Arthur W.

Toga
Paul M. Thompson
Elizabeth R. Sowell
Mapping Brain
Maturation
Human brain maturation is a complex, lifelong process that can now be examined in detail using neuroimaging tech-
niques. Ongoing projects scan subjects longitudinally with structural magnetic resonance imaging (MRI), enabling the
time-course and anatomical sequence of development to be reconstructed. Here, we review recent progress on imaging
studies of development. We focus on cortical and subcortical changes observed in healthy children, and contrast them
with abnormal developmental changes in early-onset schizophrenia, fetal alcohol syndrome, attention-decithyperac-
tivity disorder (ADHD) and Williams syndrome. We relate these structural changes to the cellular processes that under-
lie them, and to cognitive and behavioral changes occurring throughout childhood and adolescence.
(Reprinted with permission from Trends in Neurosciences March 2006; 29(3):148159)
INTRODUCTION
The dynamic course of brain maturation is one of
the most fascinating aspects of the human condi-
tion. Although brain change and adaptation are
part of a lifelong process, the earliest phases of mat-
urationduring fetal development and child-
hoodare perhaps the most dramatic and impor-
tant. Indeed, much of the potential and many of the
vulnerabilities of the brain might, in part, depend
on the rst two decades of life.
The cortex and subcortical gray-matter nuclei de-
velop during fetal life in a carefully orchestrated
sequence of cell proliferation, migration and matu-
ration. This leads to a human brain with 100
billion neurons at birth. However, the brain of a
newborn child is only one-quarter to one-third of
its adult volume, and it continues to grow and spe-
cialize according to a precise genetic program, with
modications driven by environmental inuences,
both positive and negative. With stimulation and
experience, the dendritic branching of neurons
greatly increases, as do the numbers of synaptic
connections. As layers of insulating lipids are laid
down on axons through the process of myelination,
the conduction speed of bers that interconnect
different brain regions increases 100-fold. This
exuberant increase in brain connections is followed
by an enigmatic process of dendritic pruning and
synapse elimination, which leads to a more efcient
set of connections that are continuously remodeled
throughout life.
SYNAPTIC CHANGES AND MYELINATION
When Huttenlocher began to demonstrate this
succession of events in the human brain, magnetic
resonance imaging (MRI) was still in its infancy.
His histological work in the 1980s paved the way in
showing that the time-course for synaptic bloom-
ing and pruning in the human brain varies enor-
mously by brain region. For example, in the visual
cortex, synaptic overproduction reaches a maxi-
mum at about the fourth postnatal month. Then
synapse elimination starts, and this continues until
preschool age, by which time synaptic density has
reached the adult level. But in the medial prefrontal
cortex, an area of the brain involved in executive,
attentional and regulatory functions, the peak oc-
curs at 34 years of age, and substantial decline
does not occur until mid-to-late adolescence (1).
Early work by Yakovlev and Lecours docu-
mented the progression of myelination in the de-
veloping human brain (2). More recent work by
Benes et al. with much larger samples showed sim-
ilar results, with myelination continuing well into
the third decade of life (3). Interestingly, the spatial
and temporal pattern of these cellular changes
seemed to parallel developmental changes in synap-
tic density. Simply put, myelination of the most
dorsal regions of the brain responsible for higher
cognitive functions seemed to continue well into
adolescence, and more ventral and deep brain struc-
turessome of which are responsible for relatively
more primitive functionswere myelinated earlier.
However, the sequence of myelination might be
more complex than this. A problem with many in-
terpretations of the work by Yakovlev and Lecours
(2), and of similar studies of myelogenesis, is that
dorsal and ventral systems do not ultimately my-
elinate to the same degree: on average, dorsal cor-
tex, at its most mature, is less myelinated than ven-
tral cortex. Therefore, if myelination took the same
time-course, then at any given time areas that my-
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
378
elinate more fully would appear more myelinated
than those that myelinate less. Therefore, when the
degree of myelination is observed on histologically
stained sections or in MRI scans of the brain, one
must also consider the ultimate extent of myelina-
tion for each system. Recent research studying pat-
terning molecules, and rates of proliferation and
neuronal migration, show a much more complex
pattern of cortical maturation, occurring primarily
from a rostral-lateral-ventral pole toward a dorsal-
medial-caudal pole (reviewed in (4)).
This histological evidence suggested that brain
development is a dynamic process of progressive
and regressive changes. But histology gave only
very fragmented evidence for late brain matura-
tion, given the dearth of post-mortem data from
childhood and adolescent years. By contrast,
MRI can non-invasively document these large-
scale processes of brain development, can pro-
vide insight into the sequence and timing of
these developmental processes in longitudinal
experiments, and can document how they occur
in living subjects.
INITIAL BRAIN-IMAGING STUDIES
The rst quantitative structural brain-imaging
studies in normal children using MRI were con-
ducted in the late 1980s and early 1990s. Terry
Jernigan and colleagues showed that young
adults actually had less cortical gray matter than
children, despite the fact that adults had some-
what larger overall total brain volumes (5). They
found that gray-matter volumes generally de-
clined after age 7, perhaps because the advance-
ment of white matter (i.e. myelination) through-
out childhood began to overtake the overall rate
of brain volume expansion, causing a net decre-
ment in the amount of tissue appearing gray (or
unmyelinated) on MRI. They then determined
that the timing of gray-matter loss was different
for different brain regions: loss was rst observed
in the deep motor nuclei in early childhood, then
in the parietal and frontal lobes at around pu-
berty (6). Although MRI did not assess synaptic
density per se, this was the rst in vivo evidence to
support the post-mortem ndings of Hutten-
locher and of Yakovlev and Lecours.
ANATOMICAL PARCELLATION AND MAPPING
The aforementioned studies used a method
called volumetric parcellation. In this approach, the
brain is subdivided into several, separate anatomi-
cal regions with different functions (e.g. the frontal
lobe, the deep motor nuclei or the hippocampus),
and their volumes are measured (Figure 1). Parcel-
lation enables the compilation of growth curves
that document howregional volumes vary with age.
Results are typically illustrated using scatterplots
but regional measurements are limited by anatom-
ical structures (i.e. sulcal landmarks) and can also be
reliably visualized and dened using MRI.
In the late 1990s, a few teams began to make
composite 3D maps of developing brain structures
(4, 712) showing, for example, the average pattern
of age-related change in gray-matter thickness be-
tween childhood and young adulthood. More re-
cently, these mapping techniques have become more
popular because they provide more visual detail and
group statistics regarding the rates of change and their
regional variation. Time-lapse animations also can be
computed to illustrate the changes.
Many of the basic functions of the brain, such as
vision, hearing, speech, planning and emotional
control are primarily handled in the cortex, so
much of the work mapping brain development has
focused on the cortex, although some groups have
studied subcortical gray matter. The goal of cortical
mapping is to create group average maps of cortical
features of interest such as gray-matter thickness,
gray-matter density, cortical shape, average sulcal
patterning and hemispheric asymmetries, all of
which change during development (for methods,
see (13)). Next, statistics are dened that help lo-
calize age effects on these measurements, such as
localized reductions or increases in gray-matter
density or thickness, and brain growth.
CHANGES IN GRAY AND WHITE MATTER
The earliest cross-sectional pediatric brain MRI
studies of normal developmental changes showed
that gray-matter volumes generally declined after
67 years of age and continued to decrease during
adolescence, whereas white-matter volumes in-
creased linearly over time. However, in one of the
rst studies to compile growth curves for the vol-
umes of different lobes of the brain as subjects aged
(i.e. studies that were longitudinal rather than
cross-sectional) (14), there was a clear linear in-
crease in white matter up to age 20, whereas there
were non-linear changes in cortical gray matter.
Giedd et al. demonstrated a pre-adolescent in-
crease, with developmental curves peaking at 12
years for the frontal and parietal lobe, and at 16
years for the temporal lobe. After that, gray-matter
loss occurs.
More recently, another developmental study
quantied human cortical development by measur-
ing gray-matter density in each lobe, point by point
(4). This map was constructed from serial brain
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
379
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
MRI scans of 13 children followed over a ten year
period. Children were scanned every two years for
eight years from the time they were recruited, and
were given a structured diagnostic interviewat each
visit to conrm lack of a psychiatric disorder.
This study measured which brain regions change
between 4 and 21 years of age and illustrated these
statistics as a time-lapse movie. This revealed a
shifting pattern of gray-matter loss, appearing rst
(at 48 years of age) in dorsal parietal and pri-
mary sensorimotor regions near the interhemi-
spheric margin, and spreading laterally and cau-
dally into temporal cortices and anteriorly into
dorsolateral prefrontal areas. The rst areas to ma-
ture were those with the most basic functions, such
as those processing the senses and movement. Areas
involved in spatial orientation and language (parietal
lobes) followed, around the age of puberty (1113
years). Areas with more advanced functionsinte-
grating information from the senses, reasoning and
other executive functions (e.g. prefrontal cortex)
matured last, in late adolescence. This sequence also
provided evidence that phylogenetically older cortical
areas mature earlier than the more recently evolved
higher-order association cortices, which integrate in-
formation from earlier maturing cortex.
CORTICAL THICKNESS
In another longitudinal study, changes in cortical
thickness were measured (in millimeters) in a group
of 45 normally developing children studied be-
tween 5 and 11 years of age (12). Each child was
studied twice with a two-year scanning interval.
Maps of cortical thickness were created for each
child, and the average cortical-thickness maps were
remarkably similar to maps created from postmor-
tem data by von Economo (15). The average corti-
cal thickness for the normally developing children
and from the post-mortem data are shown in Fig-
ure 2. As can be seen on the medial brain surface,
the cortex is thickest in the most dorsal aspects of
the frontal and parietal lobes (45 mm), and
thinnest in the visual cortices of the occipital lobes
surrounding the calcarine sulcus (2.02.5 mm).
These data revealed cortical thinning of 0.15
Figure 1. Typical processing steps in an analysis of MRI brain scans
(a) Atypical coronal section from a T1-weighted MRI scan of the brain. (b) The result of applying a tissue-classication approach to classify image voxels as gray
matter (green), white matter (blue) or cerebrospinal uid (CSF; red). Non-brain tissues such as scalp and meninges surrounding the brain have been digitally edited
from the image. (c) Parcellation of the brain into the frontal lobe (blue), parietal lobe (green), occipital lobe (red) and temporal lobe (yellow). This subdivision of
anatomy is performed with the aid of a cortical surface model on which sulcal landmarks separating the lobes can be reliably identied. Once partitioned in this
way, the volumes of each tissue type in the major lobes can be computed and growth curves established for each major lobe. The quantity of gray and white
matter in the brain and of CSF in the ventricles and cortical sulci can be computed and compared across subjects and overtime. Maps of these tissue types (b)
can be used to calculate shape, size and other statistics, and can also be subdivided into smaller regions to determine the amount of each tissue type in each
lobe (c).
Figure 2. Cortical thickness maps
(a) An in vivo average cortical-thickness map created from 45 normally developing
children at their rst scan (12). The brain surface is color coded according to the bar
on the right, where thickness is shown in millimeters. The average thickness map
can be compared to an adapted version of the 1929 cortical thickness map of von
Economo (15) (b). Color coding has been applied over his original stippling pattern,
respecting the boundaries of his original work, to highlight the similarities between
the two maps. Reproduced, with permission, from (12).
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
380
0.30 mm per year in normally developing children,
most prominently in right dorsal frontal and bilat-
eral parietal regions. Cortical thickening was also
observed with increases of 0.100.15 mm per
year in the classical language regions of the tempo-
ral and frontal lobes (Figure 3). The brain grew at a
rate of 0.41.5 mm per year in many regions,
most prominently in the frontal and occipital re-
gions. This pattern of results is similar to those
observed in cross-sectional and longitudinal data of
normal brain maturation (4, 16), but the new
methods enabled changes in cortical thickness to be
assessed in millimeters for the rst time. Addition-
ally, this study conrmed that dynamic but distinct
changes occur in posterior temporal and frontal
language regions, where gray matter continues to
increase in thickness. We speculate that cortical
changes in these regions could be related to chang-
ing language abilities that continue after 5 years of
age (e.g. changes in reading ability) (12); functional
and structural imaging studies are underway to test
these hypotheses.
COGNITIVE CORRELATES
Changes in cortical thickness during normal de-
velopment relate to cognitive changes as children
and adolescents mature. In the same 45 normally
developing children already described, cortical
thinning in the left dorsal frontal and parietal lobes
Figure 3. Annualized rate of change in cortical thickness
The average rate of change in cortical thickness is shown in millimeters according to the color bar on the right (maximum gray-matter loss is shown in shades of
red and maximum gray-matter gain is shown in shades of blue). Forty-ve children were studied twice (two-year scan interval) between 5 and 11 years of age.
Reproduced, with permission, from (12).
Figure 4. Brain-behavior maps for vocabulary and cortical thickness
P values are for negative correlations between change in cortical thickness (time 2 minus time 1, as shown in gure 2 of (12)) and change in vocabulary raw
scores (time 2 minus time 1). Negative P values (i.e. regions where greater thinning was associated with greater vocabulary improvement) are color-coded, and
regions in white showed no signicant association. Positive correlations were not signicant in the permutation analyses for any of the regions of interest, and are
not shown here. Reproduced, with permission, from (12).
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
381
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
correlated with improved performance on a test of
general verbal intellectual functioning (12, 17).
Greater left hemisphere gray-matter thinning was
associated with improved performance on the vo-
cabulary subtest of a standardized IQ measure (17)
(Figure 4). When permutation analyses were con-
ducted to correct for multiple comparisons in the
brain-behavior analyses, change in cortical thick-
ness was correlated with change in the behavioral
measure only in dorsal frontal and parietal cortex of
the left hemisphere. Cortex in these regions tends
to thin with increased age. Thus, the results of neg-
ative relationships between cortical thickness and
vocabulary improvements are consistent with what
would be expected. Considering these data, it seems
reasonable to speculate that the brain changes are
related to cognitive changes during development,
although few studies have addressed this issue.
CORTICAL CHANGES THROUGHOUT LIFE
The brain-mapping studies of normal brain mat-
uration described here have been limited by the age
ranges of the subjects studied. They have clearly
shown regional and temporal patterns of dynamic
maturational change continuing through child-
hood and adolescence. However, they could not
enable conclusions to be made about the endpoint
in cortical sculpting (which probably results from
increased myelination and synaptic pruning) be-
cause they did not assess individuals 30 years of
age. This issue was addressed in a study of a large
sample of normal individuals (n 176) across the
span of life (787 years of age) (11). Signicant,
non-linear age effects were observed over large areas
of the most dorsal aspects of the frontal and parietal
regions on both the lateral and interhemispheric
surfaces and in the orbitofrontal cortex. Scatter-
plots of these effects revealed a dramatic decline in
gray-matter density between 7 and 60 years of age,
with little or no decline thereafter. A sample scat-
terplot of the quadratic effect of age on gray-matter
density at one brain-surface point on the superior
frontal sulcus is shown in Figure 5, and is similar to
plots from the dorsal frontal and parietal regions.
Notably, the most lateral aspects of the brain in the
posterior temporal and inferior parietal lobes bilat-
erally showed a distinct pattern of gray-matter
change, one in which the non-linear age effects
were inverted relative to the age effects in more
dorsal cortices. A subtle increase in gray-matter
density was observed until 30 years of age, and it
then remained stable until a precipitous decline in
later decades. As these data illustrate, the trajectory
of maturational changes might continue beyond
adolescence and into young adulthood, and can be
assessed only using an extended age range. The
question of when brain maturational changes tra-
verse into the more degenerative changes of aging
becomes even more poignant when considering these
data (for a detailed discussion, see (18)). The last sys-
tems to mature are also among the rst to degenerate
in disorders such as Alzheimers disease; their high
degree of plasticity throughout life might make them
more vulnerable to neurodegeneration (1920).
SUBCORTICAL CHANGES DURING DEVELOPMENT
Most brain-mapping studies of development
have focused on the cortex, but developmental
changes in subcortical structures have also been
characterized. For example, one study mapped a
front-to-back wave of growth through the corpus
callosum at 315 years of age (22) (Figure 6).
The continuous myelination of interhemispheric
bers results in increased axonal conduction
speed and transmission of information between
the brain hemispheres. Perhaps the main surprise
from the subcortical studies is the progressive
volume loss in the deep gray-matter nuclei
especially the caudateat around puberty. In
contrast to the basal ganglia, temporal lobe gray-
matter structures (the amygdala and hippocam-
pus) seem to increase in volume during child-
hood and adolescence.
RELATIONSHIP TO CELLULAR CHANGES
We cannot conclude directly what cellular
changes are involved in the dynamic maturational
processes described here. MRI measures changes in
the volume and density of brain structures, but
lacks the resolution to characterize the cellular
mechanism (e.g. dendritic remodeling, cell death,
synaptic pruning or myelination) of such changes.
Moreover, gray-matter volume and thickness re-
ect not only the dendritic and synaptic processes
that occur in neurons but also the complex archi-
tecture of neurons, glia (i.e. myelin) and vascula-
ture. However, we can say that the gray-matter
changes described in this review correlate well with
the post-mortemsequence of regionally variable in-
creased synaptic pruning and myelination during
adolescence and early adulthood.
FETAL AND NEONATAL MRI
Mapping fetal and neonatal brains presents an-
other set of practical and technical challenges. Aside
from the obvious difculty in obtaining motion-
artifact-free images, at these ages the degree of my-
elination produces considerably less contrast in im-
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
382
ages obtained using the most commonly applied
T1 and T2 weighted pulse sequences. Nevertheless,
considerable insight into the earliest myelination
and differentiation of the telencephalon and dien-
cephalon can be seen by imaging fromfetal stages in
utero through the rst six months after birth (23).
These imaging results are consistent with post-mor-
tem measures of increased myelination before birth
(2426). A general caudal-to-rostral progression
can be appreciated in this age group. Furthermore,
morphological differentiation of several structures
such as the internal capsule (a fan-like structure of
white matter separating the lentiform nucleus from
the caudate and dorsal thalamus) precedes changes
in white-matter, often by many months. Utilizing
specic magnetic resonance pulse sequences to
identify white matter, in an approach called diffu-
sion tensor imaging (DTI), can also help character-
Figure 5. Mapping brain change overtime
Brain changes in development can be identied by tting time-dependent statistical models to data collected from subjects cross-sectionally (i.e. across a group of
subjects at a particular time), longitudinally (i.e. following individual subjects as they aged), or both. Measurements such as cortical thickness are then plotted onto
the cortex using a color code. (a,b) Trajectory of gray-matter loss over the human lifespan, based on a cohort of 176 subjects aged 787 years (11). Plots super-
imposed on the brain in (b) show how gray-matter density decreases for particular regions; (a) highlights example regions in which the gray-matter density de-
creases rapidly during adolescence (the superior frontal sulcus) or follows a more steadily declining time-course during lifespan (the superior temporal sulcus).
(c,d) Trajectory of cortical gray-matter density in 13 children scanned longitudinally every two years for eight years (4). The units used in (a) and (d) are gray-
matter density, which is dened as the proportion of tissue segmenting as gray matter within a 15-mm-diameter sphere centered at each point on the cortical
surface. This widely-used measure ranges from 0 to 1, and is highly correlated with cortical thickness (52).
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
383
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
ize the changes associated with myelination. Neil et
al. (27) found dramatic and generalized changes in
water apparent diffusion coefcients and diffusion
anisotropy. These diffusion parameters, measur-
able using DTI, are locally sensitive to the degree of
myelination (which provides resistance to water
diffusion) and to ber orientation (which con-
strains the principal directions of water diffusion to
be aligned with axons). Gilmore et al. (28) found
signicant increases in fractional anisotropy in the
anterior and posterior regions (genu and splenium)
of the corpus callosum during development that
reect underlying changes to water content and cy-
toarchitecture during maturation.
DEVELOPMENTAL DISORDERS
Statistical maps can also be created to assess
group effects such as brain structure differences be-
tween healthy children and those with neuropsy-
chiatric disorders such as schizophrenia, bipolar ill-
ness or attention-decithyperactivity disorder
(ADHD), those with genetic disorders of brain de-
velopment such as Williams syndrome (WS; a
mental retardation disorder), and those exposed to
teratogens during brain development, such as chil-
dren with fetal alcohol syndrome (FAS). The result-
ing maps can reveal where in the brain differences
are located, how signicant they are, andde-
pending on the study designwhether they are sta-
ble or progressive (29).
Statistically signicant reductions or increases in
measured gray- or white-matter structures are com-
monly found in children with developmental dis-
orders, and different brain abnormalities have been
observed in different disorders. However, most of
the abnormalities observed in the quantitative MRI
studies of children with developmental disorders
have been relatively subtle. That is, the brain struc-
tural abnormalities could not typically be observed
within individual children, but rather only when
studied in groups of children with a particular dis-
order compared with non-affected or medication-
matched controls.
CHILDHOOD-ONSET SCHIZOPHRENIA
Beginning in 1992, scientists led by Judith Rap-
oport at the National Institute of Mental Health
employed MRI technology to scan a group of 50
teenagers repeatedly as their schizophrenia devel-
oped and a group of 500 healthy children as con-
trols. The collected data were analyzed using brain-
mapping methods developed to detect subtle
changes in the cortex (30). Healthy subjects lost
gray matter at a subtle rate of 12% per year in the
parietal cortices, with very little detectable change
in the other lobes of the brain. By contrast, the
childhood-onset schizophrenia patients showed a
rapid progressive loss of gray matter in superior
frontal and temporal cortices, reaching 34% per
year in some regions. This pattern was seen in both
Figure 6. Tensor maps of growth and tensor-based morphometry
(a) The corpus callosum (indicated by a green box) of a healthy three-year-old girl in a sagittal section from a 3D MRI scan. Using a follow-up scan thee years
later, an elastic deformation eld is computed that digitally aligns, or warps, the anatomy of the earlier time-point to match its shape at the later time point. The
amount of local stretching of the anatomy is color-coded, indicating fastest growth rates (red) in the anterior corpus callosum. In a related approach (b), maps can
be compiled to represent the average expansion factor required to deform an average corpus callosum shape elastically onto each subject in a set of healthy chil-
dren and matched autistic children. This identies areas of the corpus callosum that are thinner in autistic children, pinpointing where abnormal white-matter
growth might lead to the disorder in early childhood. The letters S and R denote the splenium and rostrum of the corpus callosum, respectively. Data in (a) is
adapted, with permission, from (22); (b) is adapted, with permission, from (53).
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
384
boys and girls. Early decits in parietal brain re-
gions that support language and associative think-
ing spread forwards into the temporal lobes, sup-
plementary motor cortices and frontal eye elds.
The decits spread anatomically over a period of 5
years, consistent with the characteristic neuromo-
tor, sensory and visual search impairments in the
disease. In temporal cortices, including primary au-
ditory regions, severe gray-matter loss was absent at
disease onset but later became pervasive. Patients
with the worst brain tissue loss also had the worst
symptoms, which included hallucinations, delu-
sions, bizarre and psychotic thoughts, hearing
voices and depression.
The notion of schizophrenia as a dynamically
emerging disease of neurodevelopment is still con-
troversial. The spreading patterns of decits cor-
roborate the idea that the normal developmental
process of dendritic and synaptic pruning might be
abnormally accelerated or derailed in schizophrenia
(31). Even so, they might represent a separate pro-
cess entirely that begins in adolescencetwins
studies have hinted that the earliest decits, occur-
ring in the parietal cortex, might have a non-genetic
trigger (32). Some histological studies suggest that
there is decreased neuropil in schizophrenia, but
autopsy material is scarce and ndings are inconsis-
tent. As in normal development, the observed cor-
tical changes might also be glial or vascular in ori-
gin, rather than purely neuronal (33). Intracortical
contrast might also be affected by changes in my-
elination (34), and by changes in lipid metabolism
during the use of atypical antipsychotic agents. It is
possible that in schizophrenia, changes in gray-
matter density also reect, at least in part, increas-
ing intracortical myelination. Still, four dimen-
sional maps such as these offer a biological marker
of disease progression for drug trials, where brain
mapping can assess howwell antipsychotic medica-
tions oppose advancement of the disease (35).
WILLIAMS SYNDROME
Brain imaging has also helped to identify circum-
scribed alterations in brain structure that are asso-
ciated with developmental disorders that have ge-
netic causes. Williams syndrome (WS), for
example, is a developmental disorder associated
with a deletion of 20 genes in the 7q11.23 region
of chromosome 7. WS subjects have mild to mod-
erate mental retardation, but have remarkable pro-
ciencies in language skills, social drive and musical
ability. Figure 7 shows a range of brain regions af-
fected by this genetic deletion: in general, the cortex
is thinner, except in perisylvian language regions,
where it is thicker by 510%; cortical complexity is
also signicantly increased (36). In post-mortem
data from WS subjects, Holinger et al. (37) ob-
served larger cells in Heschls gyrus in the primary
auditory areaa nding congruent with the corti-
cal thickening seen here. WS subjects also have dis-
proportionately reduced white matter, so cortical
cells tting over a smaller white-matter area can
pile up, and the cortex can thicken owing to a
crowding effect. Some Wnt signaling genes, which
direct early cell differentiation and segmentation,
are in the deleted 7q11.23 chromosomal region,
suggesting a genetic basis for the cortical dysmor-
phology observed on MRI. Other genes deleted in
WS might normally contribute to ssure formation
in the human cortex, and their absence might ac-
count for the abnormal cortical thickening. Ongo-
ing work is examining the changes in folding com-
plexity in the same regions that show thickened
cortex, to ensure that this thickening is not simply a
consequence of increased gyrus formation (e.g. mi-
crogyria often appears as thickened cortex, when in
fact it is thinner, four-layered cortex). If cortical
thickness and complexity were positively correlated
in general, the frontal and parietal regions of greater
gyrication found in studies of WS (e.g. (38))
would be expected to show greater cortical thick-
ness, but this was not found. There is, therefore, no
simple relationship between thickness and com-
plexity, and careful study of these measures in larger
samples is warranted. Alternatively, the cortical
thickening might represent an adaptive response to
the genetic deletion, perhaps even from increased
use or overuse of specic cortical networks. Identi-
cation of such neuroanatomical characteristics of
WS is important for better understanding of the
scope and timing of the cortical anomaly in WS.
ATTENTION-DEFICITHYPERACTIVITY DISORDER
ADHD is a much more common, and less debil-
itating, disorder of childhood than childhood-on-
set schizophrenia and WS. Nonetheless, volumetric
structural brain-imaging studies of ADHDpatients
have demonstrated subtle reductions in total brain
volume, and in volumes of the right frontal lobe
and caudate nucleus (39, 40). To further under-
standing of brain abnormalities in ADHD, brain
mapping studies were conducted in 27 children
and adolescents with ADHD and in 46 age-
matched and gender-matched control subjects
(41). As predicted, abnormal morphology was ob-
served in the frontal cortices of ADHD subjects,
with reduced regional brain size localized primarily
to inferior portions of dorsal prefrontal cortices bi-
laterally. Brain size was also reduced in anterior
temporal cortices bilaterally. Gray-matter density
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
385
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
was prominently increased in large portions of the
posterior temporal and inferior parietal cortices bi-
laterally (Figure 7). The etiology and cognitive and
behavioral sequelae of this increased gray-matter
density and regionally reduced brain size are not yet
understood, but these regions of the brain are
known to be part of an action-attention network
probably involved in the symptomatology of
ADHD (discussed in (41)).
FETAL ALCOHOL SYNDROME
Brain mapping has also identied anomalies in
developmental disorders with known environmen-
tal causes. Fetal alcohol syndrome (FAS) is a severe
neurodevelopmental disorder observed in children
whose mothers abused alcohol during pregnancy
(42, 43). Not all women who abuse alcohol during
pregnancy have children with FAS, but it is quickly
becoming more evident from animal and both post-
mortem and in vivo human studies that the brain is
especially vulnerable to the teratogenic effects of alco-
hol. Children with FAS display prenatal growth de-
ciency, developmental delay, craniofacial anomalies
(i.e. microcephaly, epicanthal folds and short palpe-
bral ssures) and limb defects. Cognitive decits in
these childrenare signicant andmental retardationis
frequently observed, but a clear neuropsychological
prole has yet to be discovered (44).
In vivo quantitative MRI studies have con-
rmed brain morphological abnormalities in
children prenatally exposed to alcohol. In addi-
tion to the microcephaly frequently observed in
these children, volumetric MRI studies have
shown that within the cortex, only the parietal
lobes were signicantly reduced in volume above
and beyond the generalized brain size reduction.
White-matter hypoplasia (i.e. incomplete devel-
opment) was more signicant than gray-matter
hypoplasia, and hippocampal volume was rela-
tively intact (45). These volumetric ndings
prompted use of cortical mapping techniques to
assess brain abnormalities on a more local level.
Localized abnormalities in brain size and gray-
matter density were identied in a group of 21
children, adolescents and young adults with se-
vere prenatal alcohol exposure compared with 21
age-matched and gender-matched non-exposed
subjects (46). The resulting brain maps con-
rmed parietal lobe size reductions observed in
the volumetric studies, and revealed frontal lobe
size reductions that had not been previously ob-
served. In fact, frontal lobe abnormalities were
predicted given these childrens decits on neu-
ropsychological tests of executive functioning
(47). Furthermore, gray-matter density increases
were observed bilaterally in the inferior parietal
and posterior temporal lobes (Figure 7). These
mapping studies suggest that brain growth con-
tinues to be adversely affected long after the pre-
natal insult of alcohol exposure to the developing
brain, and the regions most prominently affected
are consistent with the behavioral and neuropsy-
chological decits in these children.
SIMILARITIES AMONG THE DEVELOPMENTAL
DISORDERS
It is interesting to note the regional correspon-
dence of gray-matter density abnormalities ob-
served in three of the four neurodevelopmental dis-
orders reviewed here. As shown in Figure 7, gray-
matter density increases were observed in perisylvian
(language) cortices in ADHD(41), WS (36) and FAS
(46). The magnitude of increase ingray matter relative
to their respective control groups varies across the dis-
orders, with maxima of 30% in ADHD, 10% in
WS and 50%in FAS. The spatial extent of these gray-
matter density increases also varies across the disor-
ders, and is bilateral in ADHD and FAS, but more
prominent in the right hemisphere in WS. These dis-
orders are distinct in terms of their etiologies and neu-
rocognitive decit proles, although various levels of
language impairment are observed in all three disor-
ders. Facial dysmorphology is present in both FAS
and WS. Of particular relevance might be that
ADHD diagnoses are common among children with
FAS (48) and WS (49). Perhaps the similar prole of
gray-matter density abnormalities observed in these
three groups are related to the common symp-
toms of ADHD. Functional and structural imag-
ing studies evaluating children that have ADHD,
WS with ADHD, WS without ADHD, FAS with
ADHD, and FAS without ADHD might yield
interesting results, although they would be dif-
cult given the relative rarity of WS in particular.
These examples are illustrative only and summa-
rize visualizations derived from different studies
(e.g. case-control studies of FAS, WS and
ADHD). Further studies are required in which
the magnitudes and spatial patterns of the effects
in the different patient groups are directly com-
pared. Such direct comparisons would have to
take into account possible confounding factors,
because the children were assessed on different
scanners at different imaging centers.
GENETIC INFLUENCES ON THE
DEVELOPING BRAIN
In the quest to understand what factors contrib-
ute to the trajectory of brain development in nor-
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
386
mally developing children and those with neurode-
velopmental disorders, genetic and imaging
methods can be combined to answer questions
about the inuence of genes and environment on
brain structurethe so-called naturenurture de-
bate. With this in mind, twins or family members
with different degrees of genetic afnity can be
compared. Of course, inuences of nature and nur-
ture in the determination of individual brain struc-
ture are not independent. Nonetheless, twin designs
can reveal the degree to which heredity is involved.
TWIN STUDIES
Probably the best known twin database is the
Finnish twin registry, which covers information on
twin births in Finland from 1940 and has been a
resource to many epidemiological studies recog-
nized worldwide. These data enabled investigation
of which brain regions develop under tight genetic
control and which are more inuenced by the en-
vironment. Before that, twins had been studied us-
ing MRI and some similarities in the volumes of
brain structures had been noted in identical twins,
with weaker resemblance in fraternal twins. In
2001, utilizing this database, the rst study of twins
using brain mapping revealed aspects of brain struc-
ture that are largely inherited (50).
Forty healthy adult subjects, consisting of ten
monozygotic and ten dizygotic twin pairs (ve male
pairs and ve female pairs in each group) were
drawn from a twin cohort consisting of all the
same-sex twins born in Finland between 1940 and
1957 for which both members of each pair were
alive and still residing in Finland. All the subjects
were scanned and 3D maps of gray matter and
models of cortical surface anatomy were derived.
Monozygotic pairs were then matched with the
dizygotic pairs. Gray-matter volumes in the frontal
parts of the brain were more closely matched in the
Figure 7. Differences in gray-matter density between subjects with three
neurodevelopmental disorders
The percentage differences in gray-matter density between subjects with Williams syndrome (WS) (a), attention-decithyperactivity disorder (ADHD) (b), fetal al-
cohol syndrome (FAS) (c) and their respective normally developing control groups are color-coded. In all maps, warmer colors represent positive differences, indi-
cating an increase in the patient group (arbitrarily coded as 1) relative to the control group (arbitrarily coded as 0), with red representing the largest group differ-
ence. Note that the maximum value varies on the three color bars, depending on the maximum group difference from each comparison. Adapted, with permission,
from (36, 41, 46).
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
387
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
identical twins than in twins who were less similar
genetically. One interpretation of the ndings is
that areas of the brain involved in process and rule
learning (e.g. frontal lobe structures) are under
more tight genetic control, at least in terms of gross
structural development, than areas of the brain that
map life experience and content (the temporal and
parietal lobes).
The nding that the gross anatomy of the frontal
cortex develops under such strong genetic control is
also of practical use for genetic studies of disease.
These genetically mediated structural differences
have even been linked, using genetic brain-map-
ping methods, with known genetic markers that are
overrepresented in patients with frontal lobe de-
cits (51).
LIMITATIONS OF BRAIN-IMAGING STUDIES
As we have shown, MRI used in both cross-sec-
tional and longitudinal studies gives precious data
on normal and pathological brain development.
Key information is now coming from the large-
scale analysis of data from children scanned longi-
tudinally using MRI. Moreover, advances in MRI
techniques other than anatomical MRI are now of-
fering additional perspectives on the functional de-
velopment of the brain. Functional MRI studies are
nowemerging as a tool to study brain development,
as are ber-mapping studies using DTI to track
connections and how they develop during matura-
tion.
However, several limitations need to be acknowl-
edged. Some are technical. For example, because it
requires considerable statistical power to detect sys-
tematic differences in brain structure, sample size is
a major issue, especially when rare disorders are
considered. MRI data can be heterogeneous. Scan
databases are only now becoming large enough to
stratify images of brain development by symptom
proles, therapeutic response and currently identi-
ed risk factors.
Figure 8. Statistical maps of cortical structure. Various maps can be made that
describe different aspects of cortical anatomy
These include maps of gyrus-pattern asymmetry (a) and how strongly genes and environmental inuence brain structure (b). Panel (a) shows the increasing gy-
rus-pattern asymmetry in groups of children, adolescents and adults (55). In these maps, the differences in asymmetry in the three age groups can be evaluated
as the angle of the left relative to the right Sylvian ssure in each age group. Note the angle is larger in the adults than in the children. The color coding enables
the displacement in millimeters between the left and right hemisphere within each age group to be visualized. Asymmetry measures can also be extended to rest
of the cortical surface and expressed in millimeters. Panel (b) shows correlations in gray matter for groups of identical (monozygotic, MZ) and fraternal (dizygotic,
DZ) twins. Some brain regions develop under tight genetic control, and these include many frontal and temporal lobe regions, such as the dorsolateral prefrontal
cortex and temporal poles (red). Other regions are more strongly inuenced by the environment as they develop (i.e. a greater proportion of the inter-subject vari-
ance is explained by non-genetic factors). (c) Average maps of gray-matter loss rates for healthy boys and girls, scanned longitudinally over ve years. Also shown
are maps of the considerably faster loss rates in age-matched and gender-matched subjects with childhood-onset schizophrenia scanned at the same ages and
intervals. The frontal cortex underwent a selective rapid loss of gray matter (up to 34% per year faster in patients than controls). These changes might be, in
some respects, an exaggeration of changes that normally occur in adolescence (10,54). By contrast, decits occurring as Alzheimers disease progresses are
shown (d) by comparing average proles of gray-matter density between 12 Alzheimers disease patients (mean age 68.4 1.9 years) and 14 age-matched con-
trols (mean age 71.4 0.9 years). In Alzheimers disease, gray-matter loss sweeps forward in the brain from limbic to frontal cortices in concert with cognitive
decline, but in the schizophrenia patients (c), the frontal cortices lose gray matter the fastest. The letters W and T denote Wernickes area and the temporal
cortex, respectively.
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
388
MRI has some limits in resolving distinctions be-
tween gray and white matter that are relevant when
assessing maturational changes. Recent algorithms
for measuring cortical thickness (13) rely on de-
nition of a boundary between gray and white mat-
ter, or at least on a proportion of tissue identied as
gray matter (in the case of gray-matter density). As
gray matter becomes more myelinated, the border
between gray and white changes, and can be dis-
placed toward the pia. One factor in the cortical
thinning seen in adolescence is not absolute dimi-
nution of the cortical gray matter, but rather a
change in the graywhite segmentation boundary
because of increasing intracortical myelin with age,
which might shift the graywhite junction further
into the cortical mantle. Other anatomical features
might also affect the accurate segmentation of gray
and white matter on magnetic resonance images.
The inner band of Baillarger, for instance, is so
richly myelinated that in brain sections it is visible
with the naked eye, close to the graywhite inter-
face in layer 5b; this might affect the ability to de-
ne an accurate cortical boundary in conventional
magnetic resonance images. Further studies with
improved tissue classication algorithms, tissue re-
laxometry and high-eld imaging (providing in-
creased resolution and contrast-to-noise ratio) are
likely to resolve some of these limitations.
Other issues concern the interpretation given to
some data. The key question is: what does a change
in the volume of a brain region mean, especially
when one considers pathology (Figure 8)? Even if
the volume and shape of brain structures are deter-
mined, to identify the underlying cellular cause of
these differences one must rely on using additional
sources of information, such as post-mortemtissue.
Nevertheless, these MRI studies are likely to help us
better understand the links between changes in the
brain and cognitive development, in addition to the
clinical course of development and treatment of
illnesses.
R E F E R E N C E S
1. Huttenlocher, P.R. (1979) Synaptic density in human frontal cortex
developmental changes and effects of aging. Brain Res. 163, 195205
2. Yakovlev, P.I. and Lecours, A.R. (1967) The myelogenetic cycles of
regional maturation of the brain. In Regional Development of the Brain in
Early Life (Minkowski, A, ed.), pp. 370, Blackwell Scientic
3. Benes, F.M. et al. (1994) Myelination of a key relay zone in the hip-
pocampal formation occurs in the human brain during childhood, ado-
lescence, and adulthood. Arch. Gen. Psychiatry 51, 477484
4. Gogtay, N. et al. (2004) Dynamic mapping of human cortical development
during childhood and adolescence. Proc. Natl. Acad. Sci. 101, 8174
8179
5. Jernigan, T.L. and Tallal, P. (1990) Late childhood changes in brain
morphology observable with MRI. Dev. Med. Child Neurol. 32, 379385
6. Jernigan, T.L. et al. (1991) Maturation of human cerebrum observed in
vivo during adolescence. Brain 114, 20372049
7. Paus, T. et al. (1999) Structural maturation of neural pathways in children
and adolescents: in vivo study. Science 283, 19081911
8. Sowell, E.R. et al. (1999) In vivo evidence for postadolescent brain
maturation in frontal and striatal regions. Nat. Neurosci. 2, 859861
9. Sowell, E.R. et al. (1999) Localizing agerelated changes in brain struc-
ture between childhood and adolescence using statistical parametric
mapping. NeuroImage 9, 587597
10. Thompson, P.M. et al. (2001) Mapping adolescent brain change reveals
dynamic wave of accelerated gray matter loss in very early-onset schizo-
phrenia. Proc. Natl. Acad. Sci. U. S. A. 98, 1165011655
11. Sowell, E.R. et al. (2003) Mapping cortical change across the human life
span. Nat. Neurosci. 6, 309315
12. Sowell, E.R. et al. (2004) Longitudinal mapping of cortical thickness and
brain growth in normal children. J. Neurosci. 24, 82238231
13. Thompson, P.M. et al. (2004) Mapping cortical change in Alzheimers
disease, brain development, and schizophrenia: special issue on math-
ematics in brain imaging. NeuroImage 23(Suppl 1), S2S18
14. Giedd, J.N. et al. (1999) Brain development during childhood and
adolescence: a longitudinal MRI study. Nat. Neurosci. 2, 861863
15. Von Economo, C.V. (1929) The Cytoarchitectonics of the Human Cerebral
Cortex, Oxford Medical Publications
16. Sowell, E.R. et al. (2001) Mapping continued brain growth and gray
matter density reduction in dorsal frontal cortex: inverse relationships
during postadolescent brain maturation. J. Neurosci. 21, 88198829
17. Wechsler, D. (1991) Manual for the Wechsler Intelligence Scale for
Children, 3rd edn, The Psychological Corporation
18. Sowell, E.R. et al. (2004) Mapping changes in the human cortex through-
out the span of life. Neuroscientist 10, 372392
19. Mesulam, M.M. (2000) A plasticitybased theory of the pathogenesis of
Alzheimers disease. Ann. N. Y. Acad. Sci. 924, 4252
20. Thompson, P.M. et al. (2003) Dynamics of gray matter loss in Alzheimers
disease. J. Neurosci. 23, 9941005
21. Toga, A.W. and Thompson, P.M. (2003) Temporal dynamics of brain
anatomy. Annu. Rev. Biomed. Eng. 5, 119145
22. Thompson, P.M. et al. (2000) Growth patterns in the developing brain
detected by using continuum-mechanical tensor maps. Nature 404,
190193
23. Martin, E. et al. (1988) Developmental stages of human brain: an MR
study. J Comp Assist Tomogr. 12, 917922
24. Grodd, W. (1993) Normal and abnormal patterns of myelin development
of the fetal and infantile human brain using magnetic resonance imaging.
Curr. Opin. Neurol. Neurosurg 6, 393397
25. Hansen, P.E. et al. (1993) MR imaging of the developing human brain,
part 1. Radiographics 13, 2136
26. Wang, Z. et al. (1998) The research of myelinization of normal fetal brain
with magnetic resonance imaging. Chin. Med. J. (Engl.) 111, 7174
27. Neil, J. et al. (2002) Diffusion tensor imaging of normal and injured
developing human braina technical review. NMR Biomed. 15, 543552
28. Gilmore, J.H. et al. (2004) 3 Tesla magnetic resonance imaging of the
brain in newborns. Psychiatry Res. 132, 8185
29. Toga, A.W. et al. (2001) Probabilistic approaches for atlasing normal and
disease-specic brain variability. Anat. Embryol. (Berl.) 204, 267282
30. Thompson, P.M. et al. (2001) Mapping adolescent brain change reveals
dynamic wave of accelerated gray matter loss in very earlyonset
schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 98, 1165011655
31. Feinberg, I. and Guazzelli, M. (1999) Schizophreniaa disorder of the
corollary discharge systems that integrate the motor systems of thought
with the sensory systems of consciousness. Br. J. Psychiatry 174,
196204
32. Cannon, T.D. et al. (2002) Cortex mapping reveals regionally specic
patterns of genetic and diseasespecic graymatter decits in twins
discordant for schizophrenia. Proc. Natl. Acad. Sci. U. S. A. 99, 3228
3233
33. Weinberger, D.R. and McClure, R.K. (2002) Neurotoxicity neuroplasticity,
and magnetic resonance imaging morphometry: what is happening in the
schizophrenic brain? Arch. Gen. Psychiatry 59, 553558
34. Bartzokis, G. et al. (2003) White matter structural integrity in aging and
Alzheimers disease: a magnetic resonance imaging study. Arch. Neurol.
60, 393398
35. Lieberman, J.A. et al. (2005) HGDH Study Group (2005). Antipsychotic
drug effects on brain morphology in rstepisode psychosis. Arch. Gen.
Psychiatry 62, 361370
36. Thompson, P.M. et al. (2005) Abnormal cortical complexity and thickness
proles mapped in Williams syndrome. J. Neurosci. 25, 41464158
37. Holinger, D.P. et al. Relative sparing of primary auditory cortex in Williams
syndrome. Brain Res. (in press)
38. Schmitt, J.E. et al. (2002) Increased gyrication in Williams syndrome:
evidence using 3D MRI methods. Dev. Med. Child Neurol. 44, 292295
39. Filipek, P.A. et al. (1997) Volumetric MRI analysis comparing subjects
TOGA
focus.psychiatryonline.org FOCUS Summer 2006, Vol. IV, No. 3
389
I
N
F
L
U
E
N
T
I
A
L
P
U
B
L
I
C
A
T
I
O
N
S
having attentiondecit hyperactivity disorder with normal controls. Neu-
rology 48, 589601
40. Semrud-Clikeman, M. et al. (2000) Using MRI to examine brainbehavior
relationships in males with attention decit disorder with hyperactivity.
J. Am. Acad. Child Adolesc. Psychiatry 39, 477484
41. Sowell, E.R. et al. (2003b) Cortical abnormalities in children and adoles-
cents with attention decit hyperactivity disorder. Lancet 362, 1699
1707
42. Jones, K.L. and Smith, D.W. (1973) Recognition of the fetal alcohol
syndrome in early infancy. Lancet 2, 9991001
43. Jones, K.L. (1975) The fetal alcohol syndrome. Addict. Dis. 2, 7988
44. Mattson, S.N. and Riley, E.P. (1998) A review of the neurobehavioral
decits in children with fetal alcohol syndrome or prenatal exposure to
alcohol. Alcohol. Clin. Exp. Res. 22, 279294
45. Archibald, S.L. et al. (2001) Brain dysmorphology in individuals with
severe prenatal alcohol exposure. Dev. Med. Child Neurol. 43, 148154
46. Sowell, E.R. et al. (2002) Regional brain shape abnormalities persist into
adolescence after heavy prenatal alcohol exposure. Cereb. Cortex 12,
856865
47. Mattson, S.N. et al. (1999) Executive functioning in children with heavy
prenatal alcohol exposure. Alcohol. Clin. Exp. Res. 23, 18081815
48. OMalley, K.D. and Nanson, J. (2002) Clinical implications of a link
between fetal alcohol spectrum disorder and attention-decit hyperac-
tivity disorder. Can. J. Psychiatry 47, 349354
49. Bawden, H.N. et al. (1997) Treatment of children with Williams syndrome
with methylphenidate. J. Child Neurol. 12, 248252
50. Thompson, P.M. et al. (2001) Genetic inuences on brain structure. Nat.
Neurosci. 4, 12531258
51. Cannon, T.D. et al. (2003) Early and late neurodevelopmental inuences
in the prodrome to schizophrenia: contributions of genes, environment,
and their interactions. Schizophr. Bull. 29, 653669
52. Thompson, P.M. et al. (2004) Structural abnormalities in the brains of
human subjects who use methamphetamine. J. Neurosci. 24, 6028
6036
53. Leow, A. et al. (2005) Brain structural mapping using a novel hybrid
implicit/explicit framework based on the levelset method. Neuro-Image
24, 910927
54. Sowell, E.R. et al. (2002) Mapping sulcal pattern asymmetry and local
cortical surface gray matter distribution in vivo: maturation in perisylvian
cortices. Cereb. Cortex 12, 172654
N O T E S
TOGA
Summer 2006, Vol. IV, No. 3 F O C U S THE J OURNAL OF LI FELONG LEARNI NG I N PS YCHI ATRY
390