Journal of Child Psychology and Psychiatry 52:4 (2011), pp 409–428 doi:10.1111/j.1469-7610.2010.02281.

Annual Research Review: Parenting and

children’s brain development: the end of the
beginning
Jay Belsky1 and Michelle de Haan2
1
Institute for the Study of Children, Families and Social Issues, Birkbeck University of London, UK; 2Institute of Child
Health, University College London, UK

After questioning the practical significance of evidence that parenting influences brain development –
while highlighting the scientific importance of such work for understanding how family experience
shapes human development – this paper reviews evidence suggesting that brain structure and function
are ‘chiselled’ by parenting. Although the generalisability of most findings is limited due to a dispro-
portionate, but understandable focus on clinical samples (e.g., maltreated children with post-traumatic
stress disorder (PTSD)) and causal inferences are difficult to draw because of the observational nature of
most of the evidence, it is noteworthy that some work with community samples and very new experi-
mental work (e.g., parent training) suggests that tentative conclusions regarding effects of parenting on
the developing brain may well be substantiated in future research. Such efforts should focus on par-
enting in the normal range, experimental manipulations of parenting, differential susceptibility to
parenting effects and pathway models linking parenting to brain development and, thereby, to
behavioural development. Research on parenting and children’s brain development may be regarded as
at ‘the end of the beginning’. Keywords: Brain, parenting, EEG, fMRI, MRI, amygdala, hippocampus,
experience, emotion.

The effect of parenting on children’s development has Unless one believes in magic, it is difficult to con-
been the subject of empirical study for more than half clude on the basis of the evidence available that par-
a century. Over this period investigations have be- enting does not affect the developing brain, in terms of
come ever more sophisticated. As appreciation either structure or function – or both. After all, how
developed for the importance of the temporal ordering else would parenting or any experience for that matter
of putative cause and effects in research designs, influence a developing organism’s behaviour, cogni-
prospective longitudinal studies in which parenting tions, emotions and even, in some instances, physi-
measures preceded the assessments of child ‘out- ology and health? From this perspective, many
comes’ came to take precedence over cross-sectional behavioural scientists, including developmentalists,
studies in which parenting ‘predictors’ and child who are not brain scientists often wonder what all the
outcome were measured at the same point in time, fuss is about with regard to brain development. Al-
including such studies in which parenting was as- though as scientists they very much appreciate the
sessed retrospectively. Owing to evidence amassed reason to investigate exactly how experiences,
principally by behaviour geneticists, students of including parenting, affect brain development and,
parenting have also come to appreciate that even well- thereby, behaviour and well-being, it seems some-
designed longitudinal studies do not afford strong – or what strange that the documentation of effects on the
at least indisputable – causal inferences due to the brain can pack more of a punch, especially of the
possibility that effects of parenting may be the result journalistic or policy-making variety, than docu-
of shared genes that influence both parenting and menting effects on behavioural, cognitive, social and/
child functioning (Plomin, DeFries, McClearn, & or emotional functioning and physical health and
McGuffin, 2008). The fact, however, that genetically well-being. After all, if parenting and other develop-
informed studies chronicle parenting effects, typically mental experiences influenced the developing brain
of the non-shared-environment variety, clearly indi- but such effects did not extend to functioning in the
cates that well-documented links between parenting real world, would there be any reason for policymak-
and child development are not all a function of genetic ers and practitioners to regard evidence of such as
effects ‘masquerading’ as environmental ones important or especially meaningful?
((Plomin et al., 2008). Even more compelling in this Two critical points follow from this rhetorical
regard is experimental research that systematically question. First, understanding of influences on the
manipulates parenting and documents its indisput- developing brain and brain development, including
ably causal influence on child behaviour and devel- parenting, carries important implications for multiple
opment (e.g., van den Boom, 1994). aspects of human development and functioning.
But, second, the chronicling of links between par-
Conflict of interest statement: No conflicts declared. enting and brain structure and function, including of
 2010 The Authors. Journal of Child Psychology and Psychiatry
 2010 Association for Child and Adolescent Mental Health.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA
410 Jay Belsky and Michelle de Haan

course causal not just statistical-associational ones, mates and rats, to illustrate points when data on
is no more important, from the perspective of either humans are not readily available. For interested
basic or applied science, than documenting such readers, a very comprehensive overview of brain
links between parenting and children’s development development can be found in Stiles (2008).
more generally. As it turns out, the study of parent-
ing and brain development is not even yet in its
Development of the cerebral cortex
infancy; it would be more appropriate to conclude
that it is still in the embryonic stage, if not that which The cerebral cortex is a flat sheet of cells about 2.5
precedes conception. cm thick covering the outer surface of the brain
Nevertheless, as we plan to show, there is ever- (Pakkenberg & Gundersen, 1997). It consists of six
emerging evidence that experience shapes the devel- layers, each made up of particular types of cells and
oping brain in humans. In this review, we focus vir- connections to and from other cells. It is estimated
tually exclusively on research that has endeavoured that the human adult neocortex contains approxi-
to measure parenting in some manner – but not mately 20 billion neurons, each connected to about
characteristics of parents known to influence par- 1,000 other neurons, thereby creating a network
enting, like maternal depression – and structural with trillions of connections! It is generally believed
and/or functional aspects of the brain. As almost no that different regions of the cortex develop and dif-
work actually involves repeated brain measurements, ferentiate at different rates, with respect to both
little of the work truly qualifies as research on brain anatomical and functional characteristics.
development; we nevertheless use such terminology The key events in human cortical development
interchangeably with that of ‘the developing brain’. Of include the production of brain cells, their migration
note, too, is that we do not consider hormones or from their birth place to their appropriate position in
genetics and their role in shaping parenting and/or the cortex, and their differentiation. In humans
brain development to any substantial degree. cortical neurons are produced before birth during
As will become apparent, most of the research the sixth to 18th week after conception. At the peak
linking individual differences in parenting with indi- of proliferation, it is estimated that 200,000 neurons
vidual differences in brain structure and function that are generated every minute! Once formed, neurons
is considered herein involves truly adverse experi- typically migrate to the correct position in the cortex
ences of the kind that no one would ever purposefully by moving along the long fibres of cells called radial
impose on fellow humans in order to evaluate causal glia, which act like ropes extending from the inner to
influence. Some of this work actually involves par- the outer surface of the brain (Rakic, 1988).
enting – in the form of maltreatment. And an even Once neurons find their way to their final position,
smaller portion involves parenting in the so-called they begin to differentiate and take on their mature
‘normal range’, measured as part of community characteristics. One aspect of differentiation is the
studies rather than clinical ones in which children are growth and branching of dendrites, which increase
selected for study because of their exposure to espe- in size and complexity with development. The den-
cially problematic environmental conditions. What drites of a neuron are like the antennae that pick up
should become clear is just how limited our knowl- signals from many other neurons and, if the cir-
edge remains today. Before proceeding to consider cumstances are right, pass the signal down their
effects of experience – and particularly effects of par- axon and on to other neurons. The pattern of
enting – on brain structure and function, a brief branching of dendrites is important because it
overview of brain development is provided to set the affects the amount and type of signals the neuron
stage for what is to come thereafter. receives. Neurons also form axons, their primary
outgoing pathways that must often extend long dis-
tances to reach their targets. Growth cones on the tip
Brain development overview
of axons help to direct axon movement by sensing
Here we offer an overview of mechanisms and mile- guidance molecules, while cell adhesion molecules
stones in brain development. We focus on aspects of help to anchor the axons to the tissue substrate.
the brain and brain functioning that figure impor- Many cortical neurons also become myelinated.
tantly in the available research to be reviewed on Myelin, a fatty sheath that surrounds neurons and
parenting effects on the brain, including the cerebral helps them transmit signals more quickly, begins to
cortex, the amygdala and hippocampus. For each we form around neurons in the third trimester.
describe prenatal development of structures to pro- Although much of this process is complete by the
vide a background as to the level of brain develop- end of the second postnatal year, it continues even
ment at point of birth and how this is achieved. We into adulthood in some areas of cortex. Myelinated
then discuss what is known about postnatal brain axons send signals at velocities that are about 50–
development and the aspects which are influenced 100 times faster than unmyelinated axons (Brinley,
by experience. In most instances we focus on studies 1980, cited in Markham & Greenough, 2004).
of humans; we do, however, include reference to The points of communication between neurons are
work with other species, including non-human pri- called synapses, and these begin to form in the cor-
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 411

tex in the early weeks of gestation (Zecevic, 1998). with advances in magnetic resonance imaging (MRI)
The generation of synapses occurs at different times measures of white matter integrity this picture may
in different cortical areas. For example, the maxi- change. With respect to grey matter, most studies
mum density of synapses is reached at about four indicate that volumes initially increase in the first
months in the visual cortex but not until about 24 years of life. For example, grey matter increases by
months after birth in the prefrontal cortex (Huttenl- 149% over the first postnatal year (Knickmeyer et al.,
ocher, 1979, 1990; Huttenlocher & Dabholkar, 2008), though after about eight years grey matter
1997). Neurons use neurotransmitters to communi- begins to decrease (Giedd et al., 1999; Giedd, 2004;
cate across synapses. The acquisition of specific Hua et al., 2009; Wilke et al., 2007). Measures of
neurotransmitters and receptors by a neuron cortical thickness also show thinning with develop-
involves both processes that are determined intrin- ment (Sowell, Thompson, & Toga, 2004). Measures
sically by the cell and those that are the result of the of cortical grey matter development do tend to vary
extracellular environment in which the neuron finds by cortical region, generally displaying a pattern
itself. The most common neurotransmitters in the whereby primary sensory and motor areas develop
cerebral cortex are the excitatory neurotransmitter earlier than frontal or superior temporal regions
glutamate, which is expressed in about half of cor- (Sowell et al., 2004; Gogtay et al., 2004).
tical neurons, and the inhibitory neurotransmitter Aspects of neuronal differentiation that occur
gamma-aminobutyric acid (GABBA) which is postnatally may be influenced by postnatal experi-
expressed in 25–40% of cortical neurons. ence, and be one of the mechanisms by which chil-
At the same time that the brain is growing and dren’s experience in the environment, including the
increasing in size and complexity, seemingly parenting to which they are exposed, shapes brain
‘regressive’ events are also occurring. One example is development. This has been well documented in rats,
the elimination of synapses. During the process of with experiments showing that animals raised in en-
synapse formation, the number of synapses riched conditions have greater dendritic arborisation,
increases above the level observed in adults and increased dendritic spine density and more synapses
remains at this level for some time. Then, synapses per neuron in a number of brain areas than animals
are eliminated until the adult number is reached. For raised in impoverished environments (summarised in
example, in certain parts of the visual cortex the Markham & Greenough, 2004). The increase in den-
density of synapses per neuron peaks at about 150% drite length contributes to the increase in cortical
of the adult level at about age four months and then thickness also observed in rats raised in enriched
starts to decrease at the end of the first year of environments (Wallace, Kilman, Withers, & Green-
postnatal life to reach the adult level by about age 4 ough, 1992). These changes persist well beyond
(Huttenlocher, 1990). The timing of this process is exposure to the enriched environment and do not re-
different for different areas of cortex. This loss of flect general changes in somatic development. Similar
synapses, referred to as ‘pruning’, does not reduce changes can also be observed in adult or older animals
the range of behaviours but may be related to sta- (Markham & Greenough, 2004).
bilisation of important networks of neurons in the There is also evidence that experience can influ-
brain and the differential strengthening of circuits ence white matter (Ullen, 2009). Studies in rats show
based on both experience and genetics. This process that environmental experience has an influence on
may provide a neural basis for ‘experience-expectant myelination in the corpus callosum, discussed more
learning’, a type of environmental influence on brain fully in the next subsection: post-weanling rats
development discussed in more detail below. raised in more complex environments have greater
To summarise, the production of cortical cells and myelination in the splenium (the part of the corpus
their migration to their appropriate position occur callosum that contains visual cortical neurons;
mainly prenatally (for discussion of neurogenesis in Juraska & Kopcik, 1988). Rhesus monkeys raised in
the adult mammalian brain see Song et al., 2005). complex environments also have larger corpus call-
Differentiation of brain cells, involving synpatogen- osa (Sanchez et al., 1998) and in humans extensive
esis and synapse elimination, dendritic branching, piano practice beginning in childhood increases
acquisition of neurotransmitters and myelination, fractional anisotropy (a measure believed to correlate
begins prenatally but continues postnatally. with degree of myelination) in cerebral white matter
These early developmental changes associated (Bengtsson et al., 2005). Rat studies suggest that
with neuronal migration, synapse formation, prun- these effects of experience on myelination may be
ing and myelination may contribute to the general limited to development, as they are not found in
changes in grey and white matter volumes that have mature rats exposed to complex environments
been used as markers of brain development during (Markham, Herting, Luszpak, Juraska, & Green-
childhood and adolescence. Generally, white matter ough, 2009). This is an important point because, if
volume increases over childhood into adulthood the same applies to humans, it could indicate that
(Giedd et al., 1999; Hua et al., 2009; Wilke, Kräge- there is a critical period during which alterations in
loh-Mann, & Holland, 2007). There is limited evi- parenting are particularly effective in influencing
dence for regional differences in this process, though this aspect of brain development.
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
412 Jay Belsky and Michelle de Haan

week to three months (Rodman, 1994; Webster et al.,

Corpus callosum
1991). For example, at a time when temporal cortex
The corpus callosum deserves special attention as it anterior to area TE remains immature, the infant
has been the focus of several studies comparing the amygdala receives additional inputs from areas
brains of maltreated and non-maltreated children posterior to TE. A functional interpretation of this
considered later. The corpus callosum is the most anatomical refinement is that the amygdala receives
prominent white matter structure in the brain and increasingly refined and detailed visual information
contains about 200 million myelinated fibres, most over this period (Payne & Bachevalier, 2009).
of which connect corresponding areas of the left and Myelination in the human amygdala begins in the
right hemispheres. It is generally thought to function first months of life, with some aspects appearing
to integrate the activities of the two hemispheres. In mature by 10 months of age (Brody, Kinney, Kloman,
line with the general trend for cortical white matter to & Gilles, 1987; Kinney, Brody, Kloman, & Gilles,
increase over childhood, research indicates that all 1988). Studies of monkeys indicate, however, that
subregions of the corpus callosum increase in vol- mature levels of myelination of amygdala output
ume over childhood into young adulthood, though fibres are not reached until at least three years after
there may be some regional variation in the rate of birth (Machado & Bachevalier, 2003). Together,
growth (Hasan et al., 2009). these findings suggest that the amygdala’s influence
over other brain areas influence increases during the
first postnatal years. Neuroimaging studies also
Subcortical development
indicate a protracted period of grey matter develop-
Generally speaking, subcortical structures develop ment from 4 to 18 years. The volume of grey matter
earlier than cortical structures. This section focuses increases over this time, though some studies find
on the subcortical structures relevant to the studies such changes are restricted to the right amygdala
described in this paper, the amygdala and hippo- and are observed only in girls (reviewed in Payne &
campus. Bachevalier, 2009; Ostby et al., 2009; Wilke et al.,
2007). The amygdala is also connected to the
Amygdala. The amygdala is a subcortical group of orbitofrontal cortex, which plays an important role in
13 interconnected nuclei located in the anterior social and emotional behaviour, though little is
portion of the medial temporal lobe. (For a more known about the development of this connection.
detailed review of amygdala development, see Payne
& Bachevalier, 2009.) The lateral nucleus is believed Hippocampus. The hippocampus is a layered
to be particularly relevant to social processing, structure located on the floor of the lateral ventricles
because anatomical studies of monkeys show that it that is believed to be important for spatial naviga-
receives input from the pulvinar, a subcortical tion, learning and memory. It consists of the CA1–4
pathway that is thought to mediate rapid orienting regions of Ammons’ horn which contain mainly
to socially relevant stimuli, and also because it pyramidal cells and the dentate gyrus. The perforant
receives, from the cortex, inputs of highly processed pathway from the entorhinal cortex provides the
visual information regarding faces, facial expression, main input to the hippocampus. Formation of hip-
gaze direction and body movements. The lateral pocampal neurons and their migration to their
nucleus projects back to cortical areas via the basal positions occurs earlier for the CA fields than the
nucleus, both to higher-order cortical areas as well dentate gyrus (Seress, 2001). In the CA fields, cells
as primary sensory ones. It is thus able to modulate are mainly formed by 15 weeks and completed by
various parts of the cortical network for social pro- 20–24 weeks, with cell migration finishing prena-
cessing, including early sensory regions such as the tally. By contrast, in the granule cell layer of the
fusiform gyrus. dentate gyrus neuron formation continues as late as
The human amygdala is first observed by five 34–36 weeks’ gestation and cell migration continues
weeks’ gestation, with its nuclei becoming discern- until one year postnatally. Overall about 30% of
ible by the early stages of the second trimester. Data granule layer neurons begin to proliferate and
regarding the development of the connections of the establish their connections only postnatally. The
amygdala come primarily from studies of monkeys. elements of the basic hippocampal circuit which
This research suggests that most of amygdale–cor- connects dentate gyrus to CA3 fields, CA3 to CA1,
tical connections are already established by the time and then CA1 providing an output route, are in place
of birth or soon afterwards (Amaral & Bennett, at birth but continue to show significant develop-
2000), including reciprocal connections with the ments until at least 5 years of age (Seress, 2001).
inferior temporal cortex (Webster, Ungerleider, & MRI studies indicate that the hippocampus
Bachevalier, 1991). However, these connections do increases in size over the first years of life, showing,
undergo some change over the first postnatal year. In for example, a 13% increase from age 1 to 2 years
infant monkeys the inferior temporal inputs to the (Knickmeyer et al., 2008). Research examining the
lateral nucleus of the amygdala are more widespread development of the hippocampus from the preschool
than in adults, and become more refined from one period into early adulthood reveals regional varia-
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 413

tions, with the posterior hippocampus subregions This process of pruning is itself regulated by com-
increasing in volume and the anterior subregions petitive interactions between neuronal connections
showing loss (Gogtay et al., 2006). (Courchesne, Chisum, & Townsend, 1994), such
Animal research demonstrates that stress and the that those neurons that remain inactive or are rarely
glucocorticoids secreted during stress can be neu- activated are eliminated, whereas those that are
rotoxic to the hippocampus. However, this has not actively stimulated by experience are strengthened
been empirically documented in human samples. and maintained (Edelman, 1987; Greenough et al.,
One small study of 14 children showed that post- 1987). This brain malleability or plasticity is a
traumatic stress disorder symptoms and cortisol function, then, of both timing and thus the basic
levels at baseline subsequently predicted hippo- developmental programme guiding the brain and the
campal volume reduction over a 12–18-month period individual’s encounters with the environment (i.e.,
(Carrion, Weems, & Reiss, 2007). Other work also experience). As a result, severe deprivation or other
indicates that parenting can influence children’s experiences that are in some sense abnormal and
neuroendocrine response to stress (e.g., Kertes et al., not species typical may adversely and enduringly
2009), providing a potential mechanism by which affect brain structure, function and development
variations in parenting could impact the developing (Black et al., 1998).
brain via influences on hormonal responses which in Experience-dependent synapse formation, the
turn influence the brain. third process under consideration, is a function of
the individually unique experience of the individual,
typically taking place later in time than experience-
expectant processes (Cicchetti, 2002). In conse-
Experience and the developing brain
quence, each person’s brain comes to reflect, at least
Even though the brain is substantially shaped by in part, his or her unique experiential history.
genetic processes (Rakic, 1988), it remains the case Moreover, ‘experience-dependent synaptogenesis is
that brain development occurs in interaction with the localised to the brain regions involved in processing
environment (Greenough, Black & Wallace, 1987; information arising from the event experienced by
O’Leary, 1989). ‘To understand neuropsychological the individual’ (Cicchetti, 2002, p. 1413). These
development is to confront the fact that the brain processes are less subject to the stringent temporal
is mutable, such that its structural organisation constraints – and opportunities – than experience-
reflects the history of the organism’ (Luu & Tucker, expectant ones. Rather than involving, then, the
1996, p. 297). The balance between genetic and overproduction and subsequent pruning of syn-
environmental influences on brain cortical develop- apses, a process restricted to early development, this
ment changes over development and also differs third brain development mechanism involves for-
across cortical regions (Lenroot et al., 2009). To be mation of new synapses and/or the modification of
appreciated as well is that cortical development and existing ones and can take place, to greater or lesser
organisation are no longer regarded as passive pro- degree, across the life course. As a result, many
cesses solely dependent on genetics and environ- sources of influence can be envisioned, ranging from
mental input, but rather are self-organising, guided everyday social experience, especially if repeatedly
by self-regulatory mechanisms (Cicchetti, 2002). encountered, to planned interventions, such as
Brain development has been described as a com- psychotherapy or occupational training.
plex scaffolding of three types of neural processes To be appreciated, of course, is that such influ-
(Black, Jones, Nelson, & Greenough, 1998). First are ences can maintain, even enhance normal func-
gene-driven processes, ones considered more or less tioning but, at the same time, adversely affect it.
impervious or insensitive to experience. Such pro- ‘Early stresses, either physiological or emotional,
cesses protect the developing brain, guide neuron may condition young neural networks to produce
migration, and target many of their synaptic con- cascading effects through later development, possi-
nections, while also determining differentiated bly constraining the child’s flexibility to adapt to new
functions (Rakic, 1988). challenging situations with new strategies rather
Second are experience-expectant processes, than with old conceptual and behavioural prototypes
taking place when the brain is primed to receive … Accordingly, abnormal perturbations at one stage
particular classes of information from the environ- of brain development hinder the creation of some
ment – and thus ‘expectant’. These correspond to new structures and functions, distort the form of
critical or sensitive periods during which an over- later-emerging ones, make possible the construction
abundance of synapses characterises the brain (or of ones that normally never become manifest, and/or
brain structure), ones which are eventually reduced. limit the elaboration and usage of structures and
The period of synapse overproduction is a genetically functions that had appeared earlier’ (Cicchetti, 2002,
programmed occurrence that allows for the emer- p. 1428). Clearly, then, there are many ways in
gence of basic skills which are then used in inter- which brain development can go awry. Not to be
action with the environment to guide the subsequent forgotten is that children vary in the extent to which
elimination of excess synapses (Huttenlocher, 1994). they prove resilient to adversity, whether it takes the
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
414 Jay Belsky and Michelle de Haan

form of poverty, community violence, parental abuse is behavioural evidence that maltreated children
and neglect, or harsh discipline. show a variety of intellectual and academic impair-
Indeed, recent theory and evidence suggests that ments (e.g., Perez & Widom, 1994; Pianta & Egeland,
individuals differ in their susceptibility to environ- 1994), as well as neuropsychological deficits,
mental influences (Belsky & Pluess, 2009a, 2009b), including impairments in executive functioning and
with some being more and some being less affected attention (Beers & De Bellis, 2002) and everyday
by conditions that both undermine well-being, memory (Moradi, Doost, Taghavi, Yule, & Dalgleish,
broadly conceived, or enhance it (Belsky, 1997, 1999). In what follows, we first consider research on
2000, 2005; Boyce & Ellis, 2005). In other words, maltreatment and how it may affect brain structure,
whereas some individuals are substantially affected, before turning to evidence pertaining to brain pro-
‘for better and for worse’, by, respectively, positive cesses. As all the research to be reviewed is non-
and negative experiences (Belsky, Bakermans- experimental, causal inferences can be made only
Kranenburg, & van IJzendoorn, 2007), others are far with extreme caution. In some cases the work to be
less affected and may even be immune to the effects reported is based on studies of children growing up
of the very same experiences. An ever-growing body in extremely deprived Romanian orphanages; in
of research indicates that phenotypic factors (e.g., some respects such experience can be likened to
temperament), endophenotypic factors (e.g., physio- extremely severe (parental) neglect. But because
logical reactivity) and genetic factors function as such deprivation involves much more than just lim-
‘plasticity markers’ and distinguish those who are itations in caregiving, it is no doubt mistaken to
more and less prone to environmental influence of equate it completely with – and thus expect exactly
perhaps many kinds (Belsky, Jonassaint et al., 2009; the same effects of – the kind of parental neglect
Belsky & Pluess, 2009b; Obradović & Boyce, 2009). typically encountered by child protection workers
This new and emerging perspective on why and dealing with troubled families.
whether, not just how, early experiences, including
parenting, may affect life-course development carries
Brain structure
implications for the study of the effects of parenting
on brain development. Most importantly, it implies Evidence that maltreatment, sometimes when cou-
that failure to distinguish those likely to be more and pled with PTSD, appears to influence brain structure
less influenced by parenting could result in comes from a series of studies carried out in the
researchers both over- and under-estimating effects laboratories of De Bellis and Teicher (see below).
on brain development, with effects on those most Some of this work focuses on corpus collosum, some
susceptible being under-estimated and effects on on the hippocampus, some on the amygdala and the
those least susceptible being over-estimated. As we most recent on cortical grey matter regions, adopting
will see, only a single study cited in the conclusion of a whole brain approach. Before saying more about
this report has considered this possibility. Thus, as this groundbreaking work, it must be appreciated
will be noted when this work is discussed, one that owing to the clinical settings in which most of
important direction for future work will be to con- this research has been conducted, much of the evi-
sider interactions between parenting and pheno- dence to be presented is limited by reliance on clin-
typic, endophenotypic and/or genetic plasticity ical samples consisting of abused individuals with a
markers when it comes to evaluating effects of par- specific form of psychopathology, often PTSD, as
enting on brain development. already noted. In consequence, much of the work
considered in this section may overestimate effects of
abuse by selecting the most adversely affected indi-
viduals, confounding abuse-related differences with
Effects of parenting on the brain
disorder-related differences or mis-identifying brain
The possibility that maltreatment (in its myriad abnormalities that were risk factors for developing a
varieties: emotional, physical, sexual, neglect) could specific disorder when exposed to trauma rather
affect brain development is suggested by research than regions altered by exposure (Tomoda, Navalta,
showing that maltreated children and adolescents Polcari, Sadato, & Teicher, 2009a). The work can be
with mood and anxiety symptoms (i.e., PTSD symp- regarded as important nevertheless, owing to the fact
toms) show evidence of altered catecholomines and that it is based on prior early-experience research
hypothalamic–pituitary–adrenal (HPA) axis activity with animals. Indeed, the research considered first
(Carrion et al., 2007; De Bellis et al., 1999a, 1999b). pertaining to the corpus collosum is informed by
And this is because elevated levels of catecholamines such work showing that stressors and experimenter
and cortisol may lead to adverse brain development handling affect this brain structure in rats, espe-
through the mechanisms of accelerated loss (or cially in males (Berrebi et al., 1988), as well as
metabolism) of neurons, delays in myelination, experiments with rhesus monkeys showing that
abnormalities in developmentally appropriate prun- males isolated from the typical social group show
ing, and/or the inhibition of neurogenesis (De Bellis substantial reduction in the midsaggital area of the
et al., 2002). Certainly consistent with this argument corpus collosum (Sanchez et al., 1998).
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 415

corpus collosum measurements and total brain size

Cortical grey and white matter: regional analyses
was also unrelated to any abuse condition.
As just mentioned, the corpus callosum is one region Subsequent work by De Bellis and associates
of white matter that has received particular attention (1999b) comparing 44 maltreated children and
in studies of the neural correlates of child maltreat- adolescents suffering from PTSD with 61 matched
ment. Although the clinical consequences of reduced controls on measurements obtained from an ana-
corpus callosum area are not fully understood, tomical MRI brain scan, as well as assessments of
reduced size is associated with diminished commu- psychiatric and neuropsychological functioning,
nication between brain hemispheres, perhaps reveals other apparent effects on brain structure and
reflecting, if not causing, reduced integration (Tei- function. Matching was done on age, gender, hand-
cher, Tomoda, & Andersen, 2006). Certainly con- edness, height, weight, pubertal status and race;
sistent with this possibility is Schiffer, Teicher, and controls had no history of disorder or of trauma or
Papanicolaou’s (1995) early work using auditory maltreatment. The participants in this research were
evoked potentials to study laterality and hemispheric judged to be similar to most studies of maltreated
integration of memory in adults with and without a children that find significantly increased rates of
history of childhood maltreatment, all of whom were internalising and externalising disorders in abused
well functioning at the time of assessment. When children. Maltreated/PTSD children had signifi-
asked to actively recall, first, a neutral or work- cantly smaller intracranial and cerebral volumes
related memory and then, with associated affect, a and, after adjusting for socioeconomic status (SES),
disturbing memory from childhood, both hemi- proportionately smaller intracranial and cerebral
spheres appeared to be equally involved in normal volumes, than comparison children. Moreover, total
individuals. In adults with childhood trauma, results midsagittal area of the corpus callosum, particularly
were entirely different. A marked suppression of the its middle and posterior regions (4, 5, 6 and 7),
evoked potential response over the left hemisphere proved smaller than controls. Of note, too, is that the
(indicative of increased left hemisphere processing) longer the duration of maltreatment, the smaller the
occurred during recall of the neutral memory but, intracranial volume and total corpus callosum, as
during recall of the disturbing childhood memory, well as its middle and posterior regions, clearly
there was a substantial shift in laterality. Evoked suggesting a dose–response relation. Important from
potential response was suppressed over the right a psychological perspective is that the smaller these
hemisphere, indicating enhanced right hemisphere structures, the greater PTSD cluster symptoms of
activation, thereby suggesting that early maltreat- intrusive thoughts, avoidance, hyperarousal and
ment is associated with increased hemispheric lat- dissociation proved to be.
erality and decreased hemispheric integration. Effects of maltreatment/PTSD have also been
Teicher et al. (2006) suggest that this may contribute reported for the lateral ventricles. Consistent with
to psychiatric illness – or reflect it. evidence that lateral ventricular enlargement is
In one of the first studies to examine potential related to many psychiatric disorders (i.e., childhood-
effects of maltreatment on the corpus callosum, and adult-onset schizophrenia, Alzheimer’s disease,
Teicher and associates (1997) examined the MRI alcoholism, bipolar disorder, major depression with
records of 51 of 115 consecutive paediatric patients psychosis), longer durations of abuse proved related
who had such records and who had normal intelli- to increased lateral ventricular enlargement (De
gence; the mean age of the children was almost 13 Bellis et al., 1999a, 1999b). There was also evidence,
years and maltreatment status was based on medi- consistent with animal studies cited earlier, that
cal records. All anatomical measurements of the many of the chronicled adverse effects of maltreat-
corpus callosum were obtained from the midsaggital ment/PTSD were more pronounced in, or even
image; regional volume of the corpus callosum was restricted to, males. Such results accord with the
corrected for brain volume (to take into account age view that males are more vulnerable to adversity.
and gender). Physically abused males and neglected In a second study of an independent sample of 28
males each showed a roughly 25% reduction in the maltreated children and adolescents with PTSD and
relative volume in region 4 of the corpus callosum 66 healthy, non-maltreated controls, De Bellis and
which largely interconnects right and left motor associates (2002) replicated – and extended – many
cortex, with neglect also being associated with a of the findings just highlighted. The maltreated/
similar reduction in region 3 which interconnects PTSD subsample had smaller intracranial, cerebral
pre-motor regions. In view of the fact that abnor- and prefrontal cortex, prefrontal white matter, and
mally large regions of the corpus callosum can also right temporal lobe volumes and areas of the corpus
reflect pathology, it seems noteworthy that physi- callosum and some of its subregions (2, 4, 5, 6 and 7)
cally abused females had a nonsignficant 17% than controls. The total midsagittal area of the cor-
increase in region 4 volume and significant increases pus callosum and middle and posterior regions
in regions 1 (89%) and 6 (51%); the latter connects remained smaller in the index group. As in the prior
superior temporal/posterior parietal regions. Neither study, brain volumes correlated negatively with
psychological nor sexual abuse proved related to duration of abuse and males seemed more vulnera-
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
416 Jay Belsky and Michelle de Haan

ble to adversity with regard to some brain measure- (2009) on 14 adolescents adopted into English
ments (e.g., lateral ventricular volume). Some of the homes before age 3.5 years, magnetic resonance
findings emanating from the investigators’ previous imaging revealed no overall or subregional differ-
inquiry could not be replicated, however, particu- ences on the midsagittal slice when comparisons
larly ones linking brain structure measurements were made to 11 agemates without such adverse
with PTSD symptoms (e.g., negative associations early experience. Perhaps the failure to ‘replicate’ the
between PTSD cluster symptoms with intracranial or abuse/neglect findings under consideration should
cerebral volumes and positive associations with have been expected given the earlier comment that
ventricular volumes). there is risk associated with equating the kind of
In a third study linking maltreatment-related neglect which child protection workers typically
paediatric PTSD with brain structure, this one of 61 encounter with what children experience in some
medically healthy children with chronic PTSD sec- orphanages.
ondary to abuse and 122 healthy controls, all 4–17 Regional analyses do indicate that formerly insti-
years of age who had participated in the prior De tutionalised children show reduced volumes in spe-
Bellis et al. (1999b, 2002) investigations, De Bellis cific regions of the cerebellum. In research by Bauer
and Keshavan (2003) detected smaller midsagittal and associates (Bauer, Hanson, Pierson, Davidson,
area of the corpus callosum subregion 7 (splenium) & Pollak, 2009), 31 9–11-year-olds raised in
in the index group. Moreover, these same mal- orphanages for 4–77 months after birth and adopted
treated PTSD children did not manifest the normal when they were between 10 and 92 months old were
age-related increases in the area of the total corpus compared to similar-aged, never-institutionalised
callosum and its region 7 which was evident in the controls. Post-institutionalised children showed
comparison sample. Most notably, males in partic- reductions in the left and right superior-lateral lobe
ular seemed adversely affected, showing smaller of the cerebellum but not in six other cerebellar
cerebral volumes and corpus callosum regions 1 regions analysed. Of developmental significance is
(rostrum) and 6 (isthmus) and greater lateral ven- that these smaller volumes predicted poorer perfor-
tricular volume increases when compared to their mance on tests of visual memory and planning.
healthy counterparts. These latter findings led the
investigators to conclude that ‘maltreated males with
Cortical grey volume: whole brain approach
PTSD show more evidence of adverse brain devel-
opment’ than their female counterparts (De Bellis & As previous work indicates that children suffering
Keshavan, 2003, p. 114). extreme neglect and deprivation in their first years of
Further evidence that maltreatment is associated life – by being raised in Romanian orphanages – have
with reduced corpus callosum area comes from Tei- smaller head circumferences than controls, it seems
cher and associates (2006). In preliminary work, unsurprising that such children, as adolescents,
Teicher et al. (1997) reported a substantial reduction also show reduced grey and white matter volumes in
in the midsaggital area of the corpus callosum in a recent MRI study (Mehta et al., 2009; see also
psychiatrically ill children with a history of abuse or Eluvathingal et al., 2006). Of note, too, is that, in the
neglect, with males again being more adversely most recently reported work being carried out in
affected than females. In later research on 51 children Teicher’s lab, attention has shifted from investigat-
admitted for psychiatric evaluation (28 with abuse or ing the size of particular brain regions to a whole
neglect) and 115 healthy controls, Teicher and his brain approach to identify, using voxel-based mor-
collaborators (2004) found, after controlling for age phometry, alterations in regional grey matter volume
and midsaggital area, that the corpus collosum area (GMV). Moreover, in order to overcome some of the
of the abused/neglected children was 17% smaller inferential problems posed by studying clinical pop-
than controls and 11% smaller than fellow psychiat- ulations, the most recent investigations recruit
ric patients who had not been mistreated; the two individuals from the community regardless of their
comparison samples did not differ from each other. psychiatric status (or lack thereof). In one such
Additional analyses revealed neglect to be the stron- study, Tomoda and associates (2009a) compared 23
gest experiential factor associated with (15–18%) unmedicated females with histories of sexual abuse
reductions in corpus callosum regions 3, 4 5 and 7; and 14 healthy controls of equivalent age (range: 18–
sexual abuse was the strongest factor associated with 22 years) and socioeconomic status with no history
reduced corpus collosum region in girls. of trauma. GMV was reduced by 12.6% and 18.1%
In light of the apparent effects of maltreatment, in, respectively, right and left primary visual (V-1)
including both abuse and neglect (and sometimes and visual association cortices of abused individu-
PTSD, too), on diverse measurements of the corpus als, results which also emerged when nine abused
callosum, it would seem somewhat surprising that females who met criteria for an Axis 1 psychiatric
no such effects emerged when the focus of inquiry disorder were excluded from analysis. Moreover, not
was on the kind of extreme neglect and deprivation only did longer exposure to sexual abuse before age
experienced in the first years of life in Romanian 12 – a cut-off based on prior human and animal work
orphanages. In this work by Mehta and associates on the development of the visual system – predict
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 417

greater reductions in grey matter, but GVM of left sexual maltreatment, as well as neglect, on brain
and right V-1 correlated positively and significantly structure. Recent research by Choi and associates
with an overall index of visual memory. A marginal (Choi, Bumseok, Rohan, Polcari, & Teicher, 2009)
relation also emerged linking GMV in left V-1 with the extends the study of maltreatment effects to a central
capacity to distinguish targets from non-targets on feature of emotional maltreatment, namely, verbal
the Go/No-Go-Stop Continuous Performance Task. abuse. Diffusion tensor imaging was used to deter-
These findings suggest that sexual abuse, by reduc- mine whether such rearing experience was associ-
ing grey matter in select areas of the brain, adversely ated with abnormalities in white matter tract
affects select aspects of psychological functioning. To integrity by comparing 16 healthy young adults
be appreciated, however, is that the investigators did reporting high levels of verbal abuse while growing
not predict exactly which grey matter areas might be (but not to other forms of maltreatment) and 16
affected by sexual abuse and were thus left to spec- healthy controls, after screening 1,271 healthy
ulate as to the basis of their findings. Perhaps, they young adults for exposure to childhood adversity.
concluded, the child’s brain may seek to reduce The specific focus of this inquiry was based on
abuse-induced stress by attenuating the develop- research summarised above pertaining to the corpus
ment of sensory systems and pathways relaying callosum, the brain’s most extensive white matter
recurrent aversive or traumatic experience. tract. Diffusion tensor imaging, which analyses the
In a second study using voxel-based morphometry restricted diffusion of water molecules, affords a
to measure regional GVM, Tomoda and associates more detailed assessment of fibre tracts than con-
(2009b) turned their attention away from sexual ventional MRI, providing a powerful technique for
abuse, indeed away from abuse per se, to harsh (but investigating the role of neural connectivity in health
apparently not abusive) corporal punishment as and disease (Catani, 2006).
reported by a community sample of 18–25-year-olds. After controlling for parent education and income,
When 23 individuals with exposure to a minimum of group differences in fractional anisotropy for three
three years of such experience, involving at a mini- white matter tract regions emerged favouring the
mum 12 episodes per year, 10 of which involved comparison cases (i.e., greater fractional anisotropy
objects, were compared to 22 healthy controls with at in each area). Indeed, analysis indicated that a his-
most minimal such experience using high-resolution tory of verbal abuse was strongly associated with
T1-weighted MRI, apparent effects of rearing history reduced fractional anisotropy in the left superior
also emerged. Relative to the latter group, children temporal gyrus (which was itself positively related to
with extensive histories of harsh corporal punish- verbal IQ and verbal comprehension), the cingulum
ment manifest GMV reductions of almost 20% in the bundle by the posterior tail of the left hippocampus
right medial frontal gyrus, almost 15% in the left (which was itself negatively correlated with ratings of
medial frontal gyrus and a little more than 15% in the depression, dissociation and limbic irritability), and
right anterior cingulated gyrus. These regions are the left body of the fornix (which was itself negatively
part of the medial rostral prefrontal cortex. Work correlated with ratings of somatisation and anxiety).
cited by the investigators indicates that this general Of note is that levels of verbal abuse reported in this
area of the human brain plays a crucial role in social particular study were not especially high, leading the
cognition, such as person perception, self-knowledge authors to surmise that the high-verbal-abuse group
and mentalising, as well as functional organisation, would qualify as controls in other studies.
including internal monitoring of one’s actions. What remains unclear, according to Choi and
Moreover, the specific regions related to rearing associates (2009), is why parental verbal abuse was
experience in this study have been linked in other associated with fractional anisotropy, as it is
imaging research to addiction, suicidal behaviour, unlikely that such treatment directly affects the
depression and dissociative disorders. GMV in these number of axons because this is generally estab-
identified regions did not correlate with psychiatric lished early in childhood. One hypothesis advanced
symptoms, however, though they did relate positively was that verbal abuse affects axon diameter,
and significantly with performance IQ. It remains microtubular structure, and the proportion of mye-
possible, of course, that the null findings pertaining linated and unmyelinated fibres that constitute a
to psychiatric symptoms reflected limited statistical component of the pathway and that these pathways
power more than anything else. Not to be forgotten in are established later in development (Keshavan
drawing conclusions from this small-sample work is et al., 2002) and appear susceptible to effects of
the general scientific principle that absence of evi- experience during preadolescence and peripubertal
dence (of the link to psychiatric symptoms) should periods (Juraska & Kopcik, 1988). In any event, as
not be regarded as evidence of absence (of such links). the authors conclude, results proved consistent with
the hypothesis that ‘the brain is chiseled in precise
ways by exposure to adverse early experiences.
White matter: whole brain approach
Diminished fibre integrity, aberrant crossing pat-
All the maltreatment work considered through this terns, alterations in axonal diameter, or extent of
point chronicles putative effects of physical and myelination along portions of these pathways’ may
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
418 Jay Belsky and Michelle de Haan

underlie some of the psychiatric and neurocognitive ume. In a preliminary report of magnetic resonance
sequelae of child abuse (Choi et al., 2009, p. 233). imaging-based measurements of hippocampal vol-
Work by Eluvathingal and associates (2006) ume in PSTD cases related to childhood physical and
examined connectivity of white matter pathways in sexual abuse, a reduction in left hippocampal vol-
the limbic systems at age 7–11 years in a group of ume was documented (Bremner et al., 1997), a
Romanian children with normal intelligence who had finding replicated by Stein and associates (1997) in a
been institutionalised and then adopted between the study of women victimised by childhood sexual
ages of 17 and 60 months, and in control children abuse. Of note also is more recent work chronicling a
matched in age who were never institutionalised. 15–18% reduction in left hippocampal volume in
Overall, previously institutionalised children showed women with a history of pre-pubertal physical and/
lower connectivity than control children, with the or sexual abuse and depression relative to healthy
difference reaching statistical significance for the control or depressed women lacking histories of
uncinate fasciuculus which connects the limbic maltreatment (Vythilingam et al., 2002). Rao and
system with the frontal lobes. The authors suggest associates (2010b) recently reported somewhat
that the lowered connectivity could be caused by related findings, examining the effect of early
lower levels of myelination, fewer fibres, or disorga- adversity, defined in part by (child and/or parent
nisation of the pathway. report of) degree of physical neglect, emotional
In view of the correlational nature of virtually all of abuse/assault, physical abuse/assault and sexual
the human evidence reviewed through this point, a abuse/assault to which adolescents had been
recent randomised controlled trial to evaluate the exposed before age 11. After controlling for chronic
effects of training parents of preterm infants in stress during adolescence, greater early adversity
reducing stressful experiences on brain development predicted smaller left hippocampal volumes in teen-
must be considered important (Milgrom et al., 2009). agers with no personal history of depression.
Evidence indicated that such training was associated The aforementioned work by De Bellis and asso-
with enhancement in DTI measures of maturation ciates (1999a, 1999b, 2002), however, failed to rep-
and connectivity of white matter at term equivalent licate these hippocampal results, with the same
age, but with no effects on grey or white matter vol- being true of Carrion et al. (2001). Teicher et al.
umes or short-term medical outcomes. Clearly, more (2006) observe that numerous hypotheses could be
work of this kind is what is required before strong advanced to account for why exposure to childhood
causal inferences can be drawn regarding ways in abuse seems to be linked to reduced hippocampal
which parenting shapes brain development. volume in adulthood, but normal hippocampal vol-
ume in childhood. One possibility that received
support in research on rats suggested that the nor-
Subcortical structures
mal pruning of synapses may lead to effects not
Attention is now turned to two subcortical struc- proving detectable until later in life (Andersen &
tures, the hippocampus and the amygdala, to fur- Teicher, 2004, 2009). Timing of maltreatment may
ther consider possible effects of parenting on brain be another factor that helps to explain the (seemingly
structure. inconsistent) findings at hand.
Indeed, when Andersen and associates (2008)
Hippocampus. Work not focused on humans and/ examined volumetric MRI scans from 26 women with
or parenting per se provides grounds for anticipating repeated episodes of childhood sexual abuse and 17
adverse effects of maltreatment on the hippocampus. healthy female comparisons, all of whom were 18–22
Classical studies by Sapolsky and McEwen demon- years of age, age of victimisation proved central to
strate that this brain structure is vulnerable to pro- illuminating which region of the brain was affected in
longed exposure to stress hormones (Sapolsky, Krey, this preliminary study. Hippocampal volume was
& McEwen, 1985). In particular, stress and corti- most reduced when childhood sexual abuse
costeroids can alter pyramidal cell morphology occurred at 3–5 years of age, but also at 11–13,
markedly (Sapolsky et al., 1990), producing pyra- periods of time associated with the overproduction
midal cell death, while suppressing the production of phases of human hippocampal grey matter (Gogtay
new granule cells (Gould & Tanapat, 1999). Just as et al., 2006). Conceivably, then, at least if studies of
importantly, stressful variations in maternal care by the rat are to serve as a guide, early stress prevents
rat dams affect neuroendocrine control mechanisms, the normal peripubertal overproduction of synapses
including adrenocortical response (Levine, 1967; in select regions of the hippocampus (CA1, CA3), but
Sapolsky & Meaney, 1986), with similar findings does not prevent pruning; and this leads to an
reported in primates (Maestripieri, Lindell, Ayala, enduring deficit in synaptic density by late adoles-
Gold, & Higley, 2005). Of note as well is that effects cence/early adulthood (Andersen & Teicher, 2004).
on hippocampal development have also been The ages associated with hippocampal reductions
chronicled in rats (Andersen & Teicher, 2004). certainly suggests that brain regions have unique
As it turns out, early clinical studies document windows of vulnerability to the effects of traumatic
links between maltreatment and hippocampal vol- stress. Indeed, the notion of early vulnerability of the
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 419

hippocampus, even if delayed in manifestation, is neglect and deprivation experienced in institutions

consistent with morphometric measures that show with that which social workers encounter in troubled
that the hippocampus has obtained 85% adult vol- families.
ume by 4 years of age (Giedd et al., 1996). The idea Another recent study, this one of mothering
that the period of 3–5 years represents a time when behaviour in the normal range, also discerned an
the hippocampus is especially influenced by experi- apparent effect of parenting on amygdala volumes
ence is echoed in a recent study examining the (Whittle et al., 2009). Adolescent boys whose mothers
influence of variation of early parental care and reacted with punishing responses to their offspring’s
environmental stimulation on hippocampal volume positive affect during pleasant event-planning inter-
in adolescence in families with no evidence of abuse actions, but not during conflictual problem-solving
or neglect (Rao et al., 2010a). This work examined a interactions, showed larger right amygdala volumes.
group of 49 children participating in a longitudinal Just as importantly, larger amygdala volumes
study in which the Home Observation for Measure- predicted, in turn, adolescents’ aggressive behaviour
ment of the Environment (HOME) scale was admin- when interacting with their mothers in a problem-
istered at 4 and 8 years of age and hippocampal solving task (Whittle et al., 2008). Evidence like this
volumes assessed at 14 years of age. Parental nur- clearly suggests how parenting comes to predict –
turance, but not environmental stimulation, at 4 and apparently influence – behavioural development
years of age were related to later hippocampal size. (i.e., parental punishment fi larger amygdale vol-
By contrast, neither HOME measurement at 8 years umes fi child aggression). Moreover, it underscores
related to later hippocampal measures. Somewhat the need extend studies of parenting and brain
counterintuitive was the finding that more parental development by including measurements of real-
nurturance at age 4 was related to smaller hippo- world performance.
campal volumes in adolescence. As noted by the
authors, this negative association between parenting Interim conclusion
and hippocampal size during childhood could be
regarded as consistent with results from studies of Clearly there is repeated indication that, as
paediatric PTSD, as these tend to show no difference expected, exposure to maltreatment, whether it
in, or sometimes larger, hippocampi in children with takes the form of physical, sexual or emotional
PTSD compared to controls. abuse, or the kind of extreme neglect and deprivation
Further evidence of the influence of early rearing demarcated by experience in Romanian orphanages,
on hippocampal volume comes from the work of is related to brain structure. Not always clear from
Buss and associates (2007). Low birth weight sig- this work, however, is whether causal inferences can
nificantly predicted smaller hippocampal volume in be drawn. This would seem to be especially the case
adulthood, but only in females retrospectively where PTSD is associated with abuse, as it is difficult
reporting poor maternal care. These findings were to determine whether PTSD is the cause or conse-
seen to suggest that good maternal care buffers the quence of the brain measurements obtained, though
adverse neurodevelopmental consequences of pre- it is not difficult to imagine multiple and reciprocal
natal risk, thereby providing evidence of a positive pathways of influence. Given how little work to date
effect of maternal care on brain development, at least focuses upon non-extreme or pathological parenting,
in the case of children otherwise at risk (of smaller that is, parenting in the ‘normal range’, it seems
hippocampal volume). noteworthy that some of the findings reviewed deal-
ing with such experience are not inconsistent with
Amygdala. Perhaps surprising and of interest is those emerging from studies of more extreme rearing
that many studies have failed to document relations conditions. This raises the possibility, though can-
between maltreatment and amygdala volume not yet confirm it, that effects and perhaps even
(Andersen et al., 2008; Bremner et al., 1997; De processes of influence detected in studies of extreme
Bellis et al., 1999b, 2002; Stein et al., 1997). environments may prove to be more or less general-
Research with children from Romanian orphanages ised to less extreme conditions. Obviously, much
adopted early in life into English homes, however, more work is called for before such a conclusion
did reveal enlarged amygdala volumes at adoles- could be embraced with any confidence.
cence, particularly in the right hemisphere, relative
to continuously family-reared controls, potentially
Brain function/process
underpinning disrupted affective processing (Mehta
et al., 2009). Especially interesting was that left Relatively little research has directly examined how
amygdala volume proved to be strongly associated variations in parenting, including extreme circum-
with time spent in institutions, such that the longer stances such as maltreatment, affect brain function.
the stay – and thus older age at time of adoption – Existing studies have mainly utilised electroen-
the smaller the volume. These findings, relative to cephalographic (EEG) and event-related potential
the null ones emanating from studies cited above, (ERP) approaches, though a few exceptions have
once again caution against equating the kind of used other brain imaging modalities.
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
420 Jay Belsky and Michelle de Haan

symptomatology often associated with maltreat-

EEG studies
ment.
Early work documented increased levels of EEG Quantitative EEG has also been studied in chil-
abnormalities, especially in fronto-temporal regions dren experiencing extreme neglect and deprivation in
and on the left side of the brain, in children with Romanian orphanages. Marshall, Reeb, Fox, Nelson,
histories of maltreatment (Davies, 1979; Ito et al., and Zeanah (2008) examined a group of children
1993). Subsequent studies used quantitative meth- aged 42 months who had, at a mean age of 2 years,
ods such as EEG coherence to investigate such been randomly assigned to placement in foster care
abnormalities in more detail. EEG coherence evalu- or remaining in institutions. Measures of EEG
ates the degree of synchrony between brain activity coherence were assessed in the theta, alpha and beta
recorded at two electrodes across a selected portion frequency ranges for fronto-temporal, fronto-occipi-
of the EEG frequency. One useful measure that can tal, fronto-central and fronto-parietal combinations.
be computed from such analysis is called short-dis- Coherence was lower in the right hemisphere for the
tance coherence. Decreases in this index are believed group in foster care compared with the group who
to reflect increased cortical differentiation and thus remained institutionalised, a result similar to that
increased brain complexity. Decreases in short- reported above for maltreated children, in that more
distance coherence have been related to enhanced nurturing environments were associated with lower
language functioning in normal children (Mundy coherence. To be noted, however, is that the findings
et al., 2003). A second useful measure is long-dis- which provide much stronger grounds for drawing
tance coherence, which is believed to increase with causal inferences than do those from observational
myelination and development of connections studies diverge from such prior work in that the
between distant cortical regions (Thatcher, 1992; differences in coherence were observed over the left
Thatcher, North, & Biver, 2005). MRI studies support hemisphere in the maltreatment study, but over the
the idea that normal development is characterised by right in the institutionalisation study. A further
a weakening of short-range functional connectivity important finding of Marshall et al. (2008) was that,
and strengthening of long-range functional connec- within the foster care group, lower short-distance
tivity (Supekar, Musen, & Menon, 2009). coherence in the alpha and beta ranges was related
In one early investigation, 15 hospitalised, 6–15- to earlier placement into foster care.
year-old victims of severe physical or sexual abuse In studies of EEG power, institutionalised Roma-
were compared to 15 non-abused agemates on nian children aged 5–31 months showed increased
measures of EEG coherence (Ito, Teicher, Gold, & low frequency band (theta) power, particularly
Ackerman, 1998). Abused children had higher left over posterior brain regions, and decreased high
hemisphere coherence in the alpha (high-frequency) frequency band (alpha and beta) power, particularly
range compared to non-abused controls, but there over fronto-temporal regions, compared to age-mat-
was no difference between the groups in right ched never-institutionalised Romanian children
hemisphere coherence. This result, together with (Marshall et al., 2004). Similar patterns of EEG
further analyses examining the rate of decay of power have been reported in a group of interna-
coherence over electrode distance, suggested that tionally adopted, post-institutionalised children
the cortex in the left hemisphere was less differenti- compared to age-matched never-institutionalised
ated in the abused children than in controls. The fact agemates (Tarullo, Chatham, & Gunnar, 2007). This
that findings were similar for physical and sexual result might reflect a delay in development, as alpha
abuse subgroups was regarded as evidence that the power normally increases with age, or it may reflect a
detected EEG abnormalities were not a consequence general hypo-activation of the brain, possibly due to
of direct physical injury. lack of appropriate stimulation (Marshall et al.,
Another way to quantify the EEG is to compute the 2004). Marshall and colleagues also found hemi-
power in defined frequency bandwidths. In one spheric asymmetries, with greater absolute theta
investigation of alpha-band activity in 44 maltreated and beta power over the right temporal and occipital
children and 43 non-maltreated children aged 6–12 regions compared to the left in the non-institu-
years, the former showed greater right relative to left tionalised group, but a lack of asymmetry in the
activity in parietal regions whereas the latter did not institutionalised group.
show this asymmetry (Curtis & Cicchetti, 2007). In a subsequent investigation, Marshall and
Although these results were considered consistent associates (2008) again examined EEG in the same
with other work documenting increased right hemi- children after they had been randomly assigned at a
sphere activity in individuals with PTSD, no assess- mean age of 2 years to foster care or remaining in
ment of PTSD was made in this inquiry. In addition, institutions. There were no overall differences in
maltreated females showed greater right than left power for either low or high frequency bands
hemisphere activity over frontal regions. This result between the groups. This apparently negative finding
was interpreted in line with prior work associating might have occurred because of a policy of non-
this EEG pattern with withdrawal and negative intervention, whereby after the initial random
affective style, and thus linked to the internalising assignment to foster care or institution, some infants
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 421

in the institution group were subsequently returned cues (Shackman, Shackman, & Pollak, 2007). Fur-
to their families or placed in foster care, with data thermore, the degree of attention allocated to the
analysed in accord with initial group assignment. angry faces statistically mediated the relationship
Despite the absence of between-group differences, between physical abuse and child-reported anxiety
analysis within the foster care group revealed earlier (Shackman et al., 2007).
foster-care placement to be related to increased al- Considered together, this body of work suggests
pha power. Given that Marshall and colleagues that maltreated children process facial emotion dif-
(2004) had previously found reduced alpha power in ferently when they are required to attend to anger,
younger children while they were living in institu- and that they process auditory information about
tions (see above), this increased alpha power in the anger differently whether or not they are required to
foster care group was regarded as evidence that attend to it. Behavioural research supports the
foster care had a positive influence on brain devel- interpretation that these maltreated children show
opment. Cognitive functioning was also superior for increased levels of attention and enhanced percep-
children placed in foster care at an earlier rather tion of anger. For example, in an attentional cuing
than later age, though no evidence emerged consis- task, physically abused children demonstrated
tent with the hypothesis that EEG variables medi- delayed disengagement when angry faces served as
ated this rearing effect on cognitive performance (i.e., invalid cues and increased attentional benefits on
foster-care placement fi EEG fi cognitive function- valid angry trials (Pollak & Tolley-Scholl, 2003). In a
ing). What this should make clear is that it cannot be different task in which children had to identify facial
presumed that just because a rearing experience emotion from perceptually impoverished stimuli,
proves related to a brain process and to a phenotypic abused children accurately recognised anger early in
outcome, the former effects mediate the latter. the formation of the facial expression, when few
physiological cues were available. The speed of
children’s recognition was associated with the degree
ERP studies
of anger/hostility reported by the child’s parent
Event-related potentials (ERPs) are a subset of the (Pollak, Messner, Kistler, & Cohn, 2009). Clearly,
EEG that reflect the brain’s processing of a discrete then, at the level of both brain and behaviour, mal-
event, such as the brief presentation of a visual treated children appear hyper-vigilant to cues to
image or sound. Several ERP studies have examined anger.
how maltreatment or institutionalisation affects Behavioural studies suggest that this altered pro-
children’s processing of emotions in faces. In one of cessing of anger is particularly related to experience
the first such inquiries, ERPs were recorded while 23 of physical abuse rather than other forms of abuse.
maltreated and 21 non-maltreated 9-year-olds For example, one investigation of emotion recogni-
viewed pictures of happy, angry and neutral faces tion found that physically abused children per-
after being directed to press a button in response to formed well, especially with angry expressions,
the happy face in one test session and to the angry suggesting a particular sensitivity to this expression.
face in another (Pollak, Cicchetti, Klorman, & Bru- By contrast, the neglected children had generally
maghim, 1997). As expected from prior work, the difficulty in differentiating facial expressions of
P300 component was larger for target than non-tar- emotion (Pollak, Cicchetti, Hornung, & Reed, 2000).
get faces. However, for non-maltreated children the These data suggest that specific kinds of experi-
size of the P300 was the same whether happy or ences, rather than simply the general presence of
angry was the target; by contrast, for maltreated stress or maltreatment, have differential effects.
children the P300 was larger when angry was the Recall, however, that the ERP studies considered
target than when happy was the target. A direct earlier did not find such a differentiation between
comparison between the groups showed that the neglected and physically abused children in terms of
P300 did not so much differ for angry targets but response to anger, possibly because of a lack of
proved smaller in the maltreated group than the statistical power to do so (Pollak et al., 2001).
non-maltreated group for happy targets. ERPs in response to facial expressions of emotion
A subsequent study used a similar task and age have also been assessed in institutionalised children
group but with happy, fearful and angry faces where who have experienced extreme neglect and depriva-
each emotion served as target and non-target equally tion. In one such study, ERPs to happy, sad, angry
often across a set of three test sessions (Pollak, and fearful faces were recorded in 5- to 31-month-
Klorman, Thatcher, & Cicchetti, 2001). In this work, old institutionalised Romanian children and a simi-
the P300 was larger for the angry target face in larly aged group of Romanian children who had
maltreated compared to non-maltreated children but never been institutionalised (Parker et al., 2005a,
there were no differences between the groups for the 2005b). Institutionalised children showed overall
happy or fearful targets. A third ERP inquiry gener- decreased amplitudes of the N170 Nc and PSW
ated evidence that physically abused children pay responses and, unlike controls, did not show a
more attention than normal to task-relevant visual decrease in Nc amplitude with age. These results
or auditory anger cues and task-irrelevant auditory appear consistent with the notion that institutiona-
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
422 Jay Belsky and Michelle de Haan

lised children show brain hypoactivation, as well as N170, Nc and PSW components and increases in
a distinctive trajectory of brain development. Insti- their P250 component compared to never-institu-
tutionalised children also showed an absence of tionalised group. Both groups of children had a
hemispheric lateralisation for the positive slow wave, normative increase in the Nc amplitude to the face of
whereas controls showed a larger response over the a stranger compared to the caregiver’s face. In con-
right hemisphere. Emotion did not influence the Nc trast, the institutionalised children displayed an
or PSW components but did influence two early abnormal pattern for the PSW component, in which
components believed to reflect perceptual process- they showed a tendency towards a decrease in
ing, a frontal ‘N170’ component (believed not to be amplitude to the stranger’s face compared to the
the face-sensitive N170 component because of its familiar caregiver’s face while the never-institu-
anterior topography) and a posterior ‘P250’ compo- tionalised children showed an opposite pattern.
nent. The control group showed the largest N170 A subsequent study comparing institutionalised
response to sad and smallest to fear, and the largest and never-institutionalised children aged 5–31
P250 response to fear and smallest to happy. The months found only a reduction in amplitude of the
institutionalised group manifest a roughly opposite P1 component, but not other components, in the
pattern, with the largest N170 response to fear and institutionalised group (Moulson, Westerlund, Fox,
smallest to happy and sad, and largest N250 Zeanah, & Nelson, 2009a), and the only evidence of
response to happy and sad and smallest to angry. recognition of the familiar face was a faster latency of
The authors speculated that these results might the P400 to the caregiver’s face in the younger (<21
reflect an overactivity of the response of the months) but not the older (>21 months) never-insti-
amygdala to fearful expressions. tutionalised children. Children were reassessed at
Subsequent work, however, failed to replicate 30 and 42 months, by which time some of the
these findings, clearly casting doubt on this inter- institutionalised children had been randomly
pretation, as no differences were detected in the ERP assigned to foster care. Results showed that by 42
correlates of emotion processing in institutionalised months the P1 amplitude was still larger in the
compared to never-institutionalised children (Moul- never-institutionalised compared to institutionalised
son, Fox, Zeanah, & Nelson, 2009b). Nevertheless, children, with the foster care group in between. All
Moulson et al. (2009b) did present evidence of hyp- groups showed evidence of face recognition at 30 and
oarousal in institutionalised children who displayed 42 months, with larger and faster Nc’s to the
smaller amplitudes for the P1 bilaterally and the strangers’ faces compared to caregivers’ faces.
N170 and P400 components over the right hemi- Overall, the EEG and ERP research reviewed
sphere compared to the never-institutionalised suggest that early adverse experience can affect
group. Latencies for the P1 and P400 components brain function. Studies of maltreated children pro-
were also longer among the institutionalised children vide evidence of altered EEG coherence and cerebral
relative to the never-institutionalised children, thus asymmetry; studies of institutionalised children also
chronicling slower information processing in chil- provide evidence of altered EEG coherence and
dren with a history of institutionalisation (Moulson cerebral asymmetry as well as reduced-amplitude
et al. 2009b). components, slower information processing and
After these initial baseline measures were col- altered cerebral asymmetry. Reductions in compo-
lected, the institutionalised children, as previously nent amplitude and alterations in EEG power
noted, were randomly assigned to be placed into observed in institutionalised children may reflect
foster care or to remain in the institution. At 30 and delayed development or general brain hypo-activa-
42 months of age, ERPs were once again collected tion. Findings of atypical cerebral asymmetry and
during presentation of emotion face stimuli. Results altered EEG coherence in both groups may reflect
revealed that P1 and P400 amplitudes of children in influences of adverse experience on brain differen-
the foster care group were midway between the tiation, though this interpretation remains to be fully
institutionalised and never-institutionalised chil- investigated. With respect to emotion processing,
dren, suggesting that foster care may be normalising studies of abused children provide consistent evi-
these components. Interestingly, neither age at dence of atypical processing of anger, but studies of
placement in foster care nor duration of time there institutionalised infants and toddlers do not provide
contributed to this outcome. consistent evidence that brain processing of facial
ERPs in institutionalised children also have been emotion or facial identity per se is atypical. A recent
recorded during face identity recognition tasks. In neuropsychological study of 8–9-year-old children
one such study, ERPs were measured while institu- adopted from institutions when they were 12
tionalised Romanian children (ranging from 5 to 31 months or older did find evidence of a mild (less than
months in age) or never-institutionalised Romanian one standard deviation from the mean) but statisti-
children viewed pictures of either their caregiver’s cally significant impairment in face learning/recog-
face (mother or preferred institutional caregiver) or a nition in this group (Pollak et al., 2010), perhaps
stranger’s face (Parker et al., 2005a). The institu- suggesting that such deficits become more marked
tionalised children displayed attenuation in their with age.
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 423

ment and institutionalisation and how they relate to

MRI studies
the developing brain. Overall, the findings of this
MRI is a noninvasive method that can provide high- work suggest that these parenting conditions of
spatial-resolution structural ( 1mm3) images of the extreme adversity, including institutionalisation, are
brain (or other body parts). Analysis of such images associated with reductions in cortical grey and white
allows segregation of grey and white matter struc- matter volumes and measures of white matter
tures and measurement of their characteristics such microstructure. Findings for subcortical structures
as volume or thickness, with recent variants of are more complex and varied, with the hippocampus
conventional MRI such as diffusion tensor imaging showing effects of early adversity only later in
(DTI) allowing even more detailed visualisation of development, beginning from the transition into
white matter. Functional magnetic resonance adulthood from adolescence, and the amygdala often
imaging allows indirect measurement of brain acti- (but not always) showing no effects. It is possible that
vation by measuring changes in brain activation. subcortical structures such as the amygdala are less
Compared to ERP measures described above, it influenced by environmental variations than the
provides more detailed spatial information about the cortex or the hippocampus, which is sometimes
regions activated, but more limited temporal reso- described as a ‘primitive cortex’. This is certainly not
lution. a conclusion that could be confidently embraced at
One fMRI study has examined longer-term rela- this point, however.
tions between early parenting experience and later With respect to function, most studies to date have
emotion processing, at least as revealed by research employed electrophysiological techniques which are
in which childhood experience is reported retro- thought to measure primarily cortical activity and
spectively in adulthood (i.e., non-longitudinal study). which do not afford conclusive identification of the
Thus, this work examined how the quality of family specific sources underlying recorded activity. This
life in childhood seemed to affect adults’ brain body of work suggests that early institutionalisation,
responses to facial emotions by comparing adults regarded here as a marker of extreme parental
who grew up in risky families characterised by harsh neglect and deprivation, is related to general brain
parenting to those who grew up in more loving, ‘non- hypo-activation, but not necessarily specific diffi-
risky’ environments (Taylor, Eisenberger, Saxbe, culties with emotion perception as assessed by face
Lehman, & Lieberman, 2006). Adults who grew up in processing. In contrast, child maltreatment, in par-
non-risky families showed amygdala activation while ticular physical abuse, is associated with enhanced
observing fearful/angry faces and activation of right neural processing of anger/threat-related stimuli.
ventral-lateral prefrontal cortex that negatively cor- Future studies using fMRI or in which EEG mea-
related with amgydala activation during emotion sures are combined with MRI measures can provide
labelling. By contrast, adults who grew up in risky more detailed information about how such func-
families showed little amygdala activation during tional abnormalities are related to structural
passive observation and a strong positive correlation abnormalities.
between right ventral-lateral prefrontal and amyg- It is not surprising that the first research exam-
dala activation during labelling. The authors con- ining parenting effects on brain development has
cluded that harsh parenting affects processing of focused on extremely adverse experiences, namely,
threat stimuli, such that there is a tuning out or lack maltreatment and institutionalisation. After all, if
of response to such stimuli when viewed passively, evidence suggestive of parenting effects on brain
and an atypical activation when forced to view them. structure and function did not emerge when expo-
Questions can be raised, however, about the source sure to such adversity was compared to parenting in
of the effect in question, as more than just harsh the normal range, one would be hard pressed to
parenting was used to define the concept of a risky expect variation within the latter to make much
family environment. difference to children’s brain development, as is
often assumed to be the case. But now that so much
progress has been made in the first stage of inquiry,
Summary, conclusions and future directions
it seems appropriate to conclude that research in
Parenting is determined by multiple forces emanat- this area of inquiry has reached ‘the end of the
ing from the child, the parent, and the social context, beginning’.
forces that both combine and interact to amplify and Work is now needed to determine whether and how
buffer each other’s effects (Belsky & Jaffee, 2006). variation in parenting in the normal range affects the
Although there is a large and ever-growing literature brain development of children not exposed to
examining how parenting influences children’s extreme adversity. Certainly Whittle et al.’s (2009)
behavioural development, relatively little is known recent effort in this regard suggests that such
about how parenting influences children’s brain enterprises should prove fruitful, thereby allowing
functioning and development. As a result, this review greater generalisation of findings and conclusions.
has, by necessity, drawn heavily, though not quite Until more work of this kind is carried out, it will be
exclusively, on research focusing on child maltreat- impossible to know whether evidence emerging from
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
424 Jay Belsky and Michelle de Haan

studies of extreme adversity apply to most parents allele had the lowest levels of hippocampal and
and children. The kind of work being called for would amygdala grey matter volume if they had been ex-
do well to move beyond just measuring constructs posed to high levels of stress, but the highest levels if
like ‘harsh parenting’ to ensure that a wide range of they had not been exposed, whereas those carrying
potential influential features of parenting are exam- only Val alleles (Val/Val) evinced no effect of early life
ined. It is certainly imaginable that effects on the stress whatsoever.
brain of warmth, harshness, sensitivity, respon- Noteworthy is the fact that the investigators failed
siveness, monitoring, to name just a few well-studied to call attention to the ‘for better and for worse’
parameters of parenting, will not all be the same. patterning of their results, just like many others
But as investigators turn, hopefully, to investi- studying GXE interactions (Belsky & Pluess, 2009b),
gating parenting and child development in the nor- thus highlighting only genetic vulnerability – in the
mal range, it would be most useful to move beyond face of adversity – of those carrying Met alleles.
observational studies which yield correlational find- Investigators in the future would do well to remain
ings. In view of the fact that a great deal of inter- open to the possibility that it is not just the case that
vention research demonstrates that parenting can be some children, for genetic or other organismic rea-
systematically manipulated, with discernible effects sons (e.g., temperament), are more susceptible to the
on child behaviour, the time would seem to have negative effects of adversity than others, but that
come to add brain measurements to such study those most vulnerable to adversity may also be those
designs. It seems critically important that Andersen most likely to benefit from supportive and enriching
and colleagues (2008) have moved in this direction. environmental conditions. This proposition could
Not only does such work afford stronger causal certainly be tested vis-à-vis parenting effects on the
inferences about effects of parenting on brain developing brain by means of experimental inter-
structure and function – that is, evidence that vention to enhance parenting.
experience ‘chisels’ the brain (Choi et al., 2009, The final point to be made concerns the perhaps
p. 233) – but it enables investigators to extend such restricted focus of this review on effects of parenting
links to behaviour. on children’s brain development, without regard to
As noted in the introduction, without (eventual) reciprocal processes whereby the experience of par-
evidence that discerned effects of parenting on brain enting influences the brain functioning of adults.
structure and function come to influence actual Work on the neural bases of maternal behaviour in
behavioural functioning, there would seem to be humans, in which mothers, non-mothers, and
grounds for questioning the significance of this sometimes fathers are presented with pictures of
entire field of study. Until such evidence emerges, their own infants or same-aged unfamiliar infants
the fundamental proposition that effects of parenting (Bartels & Zeki, 2004; Leibenluft, Gobbini, Harrison,
– or other environmental factors for that matter – on & Haxby, 2004; Nitschke et al., 2004), recorded
the brain illuminate how development operates will infant cries (Lorberbaum et al., 2002; Seifritz et al.,
remain just a hypothesis. In point of fact, even 2003), or videotapes of infants (Ranote et al., 2004),
observational-correlational studies linking parenting demonstrate that many of the same hypothalamic,
with brain measurements would do well to include limbic and coritical sites important for emotional or
behavioural ‘outcome’ measurements, broadly con- social (face) processing or for regulation of maternal
ceived, as some of the reviewed studies have done, so behaviour in other mammals are implicated in
that putative pathways of influence from parenting response to infant stimuli. When cast in individual
to brain development to behaviour could be exam- difference terms, such findings highlight the possi-
ined, even if only correlationally (e.g., path analysis). bility that variation in the parenting experience,
Future research should also consider the very real perhaps due to variation in child behaviour, could
possibility that parenting – and other environmental affect parents’ brain functioning. Quite conceivably,
factors – does not affect the brains of all children then, when the next phase of parenting–brain-
equally and thus that there is differential suscepti- development research is reviewed, scholars will be in
bility to environmental influences (Belsky & Pluess, a position to delineate not only effects of parenting in
2009a, 2009b; Boyce & Ellis, 2005; Obradović & the normal range on children’s brain development,
Boyce, 2009). Intriguingly, Gatt and associates but how variation in the parenting experience influ-
(2009), in studying the interaction of the brain- ences the brains of adults doing the parenting.
derived neurotrophic factor (BDNF) Val66Met poly-
morphism and early life stress (including ‘abuse,
neglect, family conflict’) in predicting brain and
Correspondence to
arousal pathways to syndromal depression and
anxiety, detected evidence of such. BDNF is a Jay Belsky, Institute for the Study of Children,
member of the nerve growth factor family and is Families and Social Issues, Birkbeck University of
involved in promoting neuronal differentiation, London, 7 Bedford Square, London WC1B 3RA, UK;
synaptic connectivity, and neuronal repair (Vinberg Tel: +44 (0)20-7079-0835; Fax: +44 (0)20-7323-
et al., 2009). Individuals carrying at least one Met 4735; Email: [email protected]
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 425

Key points

• Brain structures develop at different rates, so they are differentially sensitive to developmental experience,
including parenting, at different points in time.
• Most studies illuminating putative parenting ‘effects’ on brain structure and function are of children
exposed to child maltreatment and severe deprivation in institutions. Evidence clearly suggests that both
brain structure and functioning can be adversely affected by such experience.
• Experimental interventions will afford stronger causal inference than observational studies. Research on
parenting ‘in the normal range’ is needed.
• Work that incorporates phenotypic outcomes would afford evaluation of whether parenting effects on the
developing brain mediate parenting effects on behaviour, cognition, emotion and health.
• Consideration should be given to the hypothesis that children vary in their susceptibility to parenting
effects on the developing brain.

References genes or plasticity genes? Molecular Psychiatry, 14, 746–

754.
Amaral, D.G., & Bennett, J. (2000). Development of amy- Belsky, J., & Pluess, M. (2009a). The nature (and nurture?)
gdalo-cortical connections in the macaque monkey. of plasticity in early human development. Perspectives
Society for Neuroscience Abstracts, 26, 17–26. on Psychological Science, 4, 345–351.
Andersen, S.L., & Teicher, M.H. (2004). Delayed effects of Belsky, J., & Pluess, M. (2009b). Beyond diathesis-stress:
early stress on hippocampal development. Neuropsycho- Differential susceptibility to environmental influences.
pharmacology, 29, 1988–1993. Psychological Bulletin, 135, 885–908.
Andersen, S.L., & Teicher, M.H. (2009). Desperately driven Bengtsson, S.I., Nagy, Z, Skare, S., Forsman, L., Forssberg,
and no brakes: Developmental stress exposure and H., & Ullen, F. (2005). Extensive piano practice has
subsequent risk for substance abuse. Neuroscience and regionally specific effects on white matter development.
Biobehavioral Reviews, 33, 516–524. Nature Neuroscience, 8, 1148–1150.
Andersen, S.L., Tomada, A., Vincow, E., Valente, E., Berrebi, A.S., Fitch, R.H., Ralphe, D.L., et al. (1988).
Polcari, A., & Teicher, M.H. (2008). Preliminary evidence Corpus callosum: Region- specific effects of sex, early
for sensitive periods in the effect of childhood sexual experience and age. Brain Research, 438, 216–224.
abuse on regional brain development. Journal of Neuro- Black, J., Jones, T.A., Nelson, C.A., & Greenough, W.T.
psychiatry and Clinical Neurosciences, 20, 292–301. (1998). Neuronal plasticity and the developing brain. In
Bartels, A., & Zeki, S. (2004). The neural correlates of N.E. Aless, J.T. Coyle, S.I. Harrison, & S. Eth (Eds.),
maternal and romantic love. NeuroImage, 21, 1155–1166. Handbook of child and adolescent psychiatry (pp. 31–
Bauer, P.M., Hanson, J., Pierson, R.K., Davidson, R.J., & 53). New York: Wiley.
Pollak, S.D. (2009). Cerebellar volume and cognitive Boyce, W.T., & Ellis, B.J. (2005). Biological sensitivity to
functioning in children who experienced early depriva- context: I. An evolutionary-developmental theory of the
tion. Biological Psychiatry, 66, 1100–1106. origins and functions of stress reactivity. Development
Beers, S.R., & De Bellis, M.D. (2002). Neuropsychological and Psychopathology, 17, 271–301.
function in children with maltreatment-related posttrau- Bremner, J.D., Randall, P., Vermetten, E., Staib, L.,
matic stress disorder. American Journal of Psychiatry, Bronen, R.A., Mazure, C., et al. (1997). Magnetic reso-
159, 483–486. nance imaging-based measurement of hippocampal vol-
Belsky, J. (1997). Variation in susceptibility to rearing ume in posttraumatic stress disorder related to
influences: An evolutionary argument. Psychological childhood physical and sexual abuse – a preliminary
Inquiry, 8, 182–186. report. Biological Psychiatry, 41, 23–32.
Belsky, J. (2000). Conditional and alternative reproductive Brody, B.A., Kinney, H., Kloman, A., & Gilles, F.H. (1987).
strategies: Individual differences in susceptibility to Sequence of central nervous system myelination in
rearing experience. In J. Rodgers, D. Rowe, & W. Miller human infancy: I An autopsy study of myelination.
(Eds.), Genetic influences on human fertility and sexual- Journal of Neuropathology and Experimental Neurology,
ity: Theoretical and empirical contributions from the 46, 283–301.
biological and behavioral sciences (pp. 127–146). Boston: Buss, C., Lord, C., Wadiwalla, M., Hellhammer, D.H.,
Kluwer. Lupien, S.J., Meaney, M.J., & Pruessner, J.C. (2007).
Belsky, J. (2005). Differential susceptibility to rearing Maternal care modulates the relationship between
influences: An evolutionary hypothesis and some evi- prenatal risk and hippocampal volume in women but
dence. In B. Ellis, & D. Bjorklund (Eds.), Origins of the not in men. The Journal of Neuroscience, 27, 2592–
social mind: Evolutionary psychology and child develop- 2595.
ment (pp. 139–163). New York: Guilford Press. Carrion, V.G., Weems, C.F., Eliaz, S., Patwardhan, A.,
Belsky, J., Bakermans-Kranenburg, M.J., & van IJzendo- Brown, W., Ray, R.D., et al. (2001). Attenuation of frontal
orn, M.H. (2007). For better and for worse: Differential asymmetry in pediatric posttraumatic stress disorder.
susceptibility to environmental influences. Current Direc- Biological Psychiatry, 50, 943–951.
tions in Psychological Science, 16, 300–304. Carrion, V.G., Weems, C.F., & Reiss, A.L. (2007). Stress
Belsky, J., & Jaffee, S. (2006). The multiple determinants predicts brain changes in children. A pilot longitudinal
of parenting. In D. Cicchetti, & D. Cohen (Eds.), Devel- study on youth stress, posttraumatic stress disorder and
opmental psychopathology, 2nd edn, Vol. 3: Risk, disor- the hippocampus. Pediatrics, 119, 509–516.
der and adaptation (pp. 38–85). New York: Wiley. Catani, M. (2006). Diffusion tensor magnetic resonance
Belsky, J., Jonassaint, C.R., Pluess, M., Stanton, M., imaging tractography in cognitive disorders. Current
Brummett, B.H., & Williams, R.B. (2009). Vulnerability Opinion in Neurology, 19, 599–606.

 2010 The Authors

Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
426 Jay Belsky and Michelle de Haan

Choi, J., Bumseok, J., Rohan, M.L., Polcari, A., & Teicher, Dynamic mapping of normal hippocampal development.
M.H. (2009). Preliminary evidence for white matter tract Hippocampus, 16, 664–672.
abnormalities in young adults exposed to parental verbal Gould, E., & Tanapat, P. (1999). Stress and hippocampal
abuse. Biological Psychiatry, 65, 227–234. neurogenesis. Biological Psychiatry, 46, 1472–1479.
Cicchetti, D. (2002). The impact of social experience on Greenough, W., Black, J., & Wallace, C. (1987). Experience
neurobiological systems. Cognitive Development, 17, and brain development. Child Development, 58, 539–
1407–1428. 559.
Courchesne, E., Chisum, H., & Townsend, J. (1994). Hasan, K.M., Kamali, A., Iftikhar, A., Kramer, L.A., Papa-
Neural activity-dependent brain changes in develop- nicolaou, A.C., Fletcher, J.M., & Ewing-Cobbs, L. (2009).
ment: Implications for psychopathology. Development Diffusion tensor tractography quantification of the hu-
and Psychopathology, 6, 697–722. man corpus callosum fiber pathways across the lifespan.
Curtis, W.J., & Cicchetti, D. (2007). Emotion and resil- Brain Research, 1249, 91–100.
ience: A multilevel investigation of hemispheric electro- Hua, X., Leow, A.D., Levitt, J.G., Caplan, R., Thompson,
encephalogram asymmetry and emotion regulation in P.M., & Toga, A.W. (2009). Detecting brain growth
maltreated and nonmaltreated children. Development patterns in normal children using tensor-based mor-
and Psychopathology, 19, 811–840. phometry. Human Brain Mapping, 30, 209–219.
Davies, R.K. (1979). Incest: Some neuropsychiatric find- Huttenlocher, P.R. (1979). Synaptic density of human
ings. International Journal of Psychiatry Medicine, 9, frontal cortex. Developmental changes and effects of
117–121. aging. Brain Research, 163, 195–205.
De Bellis, M.D., Baum, A., Birmaher, B., Keshavan, M., Huttenlocher, P.R. (1990). Morphometric study of human
Eccard, C.H., Boring, A.M., et al. (1999a). A.E. Bennett cerebral cortex development. Neuropscyhologica, 28,
Research Award, Developmental traumatology part I: 517–527.
Biological stress systems. Biological Psychiatry, 45, Huttenlocher, P.R. (1994). Synaptogenesis, synapse elim-
1259–1270. ination, and neural plasticity in the human cerebral
De Bellis, M.D., Keshavan, M.S., Clark, D.B., et al. (1999b). cortex. In C.A. Nelson (Ed.), Threats to optimal develop-
Developmental traumatology, part II: Brain develop- ment. The Minnesota Symposia on Child Psychology (vol.
ment. Biological Psychiatry, 45, 1271–1284. 27, pp. 35–54). Hillsdale, NJ: Lawrence Erlbaum Asso-
De Bellis., M., & Keshavan, M. (2003). Sex differences in ciates.
brain maturation in maltreatment-related pediatric post- Huttenlocher, P.R., & Dabholkar, A.S. (1997). Regional
traumatic stress disorder. Neuroscience and Biobehav- differences in synaptogenesis in human cerebral cortex.
ioral Reviews, 27, 103–117. Journal of Comparative Neurology, 20, 167–178.
De Bellis, M., Keshavan, M., Shifflett, H., Iyengar, Beers, Ito, Y., Teicher, M., Gold, C.A., & Ackerman, E. (1998).
S., Hall, J., & Moritz, G. (2002). Brain structures in Preliminary evidence for aberrant cortical development
pediatric maltreatment-related posttraumatic stress dis- in abused children: A quantitative EEG study. The
order: A sociodemographically matched study. Biological Journal of Neuropsychiatry and Clinical Neurosciences,
Psychiatry, 52, 1066–1078. 10, 298–307.
Edelman, G.M. (1987). Neural Darwinism: The theory of Ito, Y., Teicher, M., Gold, C.A., Harper, D., Magnus, E., &
neuronal group selection. New York: Basic Books. Gelbard, H.A. (1993). Increased prevalence of
Eluvathingal, T.J., Chugani, H.T., Behen, M.E., Juhasz, electrophysiological abnormalities in children with psy-
C., Muzik, O., Maqbool, M., Chugani, D.C., & Makki, M. chological, physical and sexual abuse. Journal of Neuro-
(2006). Abnormal brain connectivity in children after psychiatry, 5, 401–408.
early severe socioemotional deprivation: A diffusion Juraska, J.J., & Kopcik, J.R. (1988). Sex and environmen-
tensor imaging study. Pediatrics, 117, 2093–1326. tal influences on the size and ultrastructure of the rat
Gatt, J.M., Nemeroff, C.B., Dobson-Stone, C., Paul, R.H., corpus callosum. Brain Research, 450, 1–8.
Bryant, R.A., Schofield, P.R., Gordon, E., Kemp, A.H., & Keshavan, M.S., Diwadkar, V.A., De Bellis, M., Dick, E.,
Williams, L.M. (2009). Interactions between BDNF Kotwal, R., Rosenberg, D.R., et al. (2002). Development
Val66Met polymorphism and early life stress predict of the corpus callosum in childhood, adolescence and
brain and arousal pathways to syndromal depression early adulthood. Life Science, 70, 1909–1922.
and anxiety. Molecular Psychiatry, 14, 1–15. Kertes, D.A., Donzella, B., Talge, N.M., Garvin, M.C., Van
Giedd, J.N. (2004). Structural magnetic resonance imaging Ryzin, M.J, & Gunnar, M.R. (2009). Inhibited tempera-
of the adolescent brain. Annals of the New York Academy ment and parent emotional availability differentially
of Sciences, 1021, 77–85. predict young children’s cortisol responses to novel
Giedd, J., Blumenthal, J., Jeffries, N.O., Castellanos, F.X., social and nonsocial events. Developmental Psychobiol-
Liu, H., Zijdenbs, A., Paus, T., Evans, A.C., & Rapaport, ogy, 51, 421–532.
J.L. (1999). Brain development during childhood and Kinney, H.C., Brody, B.A., Kloman, A., & Gilles, F. (1988).
adolescence: A longitudinal MRI study. Nature Neurosci- Sequence of CNS myelination in infancy. Journal of
ence, 2, 861–863. Neuropathology and Experimental Neurology, 47, 217–
Giedd, J.N., Vaituzis, A.C., Hamburger, S.D., et al. (1996). 234.
Quantitative MRI of the temporal lobe, amygdale, and Knickmeyer, R.C., Gouttard, S., Kang, C., Evans, D.,
hippocampus in normal human development: Ages 4–18 Wilber, K., Smith, J.K., Hamer, R.M., Lin, W., Gerig,
years. Journal of Comparative Neurology, 366, 223–230. G., & Gilmore, J.H. (2008). A structural MRI study of
Gogtay, N., Giedd, J.N., Lusk, L., Hayashi, K.M., Green- human brain development from birth to 2 years. Journal
stein, D., Vaituzis, A.C., Nugent, T.F. 3rd, Herman, D.H., of Neuroscience, 28, 12176–12182.
Clasen, L.S., Toga, A.W., Rapoport, J.L., & Thompson, Leibenluft, E., Gobbini, M.I., Harrison, T., & Haxby, J.V.
P.M. (2004). Dynamic mapping of human cortical devel- (2004). Mothers’ neural activation in response to pic-
opment during childhood through early adulthood. Pro- tures of their children and other children. Biological
ceedings of the National Academy of Sciences USA, 101, Psychiatry, 56, 225–2323.
8174–8179. Lenroot, R.K., Schmitt, J.E., Ordaz, S.J., Wallace, G.L.,
Gogtay, N., Nugent, T.F. 3rd, Herman, D.H., Ordonez, A., Neale, M.C., Lerch, J.P., et al. (2009). Differences in
Greenstein, D., Hayashi, K.M., Clasen, L., Toga, A.W., genetic and environmental influences on the human
Giedd, J.N., Rapoport, J.L., & Thompson, P.M. (2006). cerebral cortex associated with development during

 2010 The Authors

Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
 Parenting and children’s brain development 427

childhood and adolescence. Human Brain Mapping, 30, Nitschke, J.B., Nelson, E.E., Rusch, B.D., Fox, A.S.,
163–174. Oakes, T.R., & Davidson, R.J. (2004). Orbitofrontal
Levine, S. (1967). Maternal and environmental influences cortex tracks positive mood in others viewing pictures
on the adrenocortical response to stress in weanling rats. of their newborn infants. NeuroImage, 21, 583–592.
Science, 156, 258–260. O’Leary, D. (1989). Do cortical areas emerge from a
Lorberbaum, J.P., Newman, J.D., Horwitz, A.R., Dubno, protocortex? Trends in Neurosciences, 12, 400–406.
J.R., Lydiard, R.B., Hamner, M.B., Bohning, D.E., & Obradović, J., & Boyce, W.T. (2009). Individual differences
George, M.S. (2002). The potential role of thalamocingu- in behavioral, physiological, and genetic sensitivities to
late circuitry in human maternal behavior. Biological contexts: Implications for development and adaptation.
Psychiatry, 51, 431–445. Developmental Science, 31, 300–308.
Luu, P., & Tucker, D. (1996). Self-regulation and cortical Ostby, Y., Tamnes, C., Fjell, A.M., Westlye, L.T., Due-
development: Implications for functional studies of the Tonnessen, P., & Walhovd, K.B. (2009). Heterogeneity
brain. In R.W. Thatcher, G.R. Lyon, J. Rumsey, & N. in subcortical brain development: A structural mag-
Krasnegor (Eds.), Developmental neuroimaging: Mapping netic resonance imaging study of brain maturation
the development of brain-behaviour (pp. 298–305). San from 8 to 30 years. Journal of Neuroscience, 29,
Diego, CA; Academic Press. 11772–11782.
Machado, C.J., & Bachevalier, J. (2003). Non-human Pakkenberg, B., & Gundersen, H.J. (1997). Neocortical
primate models of childhood psychopathology: The neuron number in humans: Effects of sex and age.
promise and limitations. Journal of Child Psychology Journal of Comparative Neurology, 384, 312–320.
and Psychiatry, 44, 64–87. Parker, S.W., Nelson, C.A., & Bucharest Early Intervention
Maestripieri, D., Lindell, S.G., Ayala, A., Gold, P.W., & Project Core Group. (2005a). An event-related potential
Higley, J.D. (2005). Neurobiological characteristics of study of the impact of institutional rearing on face
rhesus macaque abusive mothers and their relation to recognition. Development and Psychopathology, 17,
social and maternal behaviour. Neuroscience and 621–639.
Biobehavavioral Review, 29, 51–57. Parker, S.W., Nelson, C.A., & Bucharest Early Intervention
Markham, J.A., & Greenough, W.T. (2004). Experience- Project Core Group. (2005b). The impact of early insti-
driven brain plasticity: Beyond the synapse. Neuron Glia tutional rearing on the ability to discriminate facial
Biology, 1, 351–363. expressions of emotion: An event–related potential
Markham, J.A., Herting, M.M., Luszpak, A.E., Juraska, study. Child Development, 76, 54–72.
J.M., & Greenough, W.T. (2009). Myelination of the Payne, C., & Bachevalier, J. (2009). Neuroanatomy of the
corpus callosum in male and female rats following developing social brain. In M. de Haan, & M.R. Gunnar
complex environment housing during adulthood. Brain (Eds.), The handbook of developmental social neurosci-
Research, 1288, 9–17. ence (pp. 38–59). New York: Guilford Press.
Marshall, P.J., Fox, N.A., & Bucharest Early Intervention Perez, C., & Widom, C.S. (1994). Childhood victimization
Project Core Group. (2004). A comparison of the electro- and longterm intellectual and academic outcomes. Child
encephalogram between institutionalized and commu- Abuse and Neglect, 18, 617–633.
nity children in Romania. Journal of Cognitive Pianta, R.C., & Egeland, B. (1994). Predictors of instability
Neuroscience, 16, 1327–1338. in children’s mental test performance at 24, 48 and 96
Marshall, P.J., Reeb, B.C., Fox, N.A., Nelson, C.A. 3rd, & months. Intelligence, 18, 145–163.
Zeanah, C.H. (2008). Effects of early intervention on EEG Plomin, R., DeFries, J.C., McClearn, G.E., & McGuffin, P.
power and coherence in previously institutionalized (2008). Behavior genetics (5th edn). New York: Freeman.
children in Romania. Development and Psychopathology, Pollak, S.D., Cicchetti, D., Hornung, K., & Reed, A. (2000).
20, 861–880. Recognizing emotion in faces: Developmental effects of
Mehta, M.A., Golembo, N.I., Nosarti, C., Colvert, E., Mota, child abuse and neglect. Developmental Psychology, 36,
A., Williams, S.C.R., Rutter, M., & Sonuga-Barke, E.J. 679–688.
(2009). Amydala, hippocampal and corpus callosum Pollak, S.D., Cicchetti, D., Klorman, R., & Brumaghim, J.T.
size following severe early institutional deprivation. (1997). Cognitive brain event-related potentials and
Journal of Child Psychology and Psychiatry, 50, 943– emotion processing in maltreated children. Child Devel-
951. opment, 68, 773–787.
Milgrom, J., Newham, C., Anderson, P.J., Doyle, L.W., Pollak, S.D., Klorman, R., Thatcher, J.E., & Cicchetti, D.
Gemmill, A.W., Lee, K., Hunt, R.W., Bear, M., & Inder, T. (2001). P3b reflects maltreated children’s reactions to
(2009). Early sensitivity training for parents of preterm facial displays of emotion. Psychophysiology, 38, 267–
infants: Impact on the developing brain. Pediatric 274.
Research. Pollak, S.D., Messner, M., Kistler, D.J., & Cohn, J.F.
Moradi, A.R., Doost, H.T.N., Taghavi, M.R., Yule, W., & (2009). Development of perceptual expertise in emotion
Dalgleish, T. (1999). Everyday memory deficits in chil- recognition. Cognition, 110, 242–247.
dren and adolescents with PTSD: Performance on the Pollak, S.D., Nelson, C.A., Schlaak, M.F., Roeber, B.J.,
Rivermead Behavioural Memory Test. Journal of Child Wewerka, S.S., Wiik, K.L., Freen, K.A., Loman, M.M., &
Psychology and Psychiatry, 40, 357–361. Gunnar, M.R. (2010). Neurodevelopmental effects of
Moulson, M.C., Fox, N.A., Zeanah, C.H., & Nelson, C.A. early deprivation in postinstitutionalized children. Child
(2009b). Early adverse experiences and the neurobiology Development, 81, 224–236.
of facial emotion processing. Developmental Psychology, Pollak, S.D., & Tolley-Scholl, S.A. (2003). Selective atten-
45, 17–30. tion to facial emotion in physically abused children.
Moulson, M.C., Westerlund, A., Fox, N.A., Zeanah, C.H., & Journal of Abnormal Psychology, 112, 323–338.
Nelson, C.A. (2009a). The effects of early experience on Rakic, P. (1988). Specification of cerebral cortex areas.
face recognition: An event related potential study of Science, 241, 170–176.
institutionalized children in Romania. Child Develop- Ranote, S., Elliott, R., Abel, K.M., Mitchell, R., Deakin, J.F.,
ment, 80, 1039–1056. & Appleby, L. (2004). The neural basis of maternal
Mundy, P., Fox, N., & Card, J. (2003). EEG coherence, joint responsiveness to infants. NeuroReport, 15, 1825–1829.
attention and language development in the second year. Rao, H., Betancourt, L., Giannetta, J.M., Brodksy, N.L.,
Developmental Science, 6, 48–54. Kroczykowski, M., Avants, B.B., Gee, J.C., Wang, J.,

 2010 The Authors

Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.
428 Jay Belsky and Michelle de Haan

Hurty, H., Detre, J.A., & Farah, M.J. (2010a). Early associated with reduced corpus callosum area. Biological
parental care is important for hippocampal maturation: Psychiatry, 56, 80–85.
Evidence from brain morphology in humans. NeuroIm- Teicher, M.H., Ito, Y., Glod, C.A., Andersen, S.L., Dumont,
age, 49, 1144–1150. N., & Ackerman, E. (1997). Preliminary evidence for
Rao, U., Chen, L., Bidesi, A., Shad, M., Thomas, M.A., & abnormal cortical development in physically and sexu-
Hammen, C.L. (2010b). Hippocampal changes associ- ally abused children using EEG coherence and MRI.
ated with early-life adversity and vulnerability to depres- Annals of the New York Academy of Sciences, 821, 160–
sion. Biological Psychiatry, 67, 357–364. 175.
Rodman, H (1994). Development of inferior temporal cortex Teicher, M.H., Tomoda, A., & Andersen, S.L. (2006).
in the monkey. Cerebral Cortex, 4, 484–498. Neurobiological consequences of early stress and child-
Sanchez, M.M., Hearn, E.F., Do, D., et al. (1998). Differ- hood maltreatment. Annals of the New York Academy of
ential rearing affects corpus callosum size and cognitive Science, 1071, 313–323.
function of rhesus monkeys. Brain Research, 812, 38– Thatcher, R.W. (1992). Cyclic cortical reorganization dur-
49. ing early childhood. Brain and Cognition, 20, 24–50.
Sapolsky, R.M., Krey, L.C., & McEwen, V. (1985). Pro- Thatcher, R.W., North, D., & Biver, C. (2005). EEG and
longed glucocorticoid exposure reduces hippocampal intelligence: Relations between EEG coherence, EEG
neuron number: Implications for aging. Journal of Neu- phase delay and power. Clinical Neurophysiology, 116,
roscience, 5, 1222–1227. 2129–2141.
Sapolsky, R.M., & Meaney, M.J. (1986). Maturation of the Tomoda, A., Navalta, C., Polcari, A., Sadato, M., &
adrenocortical stress response: Neuroendocrine control Teicher, M.H. (2009a). Childhood sexual abuse is
mechanisms and the stress hyporesponsive period. associated with reduced gray matter volume in visual
Brain Research, 396, 64–76. cortex of young women. Biological Psychiatry, 66, 642–
Sapolsky, R.M., Uno, H., Rebert, C.S., et al. (1990). 648.
Hippocampal damage associated with prolonged gluco- Tomoda, A., Suzuki, H., Rabi, K., Sheu, Y., Polcari, A., &
corticoid exposure in primates. Journal of Neuroscience, Teicher, M.H. (2009b). Reduced prefrontal cortical gray
10, 2897–2902. matter volume in young adults exposed to harsh corpo-
Schiffer, F., Teicher, M.H., & Papanicolaou, A.C. (1995). ral punishment. NeuroImage, 47, T66–T71.
Evoked potential evidence for right brain activity during Ullen, F. (2009). Is activity regulation of late myelination a
the recall of traumatic memories. Journal of Neuropsy- plastic mechanism in the human nervous system?
chiatry and Clinical Neuroscience, 7, 169–175. Neuron Glia Biology, 29, 1–6.
Seifritz, E., Esposito, F., Neuhoff, J.G., Luthi, A., Mustovic, van den Boom, D. (1994). The influence of temperament
H., Dammann, G., von Bardeleben, U., Radue, E.W., and mothering on attachment and exploration. Child
Cirillo, S., Tedeschi, G., & Di Salle, F. (2003). Differential Development, 65, 1457–1477.
sex-independent amygdale response to infant crying and Vinberg, M., Trajkovska, V., Bennike, B., Knorr, U.,
laughing in parents versus nonparents. Biological Psy- Knudsen, G.M., & Kessing, L.V. (2009). The BDNF
chiatry, 54, 1367–1375. Val66Met polymorphism: Relation to familiar risk of
Seress, L. (2001). Morphological changes of the human affective disorder, BDNF levels and salivary cortisol.
hippocampal formation from midgestation to early child- Psycheuroendocrinology, 34, 1380–1389.
hood. In C.A. Nelson, & M. Luciana (Eds.), Handbook of Vythilingam, M., Heim, C., Newport, J., Miller, A.H.,
developmental cognitive neuroscience (pp. 45–58). Cam- Anderson, E., et al. (2002). Childhood trauma associated
bridge, MA: MIT Press. with smaller hippocampal volume in women with major
Shackman, J.E., Shackman, A.J., & Pollak, S.D. (2007). depression. American Journal of Psychiatry, 159, 2072–
Physical abuse amplifies attention to threat and 2080.
increases anxiety in children. Emotion, 7, 838–852. Wallace, C.S., Kilman, V.L., Withers, G.S., & Greenough,
Song, H., Kempermann, G., Overstreet Wadiche, L., Zhao, W.T. (1992). Increases in dendritic length in occipital
C., Schinder, A.G., & Bischofberger, J. (2005). New cortex after 4 days of differential housing in weanling
neurons in the adult mammalian brain: Synaptogenesis rats. Behavioral and Neural Biology, 58, 64–68.
and functional integration. Journal of Neuroscience, 25, Webster, M.J., Ungerleider, L.G., & Bachevalier, J. (1991).
10366–10368. Connections of inferior temporal areas TE and TEO with
Sowell, E.R., Thompson, P.M., & Toga, A.M. (2004). Map- medial temporal lobe structures in infant and adult
ping changes in the human cortex throughout the span monkeys. Journal of Neuroscience, 11, 1095–1116.
of life. Neuroscientist, 10, 372–392. Whittle, S., Yap, M.B.H., Yucel, M., Fornito, A., Simmons,
Stein, M.B., Koverloa, C., Hanna, C., et al. (1997). Hippo- J.G., Barrett, A., Sheeber, L., & Allen, N.B. (2008).
campal volume in women victimized by childhood sexual Prefrontal and amygdala volumes are related to adoles-
abuse. Psychological Medicine, 27, 951–959. cents’ affective behaviors during parent–adolescent
Stiles, J (2008). The fundamentals of human brain devel- interactions. Proceedings of the National Academy of
opment. Integrating nature and nurture. Cambridge, MA: Sciences USA, 105, 3652–3657.
Harvard University Press. Whittle, S., Yap, M.B.H., Yucel, M., Sheeber, L., Simmons,
Supekar, K., Musen, M., & Menon, V. (2009). Development J.H., Pantelis, C., & Allen, N.B. (2009). Maternal responses
of large-scale functional brain networks in children. PloS to adolescent positive affect are associated with adoles-
Bio, 7, e1000157. cents’ reward neuroanatomy. Scan, 4, 247–256.
Tarullo, A., Chatham, M., & Gunnar, M.R. (2007). Electro- Wilke, M., Krägeloh-Mann, I., & Holland, S.K. (2007).
encephalogram power in post-institutionalized children. Global and local development of gray and white matter
Paper presented at the biennial meeting of the Society for volume in normal children and adolescents. Experimen-
Research in Child Development, Boston, MA. tal Brain Research, 178, 296–304.
Taylor, S.E., Eisenberger, N.K., Saxbe, D., Lehman, B.J., & Zecevic, N. (1998). Synaptogenesis in Layer 1 of the human
Lieberman, M.D. (2006). Neural responses to emotional cerebral cortex in the first half of gestation. Cerebral
stimuli are associated with childhood family stress. Cortex, 8, 245–252.
Biological Psychiatry, 60, 296–301.
Teicher, M.H., Dumont, N.L., Ito, Y., Vaituzis, C., Giedd, Accepted for publication: 29 April 2010
J.N., & Andersen, S.L. (2004). Childhood neglect is Published online: 7 October 2010
 2010 The Authors
Journal of Child Psychology and Psychiatry  2010 Association for Child and Adolescent Mental Health.