The Cognitive Neuroscience of Development

Michelle de Haan, Mark H. Johnson

• Publisher: Psychology Press
• Year: 2005
• ISBN:
  9781135431280
• Edition: 1
• Topic: Psicologia
• Subtopic: Storia e teoria della psicologia

How are the experiences of childhood incorporated into the structures of the developing brain, and how do these changes in the brain influence behaviour? This is one of the many questions motivating research in the relatively new field of developmental cognitive neuroscience. This book provides an extensive overview of the methods used to study such questions, and a thorough investigation into the emerging interface between neurobiological and psychological perspectives in the study of typical and atypical cognitive behaviour.
The Cognitive Neuroscience of Development is a collection of essays written by international experts in the field. It covers not only traditional topics such as language, attention and memory development, but also includes individual chapters covering the theories of neurocognitive development and methods of studying brain activity in young infants and children. There are additional chapters on hormonal influences on brain and behavioural development, gender differences in the brain, and genetic disorders.
This exceptional series of contributions surveys the study of both cognitive and neural development. The book takes into account brain architecture as well as the behavioural context of development, thus it succeeds in integrating the multiple methods and domains of research that have previously been studied in a more fragmented way. It will be invaluable to upper level students as well as researchers and teachers in Psychology, Neuroscience, Cognitive Science, Paediatrics and related fields.

CHAPTER ONE

Mechanisms and theories of brain development

Michelle de Haan
Institute of Child Health, University College London, London, UK

Mark H. Johnson
Centre for Brain & Cognitive Development, Birkbeck College, London, UK

INTRODUCTION

Understanding the development of the cerebral cortex is central to our understanding of psychological development, as immaturity of this structure is a major limiting factor of cognitive functioning in infants and children. Genetic factors inherent to the developing cells themselves, and interactions with external factors—at the molecular, cellular, and behavioural levels—contribute to the developmental process. Even when the baby is still in the protective environment of the womb, development does not consist simply of the unfolding of a rigid “genetic plan”. For example, the detailed folding patterns of the cerebral cortex present at birth can vary considerably, even between identical twins (Bartley et al., 1997). The comparatively long phase of postnatal brain development in humans compared with other primates or mammals provides an extended opportunity for later stages of brain development to be influenced by the external environment. The aim of this chapter is to outline the major events of prenatal and postnatal brain development, review current theories of how the cerebral cortex develops into its adult organization and function, and discuss how cortical injury affects development.

BASICS OF BRAIN DEVELOPMENT

The fundamental processes of brain development involve: (1) neural induction; (2) proliferation of neurons and glia; (3) cell migration; (4) cell death; (5) cell differentiation; (6) formation of synapses; and (7) pruning of synapses. For a more detailed discussion of how these processes unfold in the frontal cortex see Chapter 7.

Neural induction. Shortly after conception, a fertilized cell undergoes a rapid process of cell division that results in a cluster of proliferating cells called the blastocyst. Within a few days, the blastocyst differentiates into a three-layered structure. Development of the nervous system begins from just a few cells in the outermost layer of the blastocyst called the neural plate. By the third or fourth week of development, the edges of neural plate begin to fold in to form a hollow structure called the neural tube. The ventricles of the brain and the spinal canal develop from the hollow area inside the tube, and the cells that make up the brain develop from the layer of cells that line the tube. First the anterior, and then the posterior end of the tube closes, after which cerebrospinal fluid fills the tube and cells begin to divide quickly to form several layers.

Neural proliferation. Neurons begin to divide rapidly, or to proliferate, at approximately the sixth foetal week, and they continue to divide until approximately the 18th week. Division of cells in the neural tube produces clones, or groups of cells that result from division of a single precursor cell. Each of the neural precursor cells, or neuroblasts, gives rise to a definite and limited number of neurons. In some cases particular neuroblasts also give rise to particular types of neuron, whereas in other cases the distinctive morphology of the type of neuron arises as a product of its developmental interactions (Marin-Padilla, 1990). As the cells divide and their numbers increase, divisions of the brain become observable in the anterior part of the neural tube.

Cell migration. Neurons are not born in the exact position they will occupy in the adult brain. Instead, they must travel or migrate from the proliferative zone where they are formed to the position they will occupy in the mature brain. The most common form of migration outside the cerebral cortex is passive cell displacement. This occurs primarily in non-layered neural structures when new cells are slowly pushed further and further away from the proliferative zone by more recently born cells. This form of migration gives rise to an “outside-to-inside” spatiotemporal gradient, with the oldest cells pushed toward the surface of the brain and the most recently produced cells remaining on the inside.

The second form of migration is called active cell migration. This occurs when young cells actively travel past previously generated cells, thereby creating an “inside-to-outside” spatiotemporal gradient. Active migration occurs both in the cerebral cortex and in some subcortical areas that have a laminar structure. In active cell migration, neurons find their way to the correct position by clinging to the long fibres of “radial glia” that radiate from the inner to the outer surface of the brain. Special adhesion molecules on the surface of the migrating cell bind to similar molecules on the glial cell or nearby axons that act as chemical cues guiding the cells along the proper pathway.

One of the most striking features of the cerebral cortex is its three-dimensional organization of layers and columns of functionally similar neurons. According to the radial unit hypothesis (Rakic, 1988), this organization is determined by a combination of the timing and location of the birth of neurons. In this view, the layered structure of the cortex is determined by the timing of the birth of cells through the process of active cell migration. The radial structure of the cortex occurs because neurons from a given clone all climb up the same radial glial fibre. In this way all of the cells produced by a single neuroblast contribute to the same radial column of neurons within the cortex. This is how the twodimensional positional information of the proliferative cells in the ventricular zone is transformed into a three-dimensional cortical structure.

Programmed cell death. At the same time that cells are proliferating and migrating, some are also dying. Programmed cell death of both neural and glial cells is part of normal development of the nervous system (Barres et al., 1992; Cowan, Fawcett, O’Leary, & Stanfield, 1984; Oppenheim, 1991). In programmed cell death, cells die as part of a gene expression-related programme of cell differentiation. Programmed cell death is characterized by cell shrinkage, rather than by the cell swelling that occurs in necrotic cell death following brain injury. Mutation of the genes involved in programmed cell death might be one mechanism responsible for the expansion of the surface area of the cerebral cortex during evolution (Rakic, 2000).

In the human telencephalon, two distinct types of programmed cell death are observable (Rakic & Zecevic, 2000). Embryonic programmed cell death occurs during proliferation and migration of neurons and is probably not related to the establishment of neuronal circuitry. Rather, it might reflect death due to errors in cell division and could play a role in elimination of transitory areas. Foetal programmed cell death occurs during neuron differentiation and synaptogenesis and might be related to the development of connections between axons and their targets. It might be one mechanism that allows the number of neurons innervating a target to match the size of the target. A small target will produce a smaller amount of chemicals to promote the survival of neurons and thus will support relatively few neurons; conversely, a larger target will produce more and thus support more neurons. In this way, the number of neurons innervating a target can be matched precisely to the size of the cell, even if this information is not preprogrammed explicitly.

Cell differentiation. Once neurons have migrated to their final positions, they further differentiate to take on their mature characteristics. One aspect of cell differentiation is dendritic arborization. The dendrites of a neuron are like antennae, picking up signals from many neurons and, if the conditions are right, passing the signal down the axon and on to other neurons. The pattern of branching of dendrites is important because it will affect the quantity and quality of signals the neuron receives. During cortical development there is an increase in size and complexity of neurons’ dendritic trees. For example, by adulthood the length of the dendrites of neurons in the frontal cortex can increase over 30 times their length at birth. Dendritic branching occurs at different times in different areas and layers of the cortex. For example, dendritic trees of cells in layer 5 of the primary visual cortex are already at about 60 per cent of their maximum extent at birth. By contrast, the mean total length for dendrites in layer 3 is only about 30 per cent of maximum at birth. Dendrite branching is one of the properties of neurons that can be influenced by the environment during development. For example, laboratory animals that receive environmental stimulation show more dendritic branching than those that do not (Wallace, Kilman, Withers, & Greenough, 1992).

A second aspect of differentiation is axon guidance and target selection. The cerebral cortex consists of a large population of excitatory glutaminergic neurons that transmit their signals using the excitatory amino acid glutamate. These neurons are reciprocally connected with the thalamus, to each other, and to a smaller population of inhibitory GABAergic neurons that, in the main, provide local circuitry (Somogyi, Tamas, Lujan, & Buhl, 1998). To find their proper connections in the brain, axons often have to stretch long distances (sometimes over a metre) to reach their target. This process depends on the ability of receptors on the growth cones of axons to recognize cues in their environment, such as molecules on the membranes of neighbouring neurons or glia, and respond with appropriate directional movement. Once in position, axons form a synapse where signals can be communicated with the dendrites of the target cell. Cadherins, a family of calcium-dependent adherent proteins, might play an important role in this process in two ways (reviewed in Ranscht, 2000): (1) by controlling synapse positioning by guiding axons to target positions that express the same type of cadherin (Inoue & Sanes, 1997); and (2) by controlling the functions of the established synapses by modulating synaptic adhesion and stability (Fields & Itoh, 1996; Tanaka et al., 2000).

Synaptogenesis. Formation of synapses in humans begins in the early weeks of gestation, when synapses can be observed in the marginal zone (Zecevic, 1998). The density of synapses rapidly increases once the cortical plate is formed and, at least in layer l of the cortex, is occurring in synchrony in different parts of the cortex at 20 weeks gestation (Zecevic, 1998). Whereas in monkeys this synchrony in synaptogenesis across cortical areas continues postnatally (Rakic et al., 1986), in humans it is believed that the time of the peak of cortical synaptogenesis differs across cortical areas. For example, in parts of the human visual cortex, synaptic density reaches a peak of approximately 150 per cent of adult levels towards the end of the first year, whereas in the frontal cortex the peak synaptic density does not occur until approximately 24 months of age (Huttenlocher, 1979, 1990; Huttenlocher & Dabholkar, 1997; Huttenlocher & de Courten, 1987; Huttenlocher, de Courten, Garey, & Van der Loos, 1982).

Elimination of synapses. Elimination, or pruning, of synapses also occurs in normal development. For example, in the primary visual cortex the mean density of synapses per neuron peaks at a level higher than adult levels, and then starts to decrease at the end of the first year of life (Huttenlocher, 1990). In humans, most cortical regions and pathways appear to undergo this “rise and fall” in synaptic density, with the density stabilizing to adult levels at different ages during later childhood. The functional role of this process will be discussed further below.

The postnatal rise and fall developmental sequence can also be seen in other measures of brain physiology and anatomy. For example, positron emission tomography (PET) measures of brain metabolism show that, although there is an adult-like distribution of resting brain activity within and across brain regions by the end of the first year, the overall level of glucose uptake reaches a peak during early childhood, which is much higher than that observed in adults (Chugani, Phelps, & Mazzidta, 1987). The rates return to adult levels after about 9 years of age for some cortical regions.

DIFFERENTIATION OF THE CEREBRAL CORTEX

The majority of normal adults tend to have similar functions within approximately the same regions of cortex. How does this happen? We cannot necessarily infer from this consistency that this pattern of differentiation is intrinsically pre-specified (“prewired”), because most humans share very similar pre- and post-natal environments. The protocortex and protomap hyptheses are the two main explanations of how the cerebral cortex comes to be subdivided into specialised functional areas.

Protocortex hypothesis

In this view, cortical neurons are initially equipotential and areal differences are induced by outside influences such as sensory inputs from the thalamus. One type of experiment that can test this view is to rewire thalamic inputs so that they project to different regions of the cortex than usual. If thalamic inputs determine cortical divisions, then the affected cortex should take in the characteristics associated with the new, rather than the normally intended, inputs. Results from several studies support this prediction. For example, following such rewiring, auditory cortex can take on visual representations (Sur, Garraghty, & Roe, 1988; Sur, Pallas, & Roe, 1990). A second type of experiment that tests the protocortex hypothesis is transplantation of embryonic cortex to a new location. If the hypothesis is true, then the transplanted cortex should develop characteristics consistent with its new location rather than its original one. Results of several experiments support this prediction. For example, if late embryonic visual cortex is transplanted to neonatal somatosensory cortex, this “visual” cortex will develop the functional characteristics of cells in the somatosensory cortex (O’Leary & Stanfield, 1989; Schlagger & O’Leary, 1991).

Despite the impressive results of rewiring and transplantation experiments, there are indications that the embryonic cortex is not completely equipotential. First, although transplanted or rewired cortex might look very similar to the original tissue in terms of function and structure, it is rarely absolutely indistinguishable from the original. For example, in the rewired ferret cortex, when auditory cortex takes on visual function, the mapping of the azimuth (angle right or left) is at a higher resolution (more detailed) than the mapping of the elevation (angle up or down; Roe, Pallas, Hahm, & Sur, 1990). By contrast, in the normal ferret, cortex azimuth and elevation are mapped in equal detail. Second, the timing of the transplant is also important. Although the results of some transplantation studies suggested that even “late” transplants of cortical tissue could develop connections appropriate to their new locations (O’Leary & Stanfield, 1989), it was not known how many neurons actually exhibited this capacity. More recent quantitative studies have shown that the majority of neurons show a pattern of connections consistent with their site of origin rather than their new location (Garnier, Arnault, Letang, & Roger, 1996; Garnier, Arnault, & Roger, 1997). Thus, at least in some cases, postmitotic neurons retain the characteristics of their original location. Finally, it is important to note that most of the experimental neurobiological studies on cortical plasticity have been performed on rodents or cats, and not primates, and it is clear that there are species differences. For example, although in mice cortical progenitors can integrate elsewhere in the telencephalon following transplantation, in some other species, cortical progenitors appear to be restricted in their ability to do so (Frantz & McConnell, 1996). Despite these caveats, the rewiring and transplantation studies do demonstrate that it is possible for anatomical and functional changes in axonal inputs to the cortex to modify existing divisions in the cortex or to create new ones.

Protomap hypothesis

Unlike the protocortex hypothesis, which emphasizes the importance of extrinsic factors in cell differentiation, the protomap hypothesis emphasizes the importance of intrinsic factors. In this view, cells generated within the embryonic cortical cell wall already contain some intrinsic information about their prospective species-specific cortical organization (Rakic, 1988). The hypothesized protomap involves either prespecification of the proliferative zone or intrinsic molecular markers that guide the division of cortex into particular areas. For example, neurons in the embryonic cortical plate might set up a beginning map that preferentially attracts the appropriate afferents to the appropriate locations and has a capacity to respond to this input in a specific manner.

The main line of evidence in support of this view comes from descriptions of graded or restricted patterns of gene expression in embryonic cortex (see Rubenstein et al., 1999 for a review). For example, the homeobox gene Emx2 is expressed in the proliferative zone that gives rise to cortical neurons at very early stages and is thought to play an important role in the early formation of cortical regions prior to the arrival of thalamic inputs (Gulisano, Broccoli, Pardini, & Boncinelli, 1996; Yoshida et al., 1997). Emx2 “knock-out” mice show deficits in neuron migration (Mallamaci et al., 2000a) and changes in molecular markers and area-specific connections (Bishop, Goudreau, & O’Leary, 2000). The fact that absence of Emx2 affects area-specific molecules, such as cadherin 6, that normally show area-specific expression before the arrival of thalamic inputs supports the view that Emx2 is involved in arealization that occurs independently of the thalamus (Mallamaci et al., 2000b).

Genetically modified mice have also been used to investigate the consequences of disrupting thalamic inputs to the developing cortex. For example, Gbx-2 mutant mice, whose thalamic differentiation is disrupted and who lack cortical innervation by thalamic neurons, show normal region-specific expressions of a variety of genes (Miyashita-Lin et al., 1999). One limita...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• List of Contributors
• Acknowledgements
• Introduction
• Chapter One
• Chapter Two
• Chapter Three
• Chapter Four
• Chapter Five
• Chapter 6
• Chapter Seven
• Chapter Eight
• Chapter Nine
• Chapter Ten
• Glossary