. (1964) The 21-trisomy—current stage of chromosomal research. Progr. Med. Genet., 3, 144-177.

Levine, S. (1966) ‘Sex differences in the brain’, in Hormones and Reproductive Behaviour (eds. R. Silver and H. H. Feder), W. H. Freeman, London, 1979, 75-80.

Levy, J. (1969) Possible basis for the evolution of lateral specialization of the human brain. Nature, 224,614-615.

Lewis, M. (1972) State as an infant-environment interaction: an analysis of mother-infant behavior as a function of sex. Merrill Palmer Quart., 18, 95-121.

Litteria, M. and Thorner, M. W. (1974) Inhibition in the incorporation of ( 3 H) lysine in the Purkinje cells of the adult female rat after neonatal androgenization. Brain Res., 69, 170-173.

Lynch, A., Mychalkiw, W. and Hutt, S. J. (1978) Prenatal progesterone I: The effect on development and on intellectual and academic achievement. Early Hum. Devel, 2, 305-322.

Lynch, A. and Mychalkiw, W. (1978) Prenatal progesterone II: Its role in the treatment of preeclamptic toxaemia and its effect on the offsprings’ intelligence—a reappraisal. Early Hum. Devel., 2, 323-339.

McArthur, R. (1967) Sex differences in field dependence for the Eskimo. Int. J. Psychol, 2, 139-140.

Maccoby, E. E. and Jacklin, C. N. (1974) The Psychology of Sex Differences. Stanford Univ. Press, Stanford, Calif.

Marshall, W. A. (1968) Development of the Brain. Oliver and Boyd, Edinburgh.

Maurer, R. A. (1973) Unpublished Ph.D. thesis, Univ. of California, Davis. Data reproduced in: Whalen, R. E. (1974) in Sex Differences in Behaviour (eds. R. C. Friedman, R. M. Richart and R. L. Vande Wiele), John Wiley, 467-481.

McEwen, B. S. (1976) Interactions between hormones and nerve tissue. Sci. Amer., 7,48-58.

McGlone, J. (1980) Sex differences in human brain organization. Behav. Brain Sci. In press.

McGlone, J. and Kertesz, A. (1973) Sex differences in cerebral processing of visuo-spatial tasks. Cortex, 9, 313-320.

McGuiness, D. (1972) Hearing: Individual differences in perceiving. Perception, 1, 465-473.

Messer, S. and Lewis, M. (1972) Social class and sex differences in the attachment and play behaviour of the year old infant. Merrill-Palmer Quart., 18, 295-306.


Meyer-Bahlburg, H. (1974) ‘Aggression, androgens and the XYY syndrome’, in Sex Differences in Behaviour (eds. R, C. Friedman, R. M. Richart and R. L. Vande Wiele), John Wiley, 433-453.

Money, J. (1964) Two cytogenetic syndromes: psychologic comparisons. I. Intelligence and specific factor quotients. J. Psychiatr. Res., 2, 223-231.

Money, J. (1977a) ‘The “givens” from a different point of view: lessons from intersexuality for a theory of gender identity’, in The Sexual and Gender Development of Young Children: The Role of the Educator (eds. E. K. Oremland and J. D. Oremland), Ballinger Publishing Co., Cambridge, Mass.

Money, J. (1977ft) ‘Prenatal deandrogenization of human beings’, in Handbook of Sexology (eds. J. Money and H. Musaph), Excerpta Medica, Amsterdam, 259-266.

Money, J. (1977c) ‘Determinants of human gender identity/role’, in Handbook of Sexology (eds. J. Money and H. Musaph), Excerpta Medica, Amsterdam, 57-79.

Money, J. and Lewis, V. (1966) Genetics and accelerated growth: adrenogenital syndrome. Bull. Johns Hopkins Hosp., 118, 365-373.

Money, J., Ehrhardt, A. A. and Masica, D. (1968) Fetal feminization induced by androgen insensitivity in the testicular feminizing syndrome: effect on marriage and maternalism. Johns Hopkins Med. J., 123, 105-114.

Money, J. and Ehrhardt, A. A. (1972) Man and Woman, Boy and Girl. Johns Hopkins Univ. Press, Baltimore.

Money, J. and Ogunro, C. (1974) Behavioural sexology: ten cases of genetic male intersexuality with impaired pre-natal and pubertal androgenization. Arch. Sex. Behav., 3, 181-205.

Money, J. and Schwartz, M. (1977) ‘Dating, romantic and non-romantic friendship and sexuality in 17 early treated andrenogenital females, aged 16-25’, in Congenital Adrenal Hyperplasia (eds. P. A. Lee, L. P. Plotnick, A. A. Kowarski and C. J. Migeon), Univ. Park Press, Baltimore, 419-451.

Moss, H. (1967) Sex, age and state as determinants of mother-infant interaction. Merrill-Palmer Quart., 13, 19-36.

O’Connor, J. (1943) Structural Visualization. Human Engineering Lab., Boston.

Perlman, S. M. (1971) Cognitive function in children with hormone abnormalities. Unpublished doctoral dissertation, North-western University, U.S.A.

Petersen, A. C. (1979) ‘Hormones and cognitive functioning in normal development’, in Sex Related Differences in Cognitive Functioning—Developmental Issues (eds. M. A. Wittig and A. C. Petersen), Academic Press, New York, 189-214.

Pfeiffer, C. A. (1936) Sexual differences of the hypophyses and their determination by the gonads. Amer. J. Anat., 58,195-226.

Phillips, A. G. and Deol, G. (1973) Neonatal gonadal hormone manipulation and emotionality following septal lesions in weanling rats. Brain Res., 60, 55-56.

Pion, R. J., Dignam, W. J., Lamb, E. J., Moore, J. G., Frankland, M. V. and Simmer, H. H. (1965) Testicular feminization. Amer. J. Obstet. Gynaecol, 93, 1067.

Quadagno, D. M., Briscoe, R. and Quadagno, J. S. (1977) Effect of perinatal gonadal hormones on selected nonsexual behaviour patterns: a critical assessment of the non-human and human literature. Psychol. Bull, 84, 62-80.

Quadagno, D. M., Shryne, J., Anderson, C. and Gorski, R. A. (1972) Influence of gonadal hormones on social sexual, emergence and open-field behaviour in the rat. Animal Behav., 20, 732-740.

Raisman, G. and Field, P. M. (1973) Sexual dimorphism in the neuropil of the preoptic area of the rat and its dependence on neonatal androgen. Brain Res., 54, 1-29.

Reinisch, J. M. (1977) Pre-natal exposure of human foetuses to synthetic progestin and oestrogen affects personality. Nature, 266, 561-562.

Reinisch, J. M. and Gandelman, R. (1978) ‘Human research in behavioural endocrinology: methodological and theoretical considerations’, in Hormones and Brain Development (ed. G. Dorner), Elsevier/N. Holland, 71-86.

Reinisch, J. M., Gandelman, R. and Spiegel, F. S. (1979) ‘Prenatal influences on cognitive abilities: data from experimental animals and human genetic and endocrine systems’, in Sex-Related Differences in Cognitive Functioning—Developmental Issues (eds. M. A. Wittig and A. C. Petersen), Academic Press, New York, 215-239.

Reinisch, J. M. and Karow, W. G. (1977) Pre-natal exposure to synthetic progestin and estrogens: effects on human development. Arch. Sex. Behav., 6, 257-288.



Reznikov, A. C. (1978) Hormonal regulation of sexual differentiation of the brain. Fiziol. Zh., 24, (1), 126- 133.

Rheingold, H. L. and Cook, K. V. (1975) The content of boys’ and girls’ rooms as an index of parents’ behaviour. Child Devel ., 46, (2), 459-463.

Savage, M. O., Preece, M. A., Jeffcoate, S. L., Ransley, P. G., Rumsby, G., Mansfield, M. D. and Williams, D. I. (1980) Familial male pseudohermaphroditism due to deficiency of 5a-reductase. Clinical Endocrinol ., 12, 397-406.

Scotf, J. S. (1978) ‘Intersex and sex chromosome abnormalities’, in Scientific Basis of Obstetrics and Gynaecology (2nd ed.) (ed. R. R. Macdonald), Churchill Livingstone, 301-344.

Shaffer, J. W. (1962) A cognitive deficit observed in gonadal aplasia (Turner’s syndrome). J. Clin. Psychol., 18, 403-406.

Short, R. V. (1979) ‘Sexual differentiation of the brain of the sheep: effects of prenatal implantation of androgen’. General discussion in Ciba Foundation Symposium, 62, Sex, Hormones and Behaviour, Elsevier/N. Holland, 257-269.

Siann, G. (1972) Measuring field-dependence in Zambia: a cross-cultural study. Int. J. Psychol, 7, 87-96.

Stafford, R. E. (1961) Sex differences in spatial visualization as evidence of sex-linked inheritance. Perceptual and Motor Skills, 13, 428.

Sternglanz, S. H. and Serbin, L. A. (1974) Sex-role stereotypes in children’s television programmes. Devel. Psychol, 10, 710-715.

Stewart, J., Skvarenina, A. and Pottier, J. (1975) Effects of neonatal androgens and maze-learning in the pre-pubescent and adult rat. Physiol. Behav., 14, 291-295.

Theilgaard, A. (1972) Cognitive style and gender role. Dan. Med. Bull, 19, 276-282

Tomlinson-Keasey, C. and Kelly, R. R. (1979) Is hemispheric specialization important to scholastic achievement? Cortex, 15, 97-107.

Vandenberg, S. G. and Kuse, A. R. (1979) ‘Spatial ability: a critical review of the sex-linked major gene hypothesis’, in Sex Related Differences in Cognitive Functioning (ed. M. A. Wittig and A. C. Petersen), Academic Press, New York, 67-95.

Waber, D. (1976) Sex differences in cognition: a function of maturation rate? Science, 192, 572-573.

Wada, J. A., Clark, R. and Hamm, A. (1975) Cerebral hemispheric asymmetry in humans. Arch. Neurol, 32, 239-246.

Waddington, C. H. (1957) The Strategy of Genes. Allen and Unwin, London.

Whalen, R. E. (1974) ‘Sexual differentiation: models, methods and mechanisms’, in Sex Differences in Behaviour (eds. R. C. Friedman, R. M. Richart and R. L. Vande Wiele), John Wiley, New York, London, 467-481.

Walker, P. A. and Money, J. (1972) Prenatal androgenization of females. Hormones, 3,119-128.

Weitzman, L. J., Eifler, D., Hokada, E. and Ross, C. (1972) Sex role socialization in picture books for pre-school children. Amer. J. Sociol, 77, 1125-1150.

Witelson, S. (1976) Sex and the single hemisphere: specialization of the right hemisphere for spatial processing. Science, 193, 425-426.

Witelson, S. (1977) ‘Early hemisphere specialization and inter-hemisphere plasticity: an empirical and theoretical review’, in Language Development and Neurological Theory (eds. S. Segalowitz and F. Gruber), Academic Press, New York, 213-287.

Witelson, S. and Pallie, W. (1973) Left hemisphere specialisation for language in the newborn: neuroanatomical evidence of asymmetry. Brain, 96, 641-646.

Yalom, I. D., Green, R. and Fisk, N. (1973) Prenatal exposure to female hormones. Arch. Gen. Psychiatr., 28, 554-561.

Zondek, L. H. and Zondek, T. (1974) ‘The influence of complications of pregnancy and of some congenital malformations on the reproductive organs of the male foetus and neonate’, in Sexual Endocrinology of the Perinatal Period (eds. M. G. Forest and G. Bertrand), Inserm, Lyon, France, 79-96.

Zussman, J. V., Zussman, P. P. and Dalton, K. (1975) Post-pubertal effects of prenatal administration of progesterone. Paper presented at a meeting of the Society for Research in Child Development, Denver, Co.