Relations Between Brain and Cognitive Development

Relations between Brain and Cognitive Development

Kurt W.Fischer University of Denver and Haruard Unioersity

FISCHER,KURTW. Relations between Brain and Cognitive Development. CHILDDEVELOPMENT, 1987, 58, 623-632. Goldman-Rakic reports important new data on cortical development in rhesus monkeys. Some of her findings, especially concurrent cortical synaptogenesis, may be related to cognitive capacities that develop in infancy. The developmental pattern of concurrent synaptogenesis in rhesus is consistent with a straightforward model of relations between brain and cognitive development: Concurrent synaptogenesis is hypothesized to lay the primary cortical foundation for a series of developmental levels in middle infancy that have been empirically documented in both human and rhesus infants. Other general brain changes, especially in the electroencephalogram, also seem to correlate with these levels, as well as with other levels that develop at other periods. In the simplest form of the model, these several factors all show synchronous developmental discontinuities at the time of emergence of a level. Specific research methods are available for specifying when discontinuities occur in development of both brain and behavior.

Obviously brain development relates to researchers had difficulty agreeing on even cognitive development. Developmental sci- the ages of emergence of major capacities entists generally make that assumption when (Biggs & Collis, 1982; Flavell, 1982; Gelman, they consider relations between brain and be- 1978). For brain development, researchers havior. Yet it has proved to be difficult to had produced relatively few data on specific establish specific developmental relations. neural changes that might relate to significant Global correlations exist, of course: Myelina- cognitive developments in human beings or tion occurs progressively throughout infancy, other primates. childhood, and even adolescence (Gibson, 1977; ~akbvlev& Lecours, 1967). ~bthden: As Goldman-Rakic (in this issue) sug- dritic branching and synaptic density change gests, however, the field of developmental systematically during infancy (Conel, 1939- neuroscience seems to have entered a new 1963; Goldman-Rakic, 1987, in this issue; era. At the same time, the field of cognitive Huttenlocher, 1979). During the same pe- development has also advanced, with major riods, children become progressively more in- new achievements in theory and method. As telligent by most criteria. Unfortunately, such a result of these advances in the two disci- correlations say little about real links between plines, there is genuine promise of the dis- brain and behavioral development because covery of meaningful links between specific they are too global. Any factors that show gen- brain and cognitive developments. Indeed, eral change with age will correlate with each two of the major sets of findings reported by other, as every statistics book demonstrates. Goldman-Rakic seem to have important im- To be meaningful, a relation must be speci- plications for links between brain and cogni- fied more precisely. For example, what partic- tive developments. ular changes in the brain relate to what One major set of findings is the provision specific new capacities emerging in children? of a relatively complete "wiring diagram" for In the past, the state of research in both a region of the prefrontal cortex in rhesus cognitive development and brain develop- monkeys, especially in terms of its relation to ment made it difficult to test hypothesized performance on a standard cognitive assess- specific relations. For cognitive development, ment tool, the delayed-response task (Gold-

Preparation of this article was supported by a fellowship from the Cattell Fund and grants from the Spencer Foundation and the Camegie Corporation. The statements made and views expressed are solely the responsibility of the authors. I would like to thank Robert G. Anderson, Eric Fischer, Sandra Pipp, Sam Rose, and Sheldon White for their contributions to the concepts and data pre- sented and Bruce Pennington and Linda Crnic for their thoughtful editorial suggestions. The author is now at Harvard University. Address reprint requests to Kurt W. Fischer, Larsen Hall, Appian Way, Harvard University, Cambridge, MA 02138. [Child Development, 1987, 58,623-632. 8 1987 by the Society for Research in Child Development, Inc. All rights resewed. 0009392@/87/58050022$01.00] 624 Child Development man-Rakic, 1986). To my knowledge, this is be a general spurt in brain growth, as mea- the first such description of part of the neural sured by global indexes other than synap- basis of an intelligent behavior. togenesis itself. The second major set of findings is the Four levels of sensorimotor develop- phenomenon of concurrent cortical synapto- ment.-With recent advances in cognitive- genesis (Rakic, Bourgeois, Eckenhoff, Zecevic, developmental method and theory, research- & Goldman-Rakic, 1986). Neural connections ers have found clear-cut evidence for at least between parts of the cerebral cortex have al- four periods of rapid behavioral change in ready begun to form almost 2 months before infant development, often called develop- birth in rhesus, yet at birth rhesus behavior is mental levels or transitions. Data from a num- still far from the adult form. What happens ber of different laboratories have led to sim- during infancy is the formation of synapses, ilar conclusions (Comgan, 1983; Fischer & the links by which neurons communicate. In- Silvern, 1985; McCall, 1983; Seibert, Hogan, deed, synapse formation shows a sharp spurt & Mundy, 1984; Uzgiris, 1976). Based on and subsequent drop, and this pattern occurs longitudinal assessments with standardized across all areas investigated in the cortex. tests, McCall, Eichorn, and Hogarty (1977) There is no obvious developmental sequence found these rapid changes at 2-4 months, 7-8 in which one area shows synaptogenesis be- months, 12-13 months, and 18-21 months in fore another. It is this pattern that Goldman- human infants. (There may be additional pe- Rakic and her colleagues call concurrent sy- riods of rapid change in the first 4 months; see naptogenesis. Fischer, 1980, and Fischer & Hogan, in press.) As Goldman-Rakic points out, concurrent synaptogenesis relates closely in time to the Each of these periods of rapid change ap- appearance of an important cognitive advance pears to mark the emergence of a specific new on the delayed-response task, which involves cognitive capacity, as summarized in Table 1. what Piaget (193711954) called object perma- The new capacity does not lead to instanta- nence. This task requires the coordination of neous change in all skills, since it must be a number of cortical functions, and conse- applied to each individual skill to reconstruct quently it makes sense that synaptogenesis it in a more complex, advanced form. In this throughout the cortex should correspond to an way, many skills (but not all) will advance advance in performance on the task. Indeed, a within a relatively brief age period (Fischer & number of other skills develop at approxi- Pipp, 1984). Concurrent synaptogenesis may mately the same age, and all require a similar provide the cortical basis for these level- coordination of a number of cortical functions. specific capacities. Considered together with recent research At 2-4 months, human infants become in cognitive development, these findings sug- able to reliably control variations in a single gest two classes of hypotheses about relations action in the service of a simple goal. For ex- between brain and cognitive developments. ample, a 4-month-old girl can keep her gaze First, the spurt and subsequent drop in synap- on a face as it moves around in front of her, or togenesis may relate closely to certain key she can voluntarily grasp a rattle that is lying cognitive capacities that emerge in middle in- next to her, pursuing it until she gets hold of fancy. I will suggest a specific model of these it. Before 2 months, infants show little evi- relations as well as methods for testing them. dence of the capacity to reliably adopt such Second, the development of new general cog- actions. nitive capacities seems to involve concurrent changes in numerous areas of the cortex. Ap- At 7-8 months, infants demonstrate the parently, the coordination of the functions of ability to consistently relate a few actions in a diverse regions is necessary for major cogni- single unit. An 8-month-old boy can skillfully tive advances. pull a string to make a mobile jiggle, and he can look at a toy and use what he sees to guide his hand to grasp the toy in many dif- A Relation between Concurrent ferent locations. The delayed-response task Synaptogenesis and Sensorimotor studied by Diamond and Goldman-Rakic Development (1983; Diamond, 1985) involves this kind of Concurrent synaptogenesis may lay the eye-hand coordination. Other types of rela- primary cortical foundation for important cog- tions of actions also develop at this age, such nitive capacities that emerge in infancy. Dur- as the coordination of vocalization and hear- ing the period of synaptogenesis, new cogni- ing in the repetition of simple sounds such as tive capacities appear, and there also seems to "mamarna" (Rarnsay, 1984) and the coordina- Kurt W. Fischer 625

TABLE 1 FOURDOCUMENTEDSENSORIMOTOR LEVELS IN HUMANINFANTS

Modal Age of Level Some Characteristic Behaviors Emergence Single sensorimotor actions ...... Single actions and perceptions; first social- 2-4 months emotional responsiveness Relations of a few sensorimotor actions.. . . . Coordination of means and end; skilled eye-hand 7-8 months coordination; attachment relationship with caretaker Systems of several sensorimotor actions.. . . . Complex means-ends variations in action; 11-13 months location of characteristics in objects and people; use of single words Representations...... Symbolization of people and objects; vocabulary 18-21 months spurt; multiword utterances No~~.-Thistable is based on Fischer and Silvem (1985), McCall, Eichom, and Hogarty (1977), and Uzgiris (1976). tion of two or more looking acts to produce and people (Chevalier-Skolnikoff, 1977; Dia- perception of a visual pattern (Bertenthal, mond & Goldman-Rakic, 1986; Parker & Gib- Campos, & Haith, 1980). son, 1979). The most complete behavioral By 12-13 months, infants show the ca- data involve stumptail macaques, a species pacity to relate a number of actions in a com- closely related to rhesus macaques (which plex system. A 13-month-old girl can facilely were studied in Goldman-Rakic's research). move an object through many positions, using In general, both types of monkeys show de- what she sees to guide what she does with the velopmental sequences similar to those of human infants, except that they produce object, and she can experiment with different as ways of holding an object until she finds one specifies-specific behaviors such climbing that accomplishes some desired goal, such as and lip smacking and they show less explora- dropping a ball so that it falls into a small hole tion of the properties of objects. in a toy box. In the delayed-response task, The macaques show only the first three children become capable of finding the hid- of the four sensorimotor levels as evidenced den object in the face of multiple hidings, so by the developmental sequences that have long as they can actually see the object being been documented. I know of no test of covered the last time it is hidden. whether they demonstrate periods of rapid In the middle of the second year, at 18- change with the emergence of each level. 21 months, toddlers develop the ability of The first level, single actions, appears at 1 representation or symbolization, as originally month of age. The second level, simple rela- described by Piaget (194611951). Moving be- tions of actions, develops at 2 months. This is yond action systems, they can cognitively the same age that Goldman-Rakic (1987, in evoke an object, event, or person that is not this issue) finds emergence of this level in the actually present, pretending that a doll is delayed-response task in rhesus, as well as walking across the table or talking about what the age that she specifies for the start of the their uncle did a few days before (Watson & spurt in synaptogenesis. Also, she lists a long Fischer, 1977). At the same time they be- series of other behaviors that develop at this gin to speak sentences and show a large spurt age. The third level, systems of actions, ap in vocabulary (Bates, Benigni, Bretherton, pears at 4-4s months, which is the age at Camaioni, & Volterra, 1979; Bloom, 1973; which 'rhesus show the most advanced per- Conigan, 1983). formance in the delayed-response task as well as the end of the synaptogenetic spurt. Although the behavioral evidence for monkeys is less complete than that for human Synaptogenesis and relations of ac- infants, assessments of sensorimotor develop tions.-Goldman-Rakic hypothesizes that the ment similar to those used by Piaget (19371 advance in search in the delayed-response 1954) and Uzgiris and Hunt (1975) indicate task develops as a result of concurrent synap similar developmental sequences in monkeys togenesis. The formation of new synapses and 626 Child Development the subsequent deletion of extra synapses screen where they found the object on the bring the capacity to recall information about previous trial instead of under the screen where an object was hidden before, which where they just saw the object disappear. produces the change in search. She calls this This behavior in search was originally re- capacity representation, because the monkey ported by Piaget (1937/1954), and hundreds of can recall information that is not directly studies by other researchers have docu- available in the stimulus array. This sense of mented and extended his findings (Uzgiris & re~resentationis different fiom the remesen- Hunt, 1975). tion or symbolization that characterizes the fourth level of sensorimotor development. In the relation of actions for the delayed- response task, infants coordinate looking at In the cognitive-developmental litera- the disappearing object, looking at the screen, ture, the term "representation" is used in two and grasping the object under the screen; and distinct ways. Researchers with an informa- in doing so, they coordinate where they found tion-processing orientation typically use it as the object on the previous trial with where Goldman-Rakic does to refer to recall memory they search on the present trial. It is precisely in general, including the ability to remember this coordination that leads to searching where an object was hidden in the delayed- under the incorrect screen (the one correct on response task. Those in the Piagetian tradi- the previous trial). tion, on the other hand, use it to refer to a more complex recall capacity-the ability to At the end of the period of concurrent evoke an object, person, or event that is not synaptogenesis in rhesus, the third sen- present, as evidenced for example by pre- sorimotor level is emerging. Infants come to tending that a block is a dog walking across be able to control systems of actions, involv- the table. The information-processing meaning complex relations among a number of ac- ing requires that the organism bring to bear tions, and such complex relations produce a some information that is not present in the great increase in the complexity of search be- stimulus display at the moment. The Piage- havior, including coping with long delays in tian meaning, on the other hand, requires the the delayed-response task (McCall et al., evocation (interiorized recreation or imita- 1977; Uzgiris, 1976). tion) of some entity that is absent, such as the Another way of stating this relation is dog walking across the table, the uncle who that discontinuities in the synaptogenetic made a funny face yesterday, or the favorite spurt correlate with the second and third toy buck (Fischer & Jennings, 1981; Piaget, levels. The start of the peak in synaptogenesis 194611951). Rhesus never develop this repre- correlates with the second level, and the end sentational capacity of the fourth level, but at of the peak correlates with the third level. I the second and third sensorimotor levels they will argue that such discontinuities in concur- do develop representation in the sense of re- rent synaptogenesis and in other aspects of call (Fischer, 1982; Parker & Gibson, 1979). brain growth relate to each of the levels of A straightforward extension of Goldman- sensorimotor development. Recall that each Rakic's hypothesis is that concurrent syn- level is also indexed by discontinuities, spurts aptogenesis relates to the emergence and in behavioral change. Thus, it is necessary to consolidation of the second and third sensori- discuss methodological issues about testing motor levels. The delayed-response task is an for discontinuities before outlining the brain- index of the second level in infancy, relations growth model and the research it suggests. of actions. At this level, infants develop the capacity to coordinate two or more sen- Methods of Assessing for Levels sorimotor actions in a single, integrated skill Discontinuities, such as spurts and drops or scheme, including the ability to use one in developmental curves, seem to be one of action to recall a second one (Case. 1985: Fi- the most straightforward and easily assessed scher, 1980; McCall et d.,' 1977; ~zgiris, indexes of developmental levels. A discon- 1976). tinuity is an abrupt change in the slope of a According to this neo-Piagetian analysis, developmental curve. This meaning of dis- here is what happens in the delayed-response continuity refers solely to the form of the de- task. With the capacity to relate actions, in- velopmental function. It has no direct impli- fants begin to search successfully for objects cations for other issues often labeled with the placed under a screen. In carrying out that same term, such as the degree of stability search, however, they demonstrate what is (continuity) or instability (discontinuity) in called the AB error: Under certain condi- individual differences during development tions, they systematically reach under the (Bornstein & Sigrnan, 1986). Kurt W. Fischer 627 Several methods have been devised for good choice for the age period under study, investigating discontinuities and relating but its limits should be recognized. Different them across different measures. The ensuing tasks generate different developmental pat- discussion of these methods is based on the terns, even when they assess what appears to following sources, where the methods are de- be a single content domain such as search scribed in more detail: Fischer and Canfield (Jackson, Campos, & Fischer, 1978). This task (1986), Fischer, Pipp, and Bullock (1984),and effect is one of the best documented phenom- McCall (1983). Wohlwill's (1973) classic dis- ena in cognitive development (Biggs & Col- cussion of developmental methods is also rel- lis, 1982; Flavell, 1982; Uzgiris & Hunt, evant. 1975). Some research on errors suggests that a single task may only index behaviors that are Guidelines for research.-Testing for developmental discontinuities within and across close to the level of complexity of that task. domains is straightforward, provided that cer- For example, on a task that is far beyond chil- tain simple guidelines are followed. Four key dren's developmental level, they often do not methodological issues are age sampling, the demonstrate abilities that they show easily on simpler tasks (Fischer & Roberts, 1986; sensitivity of the developmental scale, the range of domains assessed, and the influence Roberts, 1981). Multi-task scales for search of testing conditions. The research from that eliminate these problems are available Goldman-Rakic's laboratory has broken new and have been used extensively in developmental research (Comgan, 1983; Uzgiris & ground, and so of course it could not be de- Hunt, 1975). signed to anticipate all these issues. Future research, however, will more satisfactorily an- Third, just as assessment of cortical func- swer important developmental questions if tioning requires testing diverse areas of the it deals with these issues. Most obviously cortex, assessment of cognitive functioning re- needed are more precise assessments of the quires testing diverse domains of behavior. developmental course of synaptogenesis and Performance in one task domain is not repre- its relation to the behavioral changes charac- sentative of cognitive development in gen- teristic of the sensorimotor levels. eral. Although each developmental level does First, detection of a spurt or drop re- seem to produce rapid changes in perform- quires a good clock, one that can detect the ance on many tasks in many domains, there is rapidity of the change. Ages must be sampled still great variability across domains. frequently enough to provide multiple assess- Fourth, a cognitive-developmental as- ments before, during, and after the period of sessment should include a range of testing the hypothesized spurt or drop. For example, conditions. Factors such as practice, familiar- if three developmental levels are hypothesized to emerge in the first year in human ity, and memory requirements produce wide babies, then assessments must be made at variations in the forms of developmental least once a month to reliably detect the pe- curves even when the same tasks are used. riod of the discontinuities. Such frequent age Only a range of testing conditions, including sampling is needed for both synaptogenesis for example both low- and high-practice con- and behavioral development. ditions, can provide an accurate general por- trait of cognitive development. Second, detection of discontinuities requires not only a good clock but also a good The developmental pattern of synap- ruler, a scale that is sensitive to develop- togenesis.-For research on synaptogenesis, mental change. The measure of synaptic den- two key issues are the sampling of ages and sity appears to provide such a ruler, but the the assessment of diversity across cortical re- delayed-response task probably does not. The gions. The data reported by Goldman-Rakic best cognitive assessments typically form a and her colleagues (Rakic et al., 1986)provide relatively continuous developmental scale of frequent-enough samples to give an initial cognitive complexity. One of the best ways of rough estimate of the developmental pattern devising such a scale is to use a number of synaptogenesis. of tasks within the same domain that vary in cognitive complexity (Comgan, 1983; Uzgiris One of the most interesting charcteristics & Hunt, 1975). of cortical synaptogenesis is its concurrence across cortical areas, as Goldman-Rakic (1987, A single task will seldom provide a good in this issue) points out: It shows approxi- ruler, primarily because it produces such a mately the same developmental function limited sample of behavior. If a single task is across all assessed areas of the cortex, even to be used, the delayed-response task is a though it would have been reasonable to ex- 628 Child Development pect large differences for different cortical A more modest extension of synaptogenesis areas. presumably occurs for the great apes as well, since they also seem to reach the level of There do seem to be important variations single representations, although in a limited in the synaptogenetic curve, however-not form (Chevalier-Skolnikoff, 1977; Parker & large differences across areas, but discon- Gibson, 1979). tinuities that, by hypothesis, relate to the cognitive-developmental levels. According to Brain-wave patterns are a good candidate Figure 6 in Goldman-Rakic's article, synaptic for testing the brain-growth hypothesis. They density first rose above the normal adult level show strong developmental change that can in rhesus between birth and 1 month of age. be reliably measured, and they index a form In some areas, it then leveled off for several of brain activity that seems to reflect changes months, while in others it continued to rise to in brain functioning. The electroencepha- a peak at approximately 2 months of age be- logram, evoked potentials, and brain metabo- fore leveling off. Next, in some areas, it rose to lism all show major changes at the ages for a new peak at 4-4% months of age. Finally, the developmental levels (Chugani & Phelps, after approximately 4% months of age, it be- 1986; Dreyfis-Brisac, 1978; Emde, Gaens- gan to gradually decrease until after many bauer, & Harmon, 1976; Kagan, 1982; Wood- months it reached the adult level. ruff, 1978). For example, in the electroencephalogram, new peak frequencies develop The available data thus indicate a peak at at approximately 4 and 8 months of age, espe- 2 months, a peak at 4-4% months, and proba- cially in the occipital-parietal area (Hagne, bly an earlier leveling off at approximately 1 month of age. Of course, limits on ages Persson, Magnusson, & Petersen, 1973). sampled as well as sample size preclude firm In addition, we have data suggesting gen- conclusions about the exact forms of the de- eral spurts in brain size during these age pe- velopmental curves and the consistency of riods. Subjects were a cross-sectional sample these apparent discontinuities across cortical of over a thousand infants measured at ages regions. Nevertheless, these three discon- distributed throughout the first year of life tinuities are consistent with the proposed by Bonnie Camp at the Colorado Health model. There seem to be discontinuities in Sciences Center. For example, the infants the synaptogenetic curves at the approximate showed a strong, statistically reliable spurt in ages for emergence of the three sensorimotor head growth between 7 and 8 months, the age levels in rhesus. of emergence of the second sensorimotor level (Fischer, Camp, Pipp, & Rose, 1986). The Brain-Growth Model That is, mean head circumference increased Several scholars have hypothesized that 2.7% (11.8 mm) between 7 and 8 months, major brain changes occur with each cog- while it increased only 1.5% (6.5 mm) be- nitive-developmental level (Epstein, 1974; tween 6 and 7 months and 0.1% (0.6 mm) be- Fischer & Pipp, 1984; Gibson, 1977; White, tween 8 and 9 months. This spurt in head 1970).The apparent connection between con- growth was the largest for any month after 4 current synaptogenesis and the three sen- months of age. (The head grows fast early in sorimotor levels in rhesus may be one such the first year and then slows down.) Note that relation. If that relation holds for human 7-8 months is the same age that the second infants as well as rhesus, then the synap- sensorimotor level emerges. togenetic spurt in human infants will show Between 3 and 4 months, when the first discontinuities at approximately 4, 8, and 13 level emerges, there was a spurt of 3.1% (12.4 months, with emergence of the three sen- mm). Analysis of week-by-week changes sug- sorimotor levels. In addition, human infants gests that this spurt is focused at 15-17 will show a fourth discontinuity at approxi- weeks. Data are not yet available to deter- mately 20 months. Available data on human mine whether a similar head-growth spurt oc- synaptogenesis are limited, but they do sug- curs with the emergence of the third level at gest that concurrent synaptogenesis peaks at 12-13 months or the fourth level at 18-21 8-12 months in the visual cortex and then months. gradually declines-a pattern similar to that predicted fiom the rhesus data (Huttenlocher, Of course, caution must be exercised de Courten, Garey, & van der Loos, 1982). In in interpreting the head-circumference data. the frontal cortex, on the other hand, it ap- First, head circumference is not a direct mea- pears to extend for a longer period-peaking sure of brain size, although there is a strong at 1-2 years and not decreasing substantially correlation between the two in human beings until after 7 years of age (Huttenlocher, 1979). (Winick & Rosso, 1969). Second, although in- Kurt W. Fischer 629 creases in brain size could be produced by Brain electrical activity changes system- concurrent synaptogenesis, they could also be atically with these later cognitive-develop- produced by many other brain events, such as mental levels too (Dustman & Beck, 1969; the growth of myelin in the cortex during the Epstein, 1980; Fischer & Pipp, 1984). These same period (Yakovlev & Lecours, 1967). findings seem to imply that there are basic Third, the spurts can be seen reliably only in changes in brain functioning, analogous to a large sample of children, not in individual concurrent synaptogenesis, that occur for growth curves. Camp has longitudinal data for each developmental level. What might these over 30 infants, and the individual curves changes be for the later levels? show wide variations, presumably arising In terms of the brain-growth model, the from factors such as illness, nutrition, hair most likely neural foundation for the later growth, and measurement error. Lamp1 and levels is formation of a group of new synapses Emde (1983) report similar individual vari- in a limited cortical region, even though over- ations in growth curves. Fourth, a note of all synaptic density is not increasing. caution is needed. A correlation between in- Al- though brain scientists used to believe that no crease in head size and onset of developmen- new synapses formed after infancy, recent tal level does not mean that head size in gen- research indicates that new synapses can eral correlates with intelligence as measured form later in development, as reported by by intelligence tests. The two correlations are Greenough, Black, and Wallace (1987, in this independent, because increase in head size at issue). Another foundation consistent with a specific age interval need not relate to total the model is a spurt in the pruning of head size. synapses rather than the formation of new ones. Finally, of course, the later levels may According to the brain-growth hy- be of a different sort from the earlier ones. pothesis, spurts in brain change will occur depending not on synaptic change but on with the emergence of each developmental other brain processes such as myelin forma- level. Research to date indicates that there are tion, which continues into early adulthood at least four sensorimotor levels in human in- (Yakovlev & Lecours, 1967). fancy and at least three in rhesus. Additional levels between those already identified seem unlikely, since careful investigations have un- Conclusion: The Prefrontal Cortex covered no evidence for additional discon- in the Whole Brain tinuities (e.g., Comgan, 1983; McCall et al., Goldman-Rakic's findings clearly provide 1977). However, before the first level (single fertile ground for relating brain development actions) additional levels have been hy- with cognitive development. Indeed, the hypothesized (Fischer, 1980; Fischer & Hogan, pothesized relations between concurrent in press). The existence of such early le- synaptogenesis and infant cognitive devel- vels would have important implications for opment lead straightforwardly to a specific, research on synaptogenesis or other brain extensive research program for mapping out changes, because it would mean that studies possible neural bases of cognitive develop- must be designed to detect a rapid-fire series ment. of discontinuities. The greater the number of developmental levels, the more frequently In the simple model I have suggested, ages must be sampled to detect those levels each developmental level is marked by not and the more discriminations developmental only behavioral discontinuities but also dis- scales must make. continuities in synaptogenesis and devel- oping brain electrical activity. Of course, the pattern may not prove to be so simple, but After infancy, human beings develop in- whatever the relations between cognitive de- tellectual capacities that are never seen in velopments and brain-growth patterns, the re- other primates. Cognitive-developmental the- sults will shed light on the neural foundations orists predict between three and ten addi- of cognitive change. tional levels after those of infancy (Biggs & Collis, 1982; Case, 1985; Fischer, 1980; Hal- Notice, however, that I have said little ford, 1982). Empirical evidence supports at about an important part of Goldman-Rakic's least four periods of rapid behavioral findings. Besides uncovering concurrent syn- change-at approximately 4 years, 6-7 years, aptogenesis, she and her colleagues have 10-12 years, and 14-16 years (Fischer & Sil- also described a basic part of the wiring dia- vern, 1985). As always, the exact ages vary gram for the delayed-response task in the pre- with the particular tasks and assessment crite- frontal cortex. They have even found a type of ria. neuron that specifically increases its firing 630 Child Development

rate during the delay period in that task. From (New Directions for Child Development No. these findings, they originally expected to dis- 21, pp. 51-64). San Francisco: Jossey-Bass, cover a synaptogenetic change specific to the 1983. prefrontal area at the age when rhesus are be- Diamond, A. (1985). Development of the ability to ginning to solve the delayed-response task. use recall to guide action, as indicated by in- Instead, they found concurrent synapto- fants' performance on AB. Child Development, genesis-spurts followed by drops in all areas 56, 868-883. of the cortex during the same age period. Diamond.. A,.. & Goldman-Rakic. P. (1983). Com- parison of performance on a Piagetian object This global pattern of cortical develop- permanence task in human infants and rhesus ment appears to make sense from a cognitive- monkeys: Evidence for involvement of pre- developmental perspective. The cognitive frontal cortex. Society for Neuroscience Ab- changes that occur with each developmental level are general ones, affecting a wide range stracts, 5, 85. Diamond, A., & Goldman-Rakic, P. (1986). Com- of behaviors. Other brain evidence also sup- ~arativedevelo~ment in human infants and ports the global pattern: There are discon- rhesus monkeys of cognitive functions that de- tinuities in the developmental functions for pend on prefrontal cortex. Society for Neuro- the electroencephalogram and even for brain science Abstracts, Vol. 6. (head) size in infancy. The prefrontal cortex Dreyfus-Brisac, C. (1978). Ontogenesis of brain is central to performance on the delayed- bioelectrical activity and sleep organization in response task. But as I believe Goldman- neonates and infants. In F. Falkner & J. M. Rakic would agree, it works with other corti- Tanner (Eds.),Human growth 3: Neurobiology cal areas to produce changes in general cognitive capacity. and nutrition. New York: Plenum. Dustman, R. E., & Beck, E. C. (1969).The effects of maturation and aging on the waveform of visu- ally evoked potentials. Electroencephalogram References and Clinical Neurophysiology, 265,2-11. Bates, E., Benigni, L., Bretherton, I., Camaioni, L., Emde, R., Gaensbauer, T., & Harmon, R. (1976). & Volterra, V. (1979). The emergence of sym- Emotional expression in infancy: A biobehav- bols. New York: Academic Press. ioral study. Psychological Issues (Vol. 10, Bertenthal, B. I., Campos, J. J., & Haith, M. M. No. 37). New York: International Universities (1980). Development of visual organization: Press. The perception of subjective contours. Child Epstein, H. T. (1974). Phrenoblysis: Special brain Deuelopment, 51, 1072-1080. and mind growth periods. Deuelopmental Psy- Biggs, J., & Collis, K. (1982).A system for evaluat- chobiology, 7,207-224. ing learning outcomes: The SOW taxonomy. Epstein, H. T. (1980). EEG developmental stages. New York: Academic Press. Developmental Psychobiology, 13, 629-631. Bloom, L. (1973).One word at a time. The Hague: Fischer, K. W. (1980).A theory of cognitive devel- Mouton. opment: The control and conshction of hier- Bornstein, M. H., & Sigman, M. D. (1986). Con- archies of skills. Psychological Reuiew, 87, tinuity in mental development from infancy. 477-531. Child Development, 57, 251-274. Fischer, K. W. (1982). Human cognitive develop- Case, R. (1985).Intellectual deuelopment: Birth to ment in the first four years. Behauioral and adulthood. New York: Academic Press. Brain Sciences, 5,282-283. Chevalier-Skolnikoff, S. (1977). A Piagetian model Fischer, K. W., Camp, B., Pipp, S. L., & Rose, S. for describing and comparing socialization in (1986).Spurts in head circumference in human monkey, ape, and human infants. In S. infants. Unpublished manuscript. Chevalier-Skolnikoff & F. E. Poirier (Eds.), Fischer, K. W., & Canfield, R. L. (1986).The ambi- Primate bio-social deuelopment: Biological, so- guity of stage and structure in behavior: Person cial, and ecological determinants (pp. 159- and environment in the development of psy- 188). New York: Garland. chological structures. In I. Levin (Ed.), Stage Chugani, H. T., & Phelps, M. E. (1986). Matura- and structure: Reopening the debate (pp. 246- tional changes in cerebral function in infants 267). New York: Plenum. determined by "FDG Positron Emission To- Fischer, K. W., & Hogan, A. (in press). Sources of mography. Science, 231, 840-843. variation in the development of actions in in- Conel, J. L. (1939-1963). The postnatal deuelop- fancy. In J. Lockman & N. Hazen (Eds.), Ac- ment of the human cerebral cortex (7 vols.). tion in social context: Perspectives on early de- Cambridge, MA: Harvard University Press. uelopment. New York: Plenum. Conigan, R. (1983). The development of repre- Fischer, K. W., & Jennings, S. (1981). The emer- sentational skills. In K. W. Fischer (Ed.), Le- gence of representation in search. Dwelop vels and transitions in children's development mental Reuiew, 1, 18-30. Kurt W. Fischer 631

Fischer, K. W., & Pipp, S. L. (1984). Processes of Kagan, J. (1982). Psychological research on the hu- cognitive development: Optimal level and skill man infant: An evaluative summary. New acquisition. In R. J. Sternberg (Ed.), Mecha- York: W.T. Grant Foundation. nisms of cognitive deuelopment (pp. 45-80). Larnpl, M., & Emde, R. N. (1983). Episodic growth San Francisco: W. H. Freeman. in infancy: A preliminary report on length, Fischer, K. W., Pipp, S. L., & Bullock, D. (1984). head circumference, and behavior. In K. W. Detecting discontinuities in development: Fischer (Ed.), Levels and transitions in chil- Method and measurement. In R. Emde & R. dren's deuelopment (New Directions for Child Harmon (Eds.), Continuities and discontinui- Development, No. 21, pp. 21-36). San Fran- ties in development (pp. 95-121). New York: cisco: Jossey-Bass. Plenum. McCall, R. B. (1983). Exploring developmental Fischer, K. W., & Roberts, R. J., Jr. (1986). The de- transitions in mental performance. In K. W. velopment of classijication skills in the pre- Fischer (Ed.), Levels and transitions in chil- school years: Developmental level and errors. dren's deuelopment (New Directions for Child Unpublished manuscript. Development, No. 21, pp. 65-80). San Fran- Fischer, K. W., & Silvern, L. (1985). Stages and in- cisco: Jossey-Bass. dividual differences in cognitive development. McCall, R. B., Eichorn, D. H., & Hogarty, P. S. Annual Review of Psychology, 36, 613-648. (1977). Transitions in early mental develop- Flavell, J. (1982). On cognitive development. Child ment. Monographs of the Society for Research Development, 53, 1-10. in Child Development, 42(3, Serial No. 171). Gelman, R. (1978). Cognitive development. Annual Parker, S. T., & Gibson, K. R. (1979). A develop- Review of Psychology, 29, 297-332. mental model of the evolution of language and Gibson, K. R. (1977). Brain shcture and intelli- intelligence in early hominids. Behavioral and gence in macaques and human infants from Brain Sciences, 2, 367-381. a Piagetian perspective. In S. Chevalier- Piaget, J. (1951). Play, dreams, and imitation in

Skolnikoff & F. E. Poirier (Eds.).~ ,, Primate bio- childhood (C. Gattegno & F. M. Hodgson, social deuelopment: Biological, social, and bans.). New York: Harcourt Brace. (Original ecological determinants (pp. 113-158). New work published 1946) York: Garland. Piaget, J. (1954). The construction of reality in the Goldman-Rakic, P. (1986). Circuitry of the prefron- child (M. Cook, trans.). New York: Basic. (Orig- tal cortex and the regulation of behavior by rep- inal work published 1937) resentational knowledge. In F. Plum & V. Rakic, P., Bourgeois, 3.-P., Eckenhoff, M. F., Mountcastle (Eds.), Handbook of physiol- Zecevic, N., & Goldman-Rakic, P. (1986). Con- ogy. Bethesda, MD: American Physiological current overproduction of synapses in diverse Society. regions of the primate cerebral cortex. Science, Goldman-Rakic, P. S. (1987). Development of corti- 232,232-235. cal circuitry and cognitive function. Child De- Ramsay, D. S. (1984). Onset of duplicated syllable velopment, 58,601-622. babbling and unimanual handedness in in- Greenough, W. T., Black, J. E., & Wallace, C. S. fancy: Evidence for developmental change in (1987). Experience and brain development. hemispheric specialization? Developmental Child Development, 58,539-559. Psychology, 20,64-71. Hagne, I., Persson, J., Magnusson, R., & Petersen, Roberts, R. J., Jr. (1981). Errors and the assessment I. (1973). Specbal analysis via fast Fourier of cognitive development. In K. W. Fischer bansform of waking EEG in normal infants. In (Ed.),Cognitive development (New Directions P. Kellaway & I. Petersen (Eds.), Automation for Child Development, No. 12, pp. 69-78). of clinical electroencephalography (pp. 103- San Francisco: Jossey-Bass. 143). New York: Raven. Seibert, J. M., Hogan, A. E., & Mundy, P. C. (1984). Halford, G. S. (1982). The development of thought. Mental age and cognitive stage in young handi- Hillsdale, NJ: Erlbaum. capped and at-risk children. Intelligence, 8, Huttenlocher, P. (1979). Synaptic density in human 11-29. frontal cortex-developmental changes and ef- Uzgiris, I. C. (1976).Organization of sensorimotor fects of aging. Brain Research, 163, 195-205. intelligence. In M. Lewis (Ed.), Origins of in- Huttenlocher, P., de Courten, C., Garey, L. J., & telligence: Infancy and early childhood (pp. van der Loos, H. (1982). Synaptogenesis in hu- 123-164). New York: Plenum. man visual cortex-evidence for synapse elim- Uzgiris, I. C., & Hunt, J. McV. (1975). Assessment ination during normal development. Neuro- in infancy: Ordinal scales of psychological science Letters, 33, 247-252. development. Urbana: University of Illinois Jackson, E., Campos, J. J., & Fischer, K. W. (1978). Press. The question of decalage between object per- Watson, M. W., & Fischer, K. W. (1977). A develop manence and person permanence. Deuelop- mental sequence of agent use in late infancy. mental Psychology, 14, 1-10. Child Development, 48,828-835. 632 Child Development

White, S. H. (1970). Some general outlines of the Woodruff, D. S. (1978). Brain electrical activity and matrix of developmental changes between five behavior relationships over the lifespan. In and seven years. Bulletin of the Orton Society, P. B. Baltes (Ed.), Life-span deuelopment and 20,41-57. behuior (Vol. 1, pp. 111-179). New York: Aca- Winick, M., & Rosso, P. (1969). Head circumfer- demic Press. ence and cellular growth of the brain in normal Yakovlev, P. I., & Lecours, A. R. (1967). The my- and marasmic children. Journal of Pediatrics, elogenetic cycles of regional maturation of the 74,774-778. brain. In A. Minkowsky (Ed.), Regional deuel- Wohlwill, J. F. (1973). The study of behuioral de- opment of the brain in early life (pp. 3-70). uelopment. New York: Academic Press. Oxford: Blackwell. http://www.jstor.org

LINKED CITATIONS - Page 1 of 3 -

You have printed the following article: Relations between Brain and Cognitive Development Kurt W. Fischer Child Development, Vol. 58, No. 3. (Jun., 1987), pp. 623-632. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198706%2958%3A3%3C623%3ARBBACD%3E2.0.CO%3B2-6

This article references the following linked citations. If you are trying to access articles from an off-campus location, you may be required to first logon via your library web site to access JSTOR. Please visit your library's website or contact a librarian to learn about options for remote access to JSTOR.

References

Development of Visual Organization: The Perception of Subjective Contours Bennett I. Bertenthal; Joseph J. Campos; Marshall M. Haith Child Development, Vol. 51, No. 4. (Dec., 1980), pp. 1072-1080. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198012%2951%3A4%3C1072%3ADOVOTP%3E2.0.CO%3B2-E

Continuity in Mental Development from Infancy Marc H. Bornstein; Marian D. Sigman Child Development, Vol. 57, No. 2. (Apr., 1986), pp. 251-274. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198604%2957%3A2%3C251%3ACIMDFI%3E2.0.CO%3B2-7

Maturational Changes in Cerebral Function in Infants Determined by 18 FDG Positron Emission Tomography Harry T. Chugani; Michael E. Phelps Science, New Series, Vol. 231, No. 4740. (Feb. 21, 1986), pp. 840-843. Stable URL: http://links.jstor.org/sici?sici=0036-8075%2819860221%293%3A231%3A4740%3C840%3AMCICFI%3E2.0.CO%3B2-9 http://www.jstor.org

LINKED CITATIONS - Page 2 of 3 -

Development of the Ability to Use Recall to Guide Action, as Indicated by Infants' Performance on AB# Adele Diamond Child Development, Vol. 56, No. 4. (Aug., 1985), pp. 868-883. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198508%2956%3A4%3C868%3ADOTATU%3E2.0.CO%3B2-P

On Cognitive Development John H. Flavell Child Development, Vol. 53, No. 1. (Feb., 1982), pp. 1-10. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198202%2953%3A1%3C1%3AOCD%3E2.0.CO%3B2-N

Development of Cortical Circuitry and Cognitive Function Patricia S. Goldman-Rakic Child Development, Vol. 58, No. 3. (Jun., 1987), pp. 601-622. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198706%2958%3A3%3C601%3ADOCCAC%3E2.0.CO%3B2-O

Experience and Brain Development William T. Greenough; James E. Black; Christopher S. Wallace Child Development, Vol. 58, No. 3. (Jun., 1987), pp. 539-559. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28198706%2958%3A3%3C539%3AEABD%3E2.0.CO%3B2-J

Transitions in Early Mental Development Robert B. McCall; Dorothy H. Eichorn; Pamela S. Hogarty; Ina C. Uzgiris; Earl S. Schaefer Monographs of the Society for Research in Child Development, Vol. 42, No. 3, Transitions in Early Mental Development. (1977), pp. 1-108. Stable URL: http://links.jstor.org/sici?sici=0037-976X%281977%2942%3A3%3C1%3ATIEMD%3E2.0.CO%3B2-B

Concurrent Overproduction of Synapses in Diverse Regions of the Primate Cerebral Cortex Pasko Rakic; Jean-Pierre Bourgeois; Maryellen F. Eckenhoff; Nada Zecevic; Patricia S. Goldman-Rakic Science, New Series, Vol. 232, No. 4747. (Apr. 11, 1986), pp. 232-235. Stable URL: http://links.jstor.org/sici?sici=0036-8075%2819860411%293%3A232%3A4747%3C232%3ACOOSID%3E2.0.CO%3B2-7 http://www.jstor.org

LINKED CITATIONS - Page 3 of 3 -

A Developmental Sequence of Agent Use in Late Infancy Malcolm W. Watson; Kurt W. Fischer Child Development, Vol. 48, No. 3. (Sep., 1977), pp. 828-836. Stable URL: http://links.jstor.org/sici?sici=0009-3920%28197709%2948%3A3%3C828%3AADSOAU%3E2.0.CO%3B2-8