AR245-NE28-01 ARI 19 May 2005 10:1

Genetics of Brain Structure

and Intelligence
Arthur W. Toga and Paul M. Thompson
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

Laboratory of Neuro Imaging, Department of Neurology, School of Medicine,

University of California, Los Angeles, California 90095; email: toga@[Link],
thompson@[Link]
by University of Sussex Library on 11/10/06. For personal use only.

Annu. Rev. Neurosci. Key Words

2005. 28:1–23
heritability, IQ, brain mapping
doi: 10.1146/
[Link].28.061604.135655
Abstract
Copyright

c 2005 by
Annual Reviews. All rights Genetic influences on brain morphology and IQ are well studied. A
reserved variety of sophisticated brain-mapping approaches relating genetic
First published online as a influences on brain structure and intelligence establishes a regional
Review in Advance on distribution for this relationship that is consistent with behavioral
January 14, 2005 studies. We highlight those studies that illustrate the complex corti-
0147-006X/05/0721- cal patterns associated with measures of cognitive ability. A measure
0001$20.00 of cognitive ability, known as g, has been shown highly heritable
across many studies. We argue that these genetic links are partly
mediated by brain structure that is likewise under strong genetic
control. Other factors, such as the environment, obviously play a
role, but the predominant determinant appears to genetic.

1
 AR245-NE28-01 ARI 19 May 2005 10:1

ture (environment) on the brain, and how do

Contents these affect intelligence?
Structural imaging of total brain gray and
INTRODUCTION . . . . . . . . . . . . . . . . . 2
white matter volumes is perhaps the most
INTELLIGENCE . . . . . . . . . . . . . . . . . . 3
obvious approach to correlate brain measures
Fluid and Crystallized Intelligence 5
with general intelligence. Brain structure
BRAIN MAPPING . . . . . . . . . . . . . . . . . 5
measured from MRI correlates with intelli-
Maps of Brain Structure . . . . . . . . . . 6
gence test scores as total brain volume (Gignac
Registration and Mapping . . . . . . . . 6
et al. 2003), as do the volumes of individual
Cortical Pattern Matching . . . . . . . . 6
lobes and aggregate gray and white matter
Statistical Maps . . . . . . . . . . . . . . . . . . 9
volumes (Posthuma et al. 2002). The quest for
GENETIC INFLUENCES
better specificity regarding regional correla-
ON BRAIN STRUCTURE . . . . . . 9
tions of brain structure with intelligence has
Related MRI Studies . . . . . . . . . . . . . 10
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

required more sophisticated analytic tech-

Candidate Genes and Brain
niques to achieve sufficient sensitivity. Voxel-
Function . . . . . . . . . . . . . . . . . . . . . . 11
by University of Sussex Library on 11/10/06. For personal use only.

based morphometry, where voxels belonging

Specific Genes and Brain Structure 11
to an area in the brain are counted and ana-
HERITABILITY OF
lyzed (Ashburner & Friston 2000, Haier et al.
INTELLIGENCE . . . . . . . . . . . . . . . 13
2004), and surface-based approaches, where
Environmental Influences on
three dimensional (3D) models of brain
Intelligence . . . . . . . . . . . . . . . . . . . 14
structures are compared across subjects
Gene x Environment Correlations . 15
(Thompson et al. 2001a), have each demon-
BRAIN STRUCTURE AND
strated regional differences in relationships
INTELLIGENCE . . . . . . . . . . . . . . . 16
to IQ.
Environmental Influences on Brain
That there is a clear relationship between
Structure. . . . . . . . . . . . . . . . . . . . . . 16
intelligence and regional brain volumes does
CONCLUSION . . . . . . . . . . . . . . . . . . . . 17
not shed light on why there are differences
across individuals. Heritability of gray matter
density (Thompson et al. 2001a) and familial
contributions to brain morphology in general
INTRODUCTION have been demonstrated repeatedly (Baaré
The relationship between genetics, brain et al. 2001a, Pfefferbaum et al. 2001). Studies
structure, and intelligence is an age-old of healthy twins (Posthuma et al. 2002) and
polemic evident in such diverse disciplines cohorts of siblings discordant and concordant
as phrenology, sociology, education, neu- for a specific disease (Cannon et al. 2002, Narr
roscience, and politics. Measures of brain et al. 2002) all provide evidence regarding
anatomy, inferred from cranial morphology the heritability of brain morphology. Finally,
(circa the 1800s) or made directly with imag- empirically we (Devlin et al. 1997) showed
ing using magnetic resonance, have been cor- that monozygotic twins reared apart are more
related with a variety of cognitive assessments alike—for many cognitive measures including
(see, e.g., Andreasen et al. 1993, McDaniel & IQ—compared with fraternal twins raised to-
Nguyen 2002). Numerous reviews of the lit- gether. This underscores the relevance of ge-
erature (Herrnstein & Murray 1994, Jensen netic factors in shaping intelligence and brain
1998)—in addition to personal experience— structure.
lead one to conclude, perhaps heretically, that Interactions with the environment also
we are not all created equal. But the ques- contribute to differences in brain mor-
tion still deserves attention: What are the rel- phology. Several animal studies show that
ative influences of nature (genetics) and nur- environmental stimulation can alter synaptic

2 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

densities in the cortex of rodents reared in logical age by 100 to obtain an intelligence
impoverished versus enriched environments quotient (or IQ—a term coined by American
(Greenough et al. 1970, Diamond 1988). scientist Lewis Terman), with scores over 100
Spearman’s g:
Furthermore, animals maintained in enriched being above average. IQ tests, among them quantified general
environments were better problem solvers the Army Alpha and Beta Tests, were subse- cognitive ability
(but not in all tests) than those not main- quently adopted by the U.S. army to help as- (intelligence); basic
tained in enriched environments (Forgays & sign jobs to vast numbers of recruits; nearly general factor of
mental ability
Forgays 1952). Thus it is clear that several, in- two million American men had taken these
terrelated factors influence cognitive function tests by 1919. Lewis Terman at Stanford Uni-
in general, and intelligence specifically. versity subsequently adapted the Binet-Simon
Here we examine the recent application of tests for the American school curriculum and
sophisticated brain-mapping approaches re- published the Revised Stanford-Binet Intel-
lating genetic influences on brain structure ligence Tests in 1937, 1960, and 1985. IQ
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

and intelligence. We highlight those studies tests began to be widely used in schools af-
that illustrate the complex cortical patterns ter World War I, largely to predict academic
by University of Sussex Library on 11/10/06. For personal use only.

associated with measures of cognitive abil- potential and to assign children to suitable
ity. Drawing on work with cohorts of subjects classes according to intellectual ability. Tra-
at risk for several genetically linked diseases, ditional intelligence tests and scholastic apti-
twins, and observations during brain matura- tude tests (SAT) remain a key part of college
tion and degeneration with age, we charac- admissions to this day. Among the tests still in
terize this interesting and important basis for use is the Wechsler Adult Intelligence Scale
human diversity. (WAIS). On the basis of work by psycholo-
gist David Wechsler in the 1930s, the WAIS
(and its counterpart for assessing children—
INTELLIGENCE the WISC) provides separate scales for verbal,
Intelligence has several meanings, largely performance, and total IQ. These scales are
based on the context in which the term is often used to assist with psychiatric diagnosis.
used. Generally referring to competence and In psychometric research, statistical anal-
accomplishment, in neuroscience intelligence ysis can distill from multiple tests a measure
is typically referred to as general cognitive of mental ability that is independent—as far
ability and quantified as Spearman’s g—after as possible—of the subject matter of the tests.
its first proponent, Charles Spearman (1927), In computing the g factor, for example, factor
the statistician who developed factor analy- analysis isolates a component of intellectual
sis. Many psychometric and twin studies have function that is common to multiple cogni-
used this cognitive measure to quantify intel- tive tests, but not specific to the task being
lectual function. performed. IQ tests come in different forms,
Intelligence testing began in 1897 with but they typically assess visuospatial, deduc-
the work of the French psychologist Alfred tive, semantic, and symbolic reasoning abil-
Binet, who, together with Theodore Simon, ity. Specific subtests may evaluate a subject’s
developed tests to identify children who ability to perform inferences, to detect simi-
needed special remedial teaching. By devel- larities and differences in geometrical patterns
oping norms for mental ability at each age, or word patterns, and to process complex in-
Binet could quantify whether a child was formation quickly and accurately.
ahead of or behind his peers, and by how People differ substantially in their per-
many years. German psychologist Wilhelm formance on these tests, but those who do
Stern noted that being a year ahead at age well on one test tend to do well on others.
5 was more significant than at age ten, so he The high correlations among scores on tests
multiplied the ratio of mental age to chrono- of spatial relations, logic, vocabulary, picture

[Link] • Genetics of Brain Structure and Intelligence 3

 AR245-NE28-01 ARI 19 May 2005 10:1

completion, and even reaction times sup- have typically pointed to physiological param-
port the notion that there may be an overar- eters in the brain that are correlated with g, in-
ching skill that underlies intellectual ability, cluding reaction times, nerve conduction ve-
rather than many distinct and independent locity, or cerebral glucose metabolism during
abilities. Scores on a range of tests can be problem solving (Haier et al. 1988). Other
factor analyzed to give g, a single summary brain-based correlates of g have been observed
measure of cognitive ability. g is composed in recent MRI studies showing that differ-
of a small number of (non-independent) sub- ences in frontal gray matter volumes correlate
factors representing more specific abilities with g (p < 0.0044; p < 0.0176 after correction
(Carroll 1993, Deary 2001), but each of these for multiple tests; Thompson et al. 2001a; see
correlates closely with g. One of the best also Haier et al. 2004).
tests for measuring “pure g” is thought to be A more modular view, to some extent im-
Raven’s Progressive Matrices, a nonverbal test plicit in brain-mapping studies, interprets in-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

of inductive reasoning. telligence as reflecting multiple abilities that

The validity of g as a single, unitary mea- may have anatomically distinct biological sub-
by University of Sussex Library on 11/10/06. For personal use only.

sure of intelligence has been hotly debated strates in the brain. Functional MRI, for ex-
by its advocates and detractors (Jensen 1969, ample, can be used to build a more mechanis-
Brand 2001, in favor; see Gould 1996, Kamin tic model of intelligence because it can localize
1997 for contrary views). Most psychometric brain systems involved during cognitive tasks.
researchers agree that the g factor is sensi- The activation of specific neural systems in
tive to individual differences in abilities to the frontal and parietal lobes correlates with
learn, reason, and solve problems. It predicts g, which suggests that these regions interact
scholastic achievement, employment, lifetime to contribute to g (Prabhakaran et al. 1997,
income, and even health-related parameters 2001; Duncan et al. 2000; Gray et al. 2002).
such as life expectancy (Gottfredson 1997). A contrary view of intelligence holds that
The ethics and validity of using IQ tests to important intellectual abilities are poorly as-
predict educational potential, and in college sessed or entirely missed by standardized in-
admissions and recruitment decisions, are telligence tests. Sternberg (1999) proposed a
still somewhat controversial. In the 1960s, triarchic theory of intelligence, in which prac-
many boards of education rejected IQ testing tical and creative intelligence are regarded on
because of concerns about possible cultural par with analytic skills. For Sternberg, ana-
biases in test questions, and there was a lytic intelligence denotes one of three primary
general backlash against psychometric testing intellectual skills, namely one that is similar
in admissions and hiring decisions, a political to the g factor—the ability to recognize and
trend that has been reversed somewhat today. apply logical relations. Equally fundamental,
From a scientific standpoint, some argue however, are practical intelligence, which de-
that the basic general factor of mental ability notes pragmatic and social skills, and creative
(g) can explain performance variations on in- intelligence, or the ability to come up with
dividual mental tests (Spearman 1927, Jensen imaginative solutions to problems rather than
1998). Most mental ability tests correlate with applying familiar logical rules or book knowl-
g, and the degree to which they do has been edge. Social or emotion-related abilities have
termed their g-loading [analogous to an oc- also been argued to be essential ingredients in
tane rating for gasoline (Jensen 1980)]. Per- mental function (Salovey et al. 2002).
formance variations on different tasks may A still broader view of intelligence
therefore depend on how much each task has been popularized by Gardner (2000).
draws on a general cognitive process under- Gardner posits at least seven types of
lying mental ability (the unitary intelligence intelligence (mathematical, spatial, musi-
theory). Advocates of unitary intelligence cal, bodily-kinesthetic, intra-personal, and

4 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

interpersonal). The case for multiple intel- widely applied to map brain activation in a
ligences has been supported by studies of variety of psychiatric and neurological dis-
brain lesions that cause very specific neuro- orders. Brain activation can be examined
Brain map: a
logical deficits but leave many cognitive abil- noninvasively while subjects perform specific relationship between
ities intact (e.g., speech or visuospatial skills). tasks or cognitive assessments [see Cabeza points in a
Gardner considers that proponents of the g & Nyberg 2000, for a review of studies us- coordinate space and
factor confuse intelligence with a highly spe- ing positron emission tomography (PET) and features or
annotations
cific type of scholastic performance. functional MRI (fMRI)]. However, the cause
describing brain
The most negative view of IQ testing is of individual differences in hemodynamic- structure and/or
that inherent biases make cognitive tests a based functional measures—their heritability, function
poor measure of individual competence. De- for example—is largely unknown. Although
tractors of IQ tests say that the ability to an- it is clear that functional imaging provides
swer some questions may depend on a person’s the link between the anatomic maps and cog-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

upbringing or cultural background, and that nitive measures, the present paucity of data
the questions assume a familiarity or agree- using fMRI may be due to the vagaries of neu-
by University of Sussex Library on 11/10/06. For personal use only.

ment with certain cultural norms. Situational rovascular coupling, the variability of the re-
factors may also impair performance (Steele & sponse or the limitations of instrumentation,
Aronson 1995; Gould 1996, p. 166; Baumeis- and protocols to date. Numerous efforts are
ter et al. 2002; Schmader & Johns 2004). underway to collect sufficient baseline data
to attempt to improve sensitivity. The Inter-
national Consortium for Brain Mapping has
Fluid and Crystallized Intelligence developed a battery of fMRI tests (Mazziotta
Even among psychometric researchers who et al. 2001) that exhibit stable baseline across
agree that there is a general factor in cogni- subjects. These ultimately can be used to
tive ability, crystallized and fluid intelligence normalize other more cognitively challeng-
are often distinguished (Cattell 1971). Crys- ing behavioral tasks in much the same way
tallized intelligence refers to the large body as structural scans are normalized for place-
of factual knowledge that an individual ac- ment into an atlas. Similarly, the Bioinfor-
cumulates over his/her lifespan, including, matics Research Network has developed a
for example, vocabulary. This ability to ap- series of tasks ([Link] specif-
ply knowledge to solve problems is largely ically designed to normalize across popula-
determined by education and experience, and tions of schizophrenic patients and their nor-
increases with age. Fluid intelligence, how- mal matched controls. Thus, it is likely that in
ever, refers to analytical reasoning ability, as the near future we will see many more studies
well as memory and information processing examining genetic influences on brain func-
speed, and it declines somewhat with age. The tion using fMRI.
fluid component of intelligence is thought to Structural brain mapping, in contrast, has
be largely genetically determined, however. already shown specific patterns related to in-
In addition, fluid intelligence is strongly as- telligence (see above) and, as with other brain-
sociated with working memory (Prabhakaran mapping studies, can provide the anatomic
et al. 2001) and is correlated with activation framework to achieve improved sensitivity in
during cognitively demanding tasks observed functional studies. For this reason, we next re-
with functional MRI (Gray et al. 2003). view the steps required to create brain maps.
These include maps of morphologic features,
such as the 3D distribution of gray and white
BRAIN MAPPING matter in the brain, and statistical maps that
Because of its promise in localizing brain compile these maps from whole populations.
function, functional brain imaging has been To examine sources of morphological and

[Link] • Genetics of Brain Structure and Intelligence 5

 AR245-NE28-01 ARI 19 May 2005 10:1

functional variability across subjects, we also tween groups, or in this case to reveal where
review methods that combine imaging and ge- individual differences in brain structure de-
Cortical pattern
matching: encoding netic statistics to compute genetic influences pend on genetic factors.
both gyral patterning on the brain (Thompson et al. 2001a, 2003).
and gray matter This combination as in correlation creates an
variation important link between genetics, brain mea- Registration and Mapping
sures, and intelligence, shedding light on the D MRI scans are first rotated and scaled
systems involved in cognition and which fac- to match a standardized brain template in
tors affect their function. stereotaxic space. This template may be either
Atlases to regionally chart the degree and an average intensity brain dataset constructed
extent of individual variations in brain struc- from a population of young normal subjects
ture require detailed descriptions of anatomy (Mazziotta et al. 2001) or one specially
to achieve a morphometric comparison rather constructed to reflect the average anatomy of
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

than the volumetric comparisons described a subgroup of defined homogeneous subjects

above. To create atlases that contain detailed (e.g., Mega et al. 2000, Thompson et al.
by University of Sussex Library on 11/10/06. For personal use only.

representations of anatomy, we have devel- 2000, Janke et al. 2001; see these papers for
oped model-driven algorithms that use ge- a discussion of disease-specific templates).
ometrical surfaces to represent structures in Once aligned, a measure of the brain scaling
the brain, such as the cortex, hippocampus, imposed is retained as a covariate for statis-
ventricles, and corpus callosum (Thompson & tical analysis. A tissue classification algorithm
Toga 1996, 2003). Anatomic models provide then segments the image data into regions
an explicit geometry for individual structures representing gray matter, white matter, cere-
in each scan, such as landmark points, curves, brospinal fluid (CSF), and nonbrain tissues.
or surfaces. These modeling approaches can Because the digital models reside in the same
also answer the following questions: How stereotaxic space as the atlas data, surface and
does anatomy differ across subjects and be- volume models stored as lists of vector coordi-
tween groups? What is the degree of indi- nates are amenable to digital transformation,
vidual variability in anatomy, and how do as well as geometric and statistical measure-
these differences link with cognitive mea- ment (Mega et al. 2000, Narr et al. 2003,
sures? What are the sources of these varia- Thompson et al. 2004, Zhou et al. 1999). The
tions, and to what degree are they influenced underlying 3D coordinate system is central to
by genes and environment? Brain mapping all atlas systems because it supports the linkage
can provide answers to these and other ques- of structure models and associated image data
tions; the answers are typically displayed in with spatially indexed neuroanatomic labels.
the form of a brain map.

Cortical Pattern Matching

Maps of Brain Structure MRI scans have sufficient resolution and tis-
First we consider the analysis steps required sue contrast, in principle, to track cortical gray
to compute the patterns of genetic influences and white matter differences in individual sub-
on brain structure, using a database of brain jects. This affords the opportunity to measure
MRIs from twins (Thompson et al. 2001). regional degrees of heritability and establish
The process can be conceived as shown in structural and even gyral/sulcal relationships
Figure 1, where a sequence of image anal- with specific cognitive measures. Even so, ex-
ysis steps are applied to brain MRI scans from treme variability in gyral patterns confounds
multiple subjects. The goal of such an analy- efforts (a) to compare between groups and (b)
sis is typically to create color-coded maps of to determine the average profile of patterns
brain regions with structural differences be- within a group. Cortical pattern matching

6 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

methods (detailed further in Figure 2) ad-

dress these challenges. They encode both gy-
ral patterning and gray matter variation. This
can substantially improve the statistical power
to localize effects of genes and environmental
factors on the brain. These cortical analyses
can also be used to measure cortical asymme-
tries (Narr et al. 2001, Sowell et al. 2001).
Briefly, a 3D geometric model of the corti-
cal surface is extracted from the MRI scan and
flattened to a two-dimensional planar format
(to avoid making cuts, a spherical topology
can be retained; Fischl et al. 2001; Thompson
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

et al. 1997, 2002). A complex deformation, or

warping transform, is then applied that aligns
by University of Sussex Library on 11/10/06. For personal use only.

the sulcal anatomy of each subject with an

average sulcal pattern derived for the group.
To improve feature alignment across subjects,
all sulci that occur consistently can be digi-
tized and used to constrain this transforma-
tion. As far as possible, this procedure ad-
justs for differences in cortical patterning and

Figure 1
Analyzing cortical data. The schematic shows a
sequence of image-processing steps that can be
used to map how development and disease, or
genetic factors, affect the cortex. Regions can also
be identified where brain variation is linked with
intelligence, specific cognitive measures, or
clinical measures. The steps include aligning MRI
data to a standard space, tissue classification, and
cortical pattern matching, as well as averaging and
comparing local measures of cortical gray matter
volumes across subjects. (These procedures are
detailed in the main text). To help compare
cortical features of subjects whose anatomy differs,
individual gyral patterns are flattened and aligned
with a group average gyral pattern. Group
variability and cortical asymmetry can also be
computed. Correlations can be mapped between
disease-related gray matter deficits and genetic risk
factors. Maps may also be generated visualizing
linkages between genes and morphology, cognitive
scores, and other effects. The only steps here that
are not currently automated are the tracing of sulci
on the cortex. Some manual editing may also be
required to assist algorithms that delete dura and
scalp from images, especially if there is very little
CSF in the subarachnoid space.

[Link] • Genetics of Brain Structure and Intelligence 7

 AR245-NE28-01 ARI 19 May 2005 10:1
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]
by University of Sussex Library on 11/10/06. For personal use only.

Figure 2
Right lateral and top views of the dynamic sequence of gray matter maturation over the cortical surface.
The side bar shows a color representation in units of gray matter volume. Constructed from MRI scans of
healthy children, these maps illustrate 15 years of brain development (ages 5–20; data from Gogtay et al.
2004). Red indicates more gray matter, blue less gray matter. Gray matter wanes in a back-to-front wave
as the brain matures and neural connections are pruned. Areas performing more basic functions mature
earlier; areas for higher-order functions (emotion, self-control) mature later. The prefrontal cortex,
which handles reasoning and other executive functions, emerged late in evolution and is among the last
to mature. Intriguingly, this sequence of brain changes is reversed in Alzheimer’s disease (see Figure 4).

shape across subjects. Cortical measures can each cortical point (Wright et al. 1995;
then be compared across subjects and groups. Bullmore et al. 1999; Sowell et al. 1999,
Sulcal landmarks are used as anchors 2003; Ashburner & Friston 2000; Mummery
because homologous cortical regions are et al. 2000; Rombouts et al. 2000; Baron et al.
better aligned after matching sulci than by 2001; Good et al. 2001; Thompson et al.
just averaging data at each point in stereotaxic 2001a,b). Given the large anatomic variability
space (see, e.g., fMRI studies by Rex et al. in some cortical regions, high-dimensional
2001; Zeineh et al. 2001, 2003; Rasser et al. elastic matching of cortical patterns
2004). Given that the deformation maps (Thompson et al. 2000, 2001b) is used
associate cortical locations with the same to associate measures of gray matter density
relation to the primary folding pattern from homologous cortical regions first across
across subjects, a local measurement of gray time and then also across subjects. One ad-
matter density is made in each subject and vantage of cortical matching is that it localizes
averaged across equivalent cortical locations. differences relative to gyral landmarks; it also
To quantify local gray matter, gray matter averages data from corresponding gyri, which
density can be measured to compare the would be impossible if data were only linearly
spatial distribution of gray matter across mapped into stereotaxic space. The effects of
subjects. This gray matter density measures age, gender, zygosity, g, and other measures
the proportion of gray matter in a small on gray matter can be assessed at each cortical
region of fixed radius (15 mm) around point (see Figure 3 for an example in twins).

8 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

Statistical Maps
An algorithm then fits a statistical model,
such as the general linear model (GLM)
(Friston et al. 1995), to the data at each cor-
tical location. This model results in a vari-
ety of parameters that characterize how gray
matter variation is linked with other variables.
The significance of these links can be plot-
ted as a significance map. A color code can
highlight brain regions where linkages are
found, allowing us to visualize the strength
of these linkages. In addition, estimated pa-
rameters can be plotted, such as (a) the lo-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

cal rates of gray matter loss with aging (see

Figure 2 for example) at each cortical loca-
by University of Sussex Library on 11/10/06. For personal use only.

tion (e.g., as a percentage change per year),

(b) regression parameters that identify heri-
tability, and even (c) nonlinearities in the rates
of brain change over time (e.g., quadratic re-
gression coefficients; Sowell et al. 2003). In
principle, any statistical model can be fitted,
including genetic models that estimate ge-
netic or allelic influences on brain structure
(Thompson et al. 2003). Finally, permutation
testing is typically used to ascribe an overall
significance value for the observed map. This
adjusts for the fact that multiple statistical tests
are performed when a whole map of statistics
is visualized. Subjects are randomly assigned Figure 3
to groups, often many millions of times on a
Heritability of gray matter. Intraclass correlation in gray matter density
supercomputer. A null distribution is built to gi,r (x) for groups of identical and fraternal twins, after cortical pattern
estimate the probability that the observed ef- matching [giving maps rMZ (φ, θ ) and rDZ (φ, θ ), in Figure 3a]. In
fects could have occurred by chance, and the behavioral genetics, a feature is heritable if rMZ significantly exceeds rDZ .
result is reported as a significance value for An estimate of its heritability h2 can be defined as 2(rMZ -rDZ ), with
standard error SE2 (h2 ) = 4[((1 − rMZ 2 )2 /nMZ ) + ((1 − rDZ 2 )2 /nDZ )].
the overall map.
Figure 3b shows a heritability map computed from the equation

h 2 (φ, θ ) = 2(r MZ (φ, θ ) − r DZ (φ, θ )).

GENETIC INFLUENCES
ON BRAIN STRUCTURE Regions in which significant genetic influences on brain structure are
detected are shown in the significance map [Figure 3b (right)] p[h2 (φ, θ )].
Statistical maps of cortical anatomy can also Genetic influences on brain structure are pronounced in some frontal and
be used to reveal genetic influences on brain temporal lobe regions, including the dorsolateral prefrontal cortex and
morphology. Figure 3 shows intersubject temporal poles [denoted by DLPFC and T in Figure 3b (left)]. These
effects were confirmed by assessing the significance of the effect size of h2
variations in cortical gray matter distribution
by permutation (this involved repeated generation of null realizations of an
and their heritability. In a study of genetic h2 -distributed random field; for details of these permutation methods, see
influences on brain structure (Thompson Thompson et al. 2004).
et al. 2001a), we began by computing

[Link] • Genetics of Brain Structure and Intelligence 9

 AR245-NE28-01 ARI 19 May 2005 10:1

the intraclass correlations in gray matter Related MRI Studies

(Figure 3a,b) in groups of monozygotic
The high heritability of gray matter volumes,
(MZ) and dizygotic (DZ) twins. Forty healthy
visualized in Figure 3, corroborates earlier
normal subjects, consisting of 10 MZ and 10
studies that revealed strong genetic influ-
age- (48.2 ± 3.4 years) and gender-matched
ences on brain structure. Studies of healthy
DZ twin pairs were drawn from a twin
twins suggest that overall brain volume is
cohort consisting of all the same-sex twins
highly genetically influenced (Bartley et al.
born in Finland between 1940 and 1957,
1997, Tramo et al. 1998). Volumes of some
inclusive, in which both members of each
brain structures are also under strong ge-
pair were alive and residing in Finland as of
netic control, including the corpus callosum
1967 (Kaprio et al. 1990). Consistent with
(Oppenheim et al. 1989, Pfefferbaum et al.
earlier studies reporting the high heritabil-
2000) and ventricles, whereas gyral patterns
ity of brain volume (Bartley et al. 1997),
are much less heritable (Bartley et al. 1997,
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

MZ within-pair gray matter differences

Biondi et al. 1998). Bartley et al. (1997)
were almost zero (intraclass r ∼ 0.9 and
reported a 94% heritability for brain vol-
higher, p < 0.0001 corrected; Figure 3a)
by University of Sussex Library on 11/10/06. For personal use only.

ume (identical twin correlation = 0.95, p <

in a broad anatomical band encompassing
0.00,001; fraternal twin correlation = 0.35,
frontal, sensorimotor, and linguistic cortices,
p = 0.09), on the basis of structural equa-
including Broca’s speech and Wernicke’s lan-
tion modeling in 10 MZ and 9 DZ pairs
guage comprehension areas. MZ twins are ge-
scanned with MRI. In elderly twins, Sullivan
netically identical, so any regional differences
et al. (2001) found that the volume of the hip-
must be attributed to environmental effects or
pocampus was less heritable (h2 = 0.4) than
gene-environment interactions. Meanwhile,
that of the adjacent temporal horns (h2 = 0.6),
sensorimotor and parietal occipital, but not
corpus callosum (h2 = 0.8), and intracranial
frontal, territories were significantly more
volume (h2 = 0.8). They suggested that en-
similar in DZ twins than random pairs. Affin-
vironmental differences, perhaps interacting
ity was greatest in the MZ pairs, suggesting
with genetic differences, may exert especially
a genetic continuum in the determination
strong or prolonged influences on hippocam-
of brain structure. In behavioral genetics, a
pal size, consistent with its lifelong plasticity
feature is heritable if the identical twin corre-
and fundamental role in learning and mem-
lation exceeds the fraternal twin correlation.
ory. A lower heritability figure for hippocam-
Comparisons of MZ and DZ correlations
pal size is consistent with its role in memory
suggested that frontal, sensorimotor, and
encoding, its vulnerability to plasma cortisol
anterior temporal cortices were under signif-
levels, and its plasticity in later life (Maguire
icant genetic control (p < 0.05, rejecting the
et al. 2000; see also Lyons et al. 2001, for a
hypothesis that h2 = 0; one-tailed). Middle
related MRI study in monkeys). In a similar
frontal regions, near Brodmann areas 9 and
vein, Baaré and colleagues (2001b) found that
46, displayed a 90%–95% genetic determi-
individual differences in lateral ventricle vol-
nation of structure (i.e., h2 ∼ 0.90–0.95).
ume were best explained by a structural equa-
Many regions are under tight genetic control
tion model containing common (58%) and
(bilateral frontal and sensorimotor regions,
unique (42%) environmental factors, indicat-
p < 0.0001; Figure 3b). Heritability es-
ing genes to be of little or no influence. The
timates were comparable with twin-based
same authors found that genetic factors almost
estimates for the most highly genetically
entirely accounted for individual differences
determined human traits, including finger-
in whole brain (90%), gray (82%), and white
print ridge count (h2 = 0.98), height (h2 =
(88%) matter volume, in a study based on a
0.66–0.92), and systolic blood pressure
sizeable sample of 54 MZ and 58 DZ twin
(h2 = 0.57).

10 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

pairs and 34 of their full siblings. In their mul- cific genes whose variations are linked with
tivariate analysis of body height and volumes brain structure and function. For example,
of gray matter, white matter, and the intracra- recent functional imaging work has shown
nial space, Baaré et al. (2001b) noted that a that a polymorphism in the human brain-
large part of the genetic influences were com- derived neurotrophic factor (BDNF) gene is
mon to the three brain measures, and a smaller associated with poor memory performance
part was shared with height. Some genes and with working memory activation mapped
may therefore have a general effect on the with fMRI (Egan et al. 2001, 2003). Diamond
brain, whereas other genes may affect specific et al. (2004) have recently shown strik-
volumes. ing specificity of COMT (catechol-o-
More recently, Pfefferbaum et al. (2001) methyltransferase) polymorphisms to some
used diffusion imaging, which is sensitive to but not other prefrontal cortex–dependent
myelination levels and fiber orientation, to tasks in children (Diamond et al. 2004).
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

quantify the microstructure of the corpus cal- Heritability of cognitive function is

losum in 15 MZ and 18 DZ pairs. They certainly complex and difficult to dissociate
by University of Sussex Library on 11/10/06. For personal use only.

found that anterior interhemispheric con- from environmental factors, among other
necting pathways, in the callosal genu, were influences. A recent review of candidate genes
more susceptible than splenial pathways to contributing to human cognition lists more
environmental influences, perhaps reflecting than 70 suspects (Morley & Montgomery
the prolonged maturation of the frontal cor- 2001). However, examining this list and
tex well into adulthood (Sowell et al. 1999, relating them to spatial patterns of gene ex-
Gogtay et al. 2004). Using bivariate genetic pression or segmenting those that are related
modeling, these authors also noted that in- to neuroanatomical regions involved in cog-
tracranial volume and corpus callosum area nition results in a far more tractable problem.
were tightly correlated, a correlation due en- For example, the prefrontal cortex, an area
tirely to shared genetic effects between these highly involved in cognition (see Winterer &
two brain structures. Wright et al. (2002) ex- Goldman 2003, among others), links only 3 of
tended this design to parcellate 92 regional the 70 identified by Morley & Montgomery
gray matter volumes in 10 MZ and 9 DZ twin (2001).
pairs scanned with MRI. Interregional rela-
tionships were summarized by principal com-
ponent analysis of the resulting genetic corre- Specific Genes and Brain Structure
lation matrix. This analysis identified shared With current databases of structural brain im-
genetic effects on the frontal-parietal cortices ages, there is now significant power to assess
and bilateral temporal cortex and insula. As the effects of specific candidate genes on brain
the size and scope of these studies increases, structure. The easiest context for evaluating
decomposition of the genetic correlation ma- these genetic influences is to examine alleles
trix is likely to be a key exploratory tool to overtransmitted to individuals with specific
identify supraregional brain systems (Wright diseases such as dementia or schizophrenia.
et al. 1999), which share common genetic in- Using statistical maps to visualize brain sys-
fluences and which may cut across conven- tems that are at genetic risk, brain images can
tional anatomic boundaries. also provide a quantitative index of disease li-
ability in individuals at increased genetic risk
for disease (Cannon et al. 2002, Narr et al.
Candidate Genes and Brain Function 2002, Thompson et al. 2003).
Heritable aspects of brain structure are For example, the apolipoprotein E4
important to identify because they provide (ApoE4) allele is found in 38% of all Alzhei-
endophenotypes to guide the search for spe- mer’s disease patients, but in only 15% of

[Link] • Genetics of Brain Structure and Intelligence 11

 AR245-NE28-01 ARI 19 May 2005 10:1

ductions in gray matter density in healthy

relatives of patients, relative to a population
of normal controls who are not related to a
schizophrenia patient (Cannon et al. 2002).
This pattern of brain structure is intriguing
because the observed deficit is associated with
the degree of genetic risk (i.e., greater in MZ
than DZ twins of a patient). There are also
Figure 4
genetically mediated structural deficits in the
hippocampus, and in the corpus callosum, in
Patterns of brain structure associated with genetic risk for Alzheimer’s
disease. Brain structure is often significantly different from normal in schizophrenic patients and in their healthy
subjects who are at genetic risk for a brain disorder but who are at-risk relatives (Narr et al. 2003). But,
cognitively normal. Typical MRI scans are shown from healthy elderly unlike the ApoE4 example, the specific genes
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

subjects with zero, one, and two ε4 alleles of the ApoE gene, each of involved in schizophrenia are currently un-
which confers increased risk for late-onset Alzheimer’s disease (data
known. By correlating alleles overtransmitted
courtesy of G. Small, UCLA Center on Aging). Note the hippocampal
by University of Sussex Library on 11/10/06. For personal use only.

atrophy and ventricular enlargement in those at risk. The ε3 allele is most to patients with these structural variations,
prevalent and is considered normal. Subjects at genetic risk may display specific polymorphic genetic loci that medi-
metabolic and structural deficits before overt cognitive symptoms appear, ate these deficits may be identified (Cannon
which suggests that genetic and imaging information may identify et al. 2003). In this respect, brain mapping
candidates for early treatment in dementia (Small et al. 2000).
can assist in the search for genes involved in
a disorder by generalizing genetic methods
controls (Roses 1996). As shown in Figure 4, such as Haseman-Elston mapping to brain
medial temporal brain structure shows pro- images.
found atrophic changes in healthy ApoE4- Conversely, brain mapping may also help
positive individuals, and the ventricles ex- to establish the scope of brain abnormalities
pand, even before overt cognitive deficits can in diseases in which the genetic causes are al-
be detected. However, some brain regions ready well understood. Williams syndrome,
are comparatively protected (e.g., frontal cor- for example, results from a known genetic
tices in ApoE4 subjects with Alzheimer’s deletion in the 7q11.23 chromosomal region
disease; Geroldi et al. 1999, Hashimoto (Korenberg et al. 2000). The syndrome is as-
et al. 2001). Because neuroprotective drugs sociated with disrupted cortical development
are effective in early dementia (Lehtovirta and mild-to-moderate mental retardation
et al. 1995, 2000; Small et al. 2000) there (Bellugi et al. 2000). By statistically averaging
is interest in associating patterns of brain cortical thickness maps from 166 brain hemi-
change with specific genetic markers, es- spheres and comparing Williams syndrome
pecially if these patterns predict imminent patients with healthy controls, we recently
disease onset among individuals at genetic identified a sharply delimited region of lan-
risk. guage cortex with increased cortical thickness,
Gray matter deficits are also found in revealing the cortical territory affected by the
healthy first-degree relatives of schizophre- genetic deletion (Thompson et al. 2004). This
nia patients. Because these relatives are at selective augmentation of brain structure may
increased genetic risk for developing the underlie the relative strengths patients exhibit
disorder themselves, there is great interest in in language function. These maps also refine
understanding what factors promote or resist our knowledge of how the genetic deletion
disease onset or progression (Weinberger impacts the brain, providing new leads for
et al. 1981, Suddath et al. 1990, Cannon molecular work on Williams syndrome and a
et al. 2002, Thompson et al. 2003). Figure 5 link between genetic and behavioral findings
shows brain regions with significant re- in the disorder.

12 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]
by University of Sussex Library on 11/10/06. For personal use only.

Figure 5
Patterns of brain structure associated with genetic risk for schizophrenia and with genetic deletion in
Williams syndrome. Top row, last panel: Statistical combinations of brain scans from at-risk relatives of
schizophrenia patients show that relatives have abnormally reduced gray matter density in the frontal
cortex (green; data adapted from Cannon et al. 2002). In a twin design, statistical comparison of
schizophrenia patients with their healthy identical twins reveals regions (red) in parietal and frontal
cortices that have reduced gray matter density in disease. The source of these differences must be
environmental in origin because the differences are based on averages of maps that subtract data from
genetically identical twins. Bottom Row: Finally, group average maps of cortical thickness are shown for
43 subjects with Williams syndrome (a), and 40 matched healthy controls (b), which revealed that
perisylvian language cortex is 10% thicker in the patients. Williams syndrome results from a known
genetic deletion on chromosome 23q11. Composite brain maps such as these can help identify
circumscribed cortical regions whose formation or maturation is influenced by the genetic lesion,
perhaps during gyrogenesis.

HERITABILITY OF bate regarding genetic influences on IQ be-

INTELLIGENCE came increasingly vitriolic after an article
argued that there are racial differences in
Before reviewing some of the brain substrates
intelligence that may be genetic in origin
that correlate with intelligence, it is worth ex-
(Jensen 1969). Most behavioral geneticists
amining the evidence that there are genetic
now agree that heredity plays a role in individ-
influences on intelligence; we argue that these
ual differences in intelligence, but some have
genetic links are partly mediated by brain
argued that group differences in IQ include
structure that is under strong genetic control.
environmental influences or cultural biases in
We review this literature only briefly because
the tests (Lewontin 1975; see Jensen & Miele
it has been examined thoroughly elsewhere
2002, Gray & Thompson 2004, for reviews of
(Herrnstein & Murray 1994, Gould 1996,
arguments on this topic).
Jensen 1998, Pinker 2002). In 1969, the de-

[Link] • Genetics of Brain Structure and Intelligence 13

 AR245-NE28-01 ARI 19 May 2005 10:1

Correlations between related individuals biological rather than purely statistical basis
show that both nature and nurture influ- for g.
ence intelligence. Adopted MZ twins—raised
apart—still correlate 0.72 for intelligence, i.e.,
one twin’s intelligence strongly predicts the Environmental Influences on
other’s, despite their different rearing envi- Intelligence
ronments. This suggests an undeniable ge- Many environmental factors are known to
netic component to intelligence. A popular influence intelligence favorably or adversely
line of attack against this argument states (Ceci & Williams 1997, Neisser 1998). By
that several nongenetic factors could con- comparing identical twins reared apart and
found this association by making MZ twins reared together, effects of different rearing
more similar. For example, identical twins environments can be established. Bouchard
might be adopted into similar homes (selec- et al. (1990) found that growing up in the
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

tive placement). Sharing the same fetal en- same family increased the IQ similarities for
vironment might also make identical twins all types of relatives: Individual IQs correlated
by University of Sussex Library on 11/10/06. For personal use only.

more alike cognitively or perhaps even less more highly with their MZ twins, siblings,
alike (via twin-twin competition for nutrition, and parents (0.86, 0.47, 0.42) if they grew
transfusion effects, and so on). Also, frater- up together than if they did not (0.72, 0.24,
nal twins may inadequately control for the ef- 0.22). Adopted children’s IQs also correlate
fects of shared family environments (see Vogel with their siblings (0.34) and adoptive parents
& Motulsky 1997, Kamin & Goldberger (0.19), so 20%–35% of the observed popula-
2002). tion differences in IQ are thought to be due
Nonetheless, adoption and family studies to differences in family environments. In-
using sophisticated genetic model-fitting have triguingly, these shared family environmental
shown g to be highly heritable across many influences on IQ dissipate once young chil-
studies, even more so than specific cognitive dren leave home: As adults, adoptive relatives
abilities [h2 = 0.62, McClearn et al. (1997), only correlated –0.01 for IQ (McGue et al.
Feldman & Otto (1997); h2 = 0.48, Devlin 1993), showing no lasting influence of shared
et al. (1997); h2 = 0.6–0.8, Finkel et al. (1998), upbringing on IQ. Those environmental
Swan et al. (1990), Loehlin (1989), Chipuer influences on IQ that do last are thought to be
et al. (1990), Plomin & Petrill (1997)]. The experiences that an individual does not share
heritability of intelligence also increases with with others, interpreted broadly to include
age: As we grow older, phenotype reflects the chemical environment in the womb and
genotype more closely. A strictly environ- the multitude of random events in human ex-
mental theory would predict the opposite. perience that are hard to quantify or control.
Some IQ-related genes may not be switched Heritability does not imply inevitability
on until adolescence or adulthood, but a because the environment can determine the
more plausible explanation may be the ex- relative impact of genetic variation (gene x
istence of a gene by environment interac- environment interaction). For example, in a
tion (Boomsma et al. 1999, Rowe & Jacobson recent study of 320 pairs of twins who were
1999). As individuals select or create envi- born in the 1960s and given IQ tests at age
ronments that foster their genetic propensi- 7, Turkheimer et al. (2003) found that envi-
ties throughout life, the genetic differences ronmental factors made a much bigger differ-
in cognition are greatly amplified (Plomin ence in the determination of childhood IQ in
1999). Jensen (1998) hypothesized that the impoverished families relative to those with
more a mental test score correlates with gen- higher socioeconomic status. The heritabil-
eral intelligence, or g, the higher its heri- ity of IQ at the low end of the wealth spec-
tability is. If true, this hypothesis supports a trum was just 0.10 on a scale of zero to one,

14 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

but it was 0.72 for families of high socioeco- ing infancy is still associated with enhanced
nomic status. The importance of environmen- childhood cognitive development (by 2–5
tal influences on IQ was four times stronger IQ points for full-term infants and 8 points
in the poorest families than in the higher sta- for those with low birth weight; Drane &
tus families, which suggests that nature mat- Logemann 2000).
ters more on the high end of socioeconomic
status and nurture matters more on the low
end. The genetic contribution to intelligence Gene x Environment Correlations
therefore differs in different environments— The significant influence of heredity on IQ
a caveat against general inferences based on has been misinterpreted to imply that there is
heritability data. The same could be said little point trying to educate or be educated,
of certain physical attributes such as height, or that IQ is somehow impervious to change.
which is heritable when nutrition is not This is a fallacy because many environmen-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

limiting. tal factors, including family rearing environ-

Population-level increases in intelligence ments, socioeconomic status, diet, and school-
by University of Sussex Library on 11/10/06. For personal use only.

test scores have also been observed in re- ing, influence IQ. As noted elsewhere (Plomin
cent decades. Dutch 18-year-old men tested & Kosslyn 2001), gray matter volume may
in 1982 scored 20 IQ points (standard devia- be correlated with intelligence partly because
tion = 15) higher than did 18-year-old men more intelligent individuals seek out mentally
tested in 1952 (Dickens & Flynn 2001). This challenging activities that increase the vol-
widely replicated population-level increase in ume of their gray matter. Such strong gene x
intelligence is known as the Flynn Effect. Be- environment correlations may draw individu-
cause genetic variation remained fairly stable als with higher genetic potential into learn-
over such a short time frame, these relatively ing environments more conducive to intel-
rapid increases are attributed to nongenetic lectual advancement. Gifted individuals might
factors such as improved schooling and tech- either create or evoke situations that further
nology, better access to education, and im- promote their intellectual ability (termed ac-
proved nutrition. There has also been a re- tive and reactive genotype-environment (GE)
duction in some environmental toxins (such as correlation, respectively; Plomin et al. 1977).
lead) and hazards that are detrimental to IQ. These correlations make it impossible to con-
Dickens & Flynn (2001) also proposed pow- ceptually differentiate effects of nature and
erful gene-environment interactions to rec- nurture (Ridley 2003).
oncile the paradox that IQ is highly herita- If environments are not randomly assigned
ble even though average scores have increased to each individual but are, in part, individ-
significantly in recent decades. ually selected on the basis of genetically in-
Positive environmental influences on in- fluenced preferences (GE autocorrelation), it
telligence are hard to identify, in part, be- becomes impossible to discern which genetic
cause of the inevitable confounding of vari- effects act directly on intellectual function and
ables in large-scale epidemiological studies of which result from the action of environmental
cognition. For example, duration of breast- variation causally linked with genetic differ-
feeding during infancy has been associated ences (Block 1995). One form of GE corre-
with higher IQ in a group of more than lation can be estimated explicitly in adoption
2000 children assessed at age 6 (Oddy et al. designs: the environment that parents provide
2003). However, this association has been their offspring (Neale 1997). Active and reac-
contested because it is confounded by mater- tive correlations are more difficult to estimate,
nal age, intelligence, and education, as well as leading to suggestions that the notion of heri-
smoking during pregnancy. After adjusting for tability conflicts with common sense (Sesardic
these confounding factors, breastfeeding dur- 2002).

[Link] • Genetics of Brain Structure and Intelligence 15

 AR245-NE28-01 ARI 19 May 2005 10:1

BRAIN STRUCTURE AND used voxel-based morphometry in two

INTELLIGENCE independent samples to identify substantial
gray matter correlates of IQ. More gray
If specific features of brain structure are under
ApoE4: matter was associated with higher IQ in all
apolipoprotein E4 strong genetic control, investigators should
lobes, underscoring a distributed model of
determine whether any of these features are
BDNF: the neural basis of intelligence. Intriguingly,
brain-derived correlated with intelligence. If so, this corre-
the strongest correlations are typically found
neurotrophic factor lation may not only reveal why IQ has repeat-
between IQ and frontal gray matter volumes
COMT: catechol-o- edly been found to be highly heritable, but
(Thompson et al. 2001a, Haier et al. 2004),
methyltransferase also yield insight into possible neural mecha-
the same brain regions that are under greatest
CSF: cerebrospinal nisms. To help understand this approach, we
genetic control. Frontal brain regions play
fluid first review evidence that brain structure and
a key role in working memory, executive
DZ: dizygotic intelligence are correlated before discussing
function, and attentional processes, and their
evidence for the existence of genetic correla-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

fMRI: functional structure has rapidly expanded in recent

magnetic resonance tions between brain structure and intelligence
primate evolution, consistent with their role
imaging (which means that the same sets of genes
by University of Sussex Library on 11/10/06. For personal use only.

in reasoning and intellectual function.

g: a measure of are implicated in determining both; Posthuma
cognitive ability et al. 2002).
GLM: general A recent meta-analysis (including a total of
1375 subjects) found that total brain volume
Environmental Influences
linear model
and IQ were correlated significantly in all but
on Brain Structure
MZ: monozygotic
1 of 28 MRI studies, with an estimated cor- Neural plasticity in humans may also lead to
PET: positron
emission relation of 0.33 (McDaniel & Nguyen 2002). use-dependent structural adaptation in cere-
tomography This finding implies that ∼10% of the popu- bral gray matter in response to environmen-
SAT: scholastic lation variability in IQ can be predicted from tal demands. At the gross level observable
aptitude test brain volume measures alone. Some studies with MRI, there is already evidence that the
WAIS: Wechsler have quoted slightly higher figures for these human brain may adapt dynamically to re-
Adult Intelligence correlations (e.g., 0.41; Andreasen et al. 1993), flect the cognitive demands of the environ-
Scale and the exact value obtained will depend on ment. Neuroimaging studies have observed
WISC: Wechsler the measurement error of the technique be- structural plasticity after training on diffi-
Intelligence Scale for cause measurement errors will tend to dimin- cult motor tasks such as juggling (Draganski
Children ish any observed correlation (relative to the et al. 2004). Increased hippocampal volumes
true correlation). have also been found in taxi drivers with
Linkages between brain structure and IQ enhanced spatial navigation skills (Maguire
also can be further localized by parcellating et al. 2000). Gaser & Schlaug (2003) also
the brain into subregions or by creating found gray matter increases in motor, audi-
maps of the correlations between gray matter tory, and visual-spatial brain regions when
and IQ. Recently, we found that intellectual comparing professional musicians (keyboard
function (g) was significantly linked with players) with a matched group of amateur mu-
differences in frontal gray matter volumes, sicians and nonmusicians. Brain structure is
which were determined primarily by genetic by no means unchanging even in health. Dy-
factors (Thompson et al. 2001a). Posthuma namic regional changes over the entire life
et al. (2002) extended these findings using span can be mapped (Figure 2), showing a
a cross-twin cross-trait (bivariate genetic) progressive change in cortical volume. The
analysis to compute genetic correlations. heritability of brain structure, although cer-
They demonstrated that the linkage between tain, is neither final nor static. However, with-
gray matter volumes and g is mediated by out genetic brain-mapping techniques (de-
a common set of genes. Haier et al. (2004) scribed in this review), strictly speaking it is

16 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

not certain whether these brain differences are of the hard data to establish a basis for why
attributable to innate predisposition or due to people vary in their general mental capac-
adaptations in response to skill acquisition and ity. This review illustrates the bridge afforded
repetitive rehearsal of those skills. by structural imaging between genetics and
Intelligence therefore depends, to some behavior.
extent, on structural differences in the brain Nature is not democratic. Individuals’ IQs
that are under very strong genetic con- vary, but the data presented in this review and
trol. This indicates a partly neuroanatomical elsewhere do not lead us to conclude that our
(structural) explanation for the high heritabil- intelligence is dictated solely by genes. In-
ity of intelligence. These methods are cur- stead genetic interactions with the environ-
rently being applied to large databases that ment suggest that enriched environments will
assess the impact of candidate genes on brain help everyone achieve their potential, but not
structure, which allows causal pathways be- to equality. Our potential seems largely pre-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

tween genes, brain, and behavior to be pur- determined.

sued at the allelic level. That our interpretation of intelligence, the
by University of Sussex Library on 11/10/06. For personal use only.

brain, and heritability has succumbed to a

variety of political and social pressures is un-
CONCLUSION deniable. How the public chooses to use sci-
Currently, the most fruitful combination of entific findings in the establishment of pol-
genetics and imaging is perhaps their applica- icy, particularly in regards to education and
tion to large patient populations. This shows law, however, is not the stuff of a chapter in
great promise for seeking out genetic mark- Annual Review of Neuroscience. As our under-
ers that are linked with brain structure, as well standing of the complex relationships between
as intellectual function and cognition, more genes, brain, and intelligence improves, what
generally. Brain mapping can provide some becomes of this knowledge remains to be seen.

ACKNOWLEDGMENTS
Funding support for this work was provided by grant P41 RR13642 from the National Cen-
ter for Research Resources, the Center for Computational Biology, U54 RR21813, and a
Human Brain Project grant and the International Consortium for Brain Mapping, funded
by the National Institute of Mental Health and the National Institute on Drug Abuse (P20
MH/DA52176). Additional support was provided by grants from the National Institute for
Biomedical Imaging and Bioengineering and the National Center for Research Resources
(R21 EB01651, R21 RR019771; to P.T.).

LITERATURE CITED
Andreasen NC, Flaum M, Swayze V 2nd, O’Leary DS, Alliger R, et al. 1993. Intelligence and
brain structure in normal individuals. Am. J. Psychiatry 150:130–34
Ashburner J, Friston KJ. 2000. Voxel-based morphometry—the methods. NeuroImage
11(6):805–21
Baare WF, Hulshoff Pol HE, Boomsma DI, Posthuma D, de Geus EJ, et al. 2001a. Quanti-
tative genetic modeling of variation in human brain morphology. Cereb. Cortex 11:816–
24
Baare WF, van Oel CJ, Hulshoff Pol HE, Schnack HG, Durston S, et al. 2001b. Volumes
of brain structures in twins discordant for schizophrenia. Arch. Gen. Psychiatry 58(1):33–
40

[Link] • Genetics of Brain Structure and Intelligence 17

 AR245-NE28-01 ARI 19 May 2005 10:1

Baron JC, Chetelat G, Desgranges B, Perchey G, Landeau B, et al. 2001. In vivo mapping of
gray matter loss with voxel-based morphometry in mild Alzheimer’s disease. NeuroImage
14:298–309
Bartley AJ, Jones DW, Weinberger DR. 1997. Genetic variability of human brain size and
cortical gyral patterns. Brain 120(2):257–69
Baumeister RF, Twenge J, Nuss CK. 2002. Effects of social exclusion on cognitive processes:
Anticipated aloneness reduces intelligent thought. J. Pers. Soc. Psychol. 83:817–27
Bellugi U, Lichtenberger L, Jones W, Lai Z, St. George M. 2000. I. The neurocognitive profile
of Williams Syndrome: a complex pattern of strengths and weakenesses. J. Cogn. Neurosci.
12(Suppl. 1):7–29
Biondi A, Nogueira H, Dormont D, Duyme M, Hasbon D, et al. 1998. Are the brains
of monozygotic twins similar? A three-dimensional MR study. Am. J. Neuroradiol.
19(7):1361–67
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

Block N. 1995. How heritability misleads about race. Cognition 56:99–128

Boomsma DI, de Geus EJ, van Baal GC, Koopmans JR. 1999. A religious upbringing reduces
the influence of genetic factors on disinhibition: evidence for interaction between genotype
by University of Sussex Library on 11/10/06. For personal use only.

and environment on personality. Twin Res. 2:115–25

Bouchard TJ Jr, Lykken DT, McGue M, Segal NL, Tellegen A. 1990. Sources of human
psychological differences: the Minnesota Study of Twins Reared Apart. Science 250:223–
28
Brand C. 2001. The G factor. [Link]
Bullmore ET, Suckling J, Overmeyer S, Rabe-Hesketh S, Taylor E, Brammer MJ. 1999. Global,
voxel, and cluster tests, by theory and permutation, for a difference between two groups
of structural MR images of the brain. IEEE Trans. Med. Imag. 18:32–42
Cabeza R, Nyberg L. 2000. Imaging cognition II: an empirical review of 275 PET and fMRI
studies. J. Cogn. Neurosci. 12(1):1–47
Cannon TD, Thompson PM, van Erp TGM, Toga AW, Poutanen V-P, et al. 2002. Cortex
mapping reveals regionally specific patterns of genetic and disease-specific gray-matter
deficits in twins discordant for schizophrenia. Proc. Natl. Acad. Sci. USA 99(5):3228–33
Cannon TD, van Erp TG, Bearden CE, Loewy R, Thompson P, et al. 2003. Early and late
neurodevelopmental influences in the prodrome to schizophrenia: contributions of genes,
environment, and their interactions. Schizophr. Bull. 29(4):653–69
Carroll J. 1993. Human Cognitive Abilities: A Survey of Factor-Analytic Studies. Cambridge, UK:
Cambridge Univ. Press
Cattell RB. 1971. Abilities: Their Structure, Growth, and Action. Boston: Houghton Mifflin
Ceci SJ, Williams WM. 1997. Schooling, intelligence, and income. Am. Psychol. 52:1051–58
Chipuer HM, Rovine MJ, Plomin R. 1990. LISREL modeling: genetic and environmental
influences on IQ revisited. Intelligence 14:11–29
Deary IJ. 2001. Human intelligence differences: a recent history. Trends Cogn. Sci. 5:127–30
Devlin B, Daniels M, Roeder K. 1997. The heritability of IQ. Nature 388(6641):468–71
Diamond A, Briand L, Fossella J, Gehlbach L. 2004. Genetic and neurochemical modulation
of prefrontal cognitive functions in children. Am. J. Psychiatry 161:125–32
Diamond MC. 1988. Enriching Heredity. New York: Free Press
Dickens WT, Flynn JR. 2001. Heritability estimates versus large environmental effects: the
IQ paradox resolved. Psychol. Rev. 108:346–69
Draganski B, Gaser C, Busch V, Schuierer G, Bogdahn U, May A. 2004. Neuroplasticity:
changes in grey matter induced by training. Nature 427(6972):311–12
Drane DL, Logemann JA. 2000. A critical evaluation of the evidence on the association between
type of infant feeding and cognitive development. Paediatr. Perinat. Epidemiol. 14:349–56

18 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

Duncan J, Seitz RJ, Kolodny J, Bor D, Herzog H, et al. 2000. A neural basis for general
intelligence. Science 289(5478):457–60
Egan MF, Goldberg TE, Kolachana BS, Callicott JH, Mazzanti CM, et al. 2001. Effect of
COMT Val 108/158 Met genotype on frontal lobe function and risk for schizophrenia.
Proc. Natl. Acad. Sci. USA 98:6917–22
Egan MF, Kojima M, Callicott JH, Goldberg TE, Kolachana BS, et al. 2003. The BDNF
val66met polymorphism affects activity-dependent secretion of BDNF and human mem-
ory and hippocampal function. Cell 112:257–69
Feldman MW, Otto SP. 1997. Twin studies, heritability, and intelligence. Science 278:1383–84
Finkel D, Pedersen NL, Plomin R, McClearn GE. 1998. Longitudinal and cross-sectional twin
data on cognitive abilities in adulthood: the Swedish Adoption/Twin Study of Aging. Dev.
Psychol. 34:1400–13
Fischl B, Liu A, Dale AM. 2001. Automated manifold surgery: constructing geometrically
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

accurate and topologically correct models of the human cerebral cortex. IEEE Trans Med.
Imag. 20(1):70–80
Forgays DG, Forgays JW. 1952. The nature of the effect of free-environmental experience in
by University of Sussex Library on 11/10/06. For personal use only.

the rat. J. Comp. Physiol. Psychol. 45:322–28

Friston KJ, Frith CD, Frackowiack RS, Turner R. 1995. Characterizing dynamic brain re-
sponses with fMRI: a multivariate approach. NeuroImage 2(2):166–72
Gardner H. 2000. Intelligence Reframed. New York: Basic Books
Gaser C, Schlaug G. 2003. Gray matter differences between musicians and nonmusicians. Ann.
N. Y. Acad. Sci. 999:514–17
Geroldi C, Pihlajamaki M, Laakso MP, DeCarli C, Beltramello A, et al. 1999. APOE-epislon4
is associated with less frontal and more medial temporal lobe atrophy in AD. Neurology
53(8):1825–32
Gignac G, Vernon PA, Wickett JC. 2003. Factors influencing the relationship between brain
size and intelligence. In The Scientific Study of General Intelligence, ed. H Nyborg, pp.
93–106. Amsterdam: Pergamon
Gogtay N, Lusk L, Hayashi KM, Giedd JN, Greenstein D, et al. 2004. Dymanic mapping of
human cortical development during childhood and adolescence. Proc. Natl. Acad. Sci. USA
101(21):8174–79
Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS. 2001. A voxel-
based morphometric study of ageing in 465 normal adult human brains. NeuroImage 14(1
Pt. 1):21–36
Gottfredson LS. 1997. Why g matters: the complexity of everyday life. Intelligence 24:79–132
Gould SJ. 1996 [1981]. The Mismeasure of Man. New York: Norton
Gray JR, Braver TS, Raichle ME. 2002. Integration of emotion and cognition in the lateral
prefrontal cortex. Proc. Natl. Acad. Sci. USA 99:4115–20
Gray JR, Chabris CF, Braver TS. 2003. Neural mechanisms of general fluid intelligence. Nat.
Neurosci. 6:316–22
Gray JR, Thompson PM. 2004. Neurobiology of intelligence: science and ethics. Nat. Rev.
Neurosci. 5:471–82
Greenough WT, Fulcher JK, Yuwiler A, Geller E. 1970. Enriched rearing and chronic elec-
troshock: effects on brain and behavior in mice. Physiol. Behav. 5(3):371–73
Haier RJ, Jung RE, Yeo RA, Head K, Alkire MT. 2004. Structural brain variation and general
intelligence. NeuroImage 23(1):425–33
Haier RJ, Siegel BV, MacLachlan A, Soderling E, Lottenberg S, Buchsbaum MS. 1992. Re-
gional glucose metabolic changes after learning a complex visuospatial/motor task: a
positron emission tomography study. Brain Res. 570:134–43

[Link] • Genetics of Brain Structure and Intelligence 19

 AR245-NE28-01 ARI 19 May 2005 10:1

Haier RJ, Siegel BV, Nuechterlein KH, Hazlett E, Wu JC, et al. 1998. Cortical glucose
metabolic-rate correlates of abstract reasoning and attention studied with positron emis-
sion tomography. Intelligence 12:199–217
Haier RJ, Siegel B, Tang C, Abel L, Buchsbaum MS. 1992. Intelligence and changes in regional
cerebral glucose metabolic-rate following learning. Intelligence 16:415–26
Haier RJ, White NS, Alkire MT. 2003. Individual differences in general intelligence correlate
with brain function during nonreasoning tasks. Intelligence 31:429–41
Hashimoto M, Yasuda M, Tanimukai S, Matsui M, Hirono N, et al. 2001. Apolipoprotein
E epsilon 4 and the pattern of regional brain atrophy in Alzheimer’s disease. Neurology
57(8):1461–66
Herrnstein RJ, Murray C. 1994. The Bell Curve. New York: Free Press
Janke AL, de Zubicaray GI, Rose SE, Griffin M, Chalk JB, Galloway GJ. 2001. 4D deforma-
tion modeling of cortical disease progression in Alzheimer’s dementia. Magn. Reson. Med.
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

46:661–66
Jensen AR. 1969. How much can we boost I.Q. and scholastic achievement? Harvard Educ.
Rev. 39(1):1–123
by University of Sussex Library on 11/10/06. For personal use only.

Jensen AR. 1980. Bias in Mental Testing. New York: Free Press
Jensen AR. 1998. The G Factor: The Science of Mental Ability. Westport, CT: Praeger
Jensen AR, Miele F. 2002. Intelligence, Race, and Genetics: Conversations with Arthur R. Jensen.
Boulder, CO: Westview
Kamin LJ. 1997. Twin studies, heritability, and intelligence. Science 278:1385
Kamin LJ, Goldberger AS. 2002. Twin studies in behavioral research: a skeptical view. Theor.
Popul. Biol. 61(1):83–95
Kaprio J, Koskenvuo M, Rose RJ. 1990. Population-based twin registries: illustrative appli-
cations in genetic epidemiology and behavioral genetics from the Finnish Twin Cohort
Study. Acta Genet. Med. Gemellol. 39(4):427–39
Korenberg JR, Chen XN, Hirota H, Lai Z, Bellugi U, et al. 2000. VI. Genome structure and
cognitive map of Williams Syndrome. J. Cogn. Neurosci. 12(Suppl. 1):89–107
Lehtovirta M, Helisalmi S, Mannermaa A, Soininen H, Koivisto K, et al. 1995. Apoliopoprotein
E polymorphism and Alzheimer’s disease in eastern Finland. Neurosci Lett. 185(1):13–15
Lehtovirta M, Partanen J, Konomen M, Hiltunen J, Helisalmi S, et al. 2000. A longitudinal
quantitative EEG study of Alzheimer’s disease: relation to apolioprotein E polymorphism.
Dement. Geriatr. Cogn. Disord. 11(1):29–35
Lewontin RC. 1975. Genetic aspects of intelligence. Annu. Rev. Genet. 9:387–405
Loehlin JC. 1989. Partitioning environmental and genetic contributions to behavioral devel-
opment. Am. Psychol. 44:1285–92
Lyons DM, Yang C, Sawyer-Glover AM, Moseley ME, Schatzberg AF. 2001. Early life
stress and inherited variation in monkey hippocampal volumes. Arch. Gen. Psychiatry
58(12):1145–51
Maguire EA, Gadian DG, Johnsrude IS, Good CD, Asburner J, et al. 2000. Navigation-related
structural change in the hippocampi of taxi drivers. Proc. Natl. Acad. Sci. USA 97(8):4398–
403
Mazziotta J, Toga AW, Evans A, Fox P, Lancaster J, et al. 2001. A probabilistic atlas and refer-
ence system for the human brain: International Consortium for Brain Mapping (ICBM).
Phil. Trans. R. Soc. Lond. (R. Soc.) 356:1293–322
McClearn GE, Johansson B, Berg S, Pedersen NL, Ahren F, et al. 1997. Substantial genetic
influence on cognitive abilities in twins 80 or more years old. Science 276:1560–63
McDaniel MA, Nguyen NT. 2002. A meta-analysis of the relationship between MRI-assessed brain
volume and intelligence. Presented at Proc. Int. Soc. Intel. Res., Nashville, TN

20 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

McGue M, Bouchard TJ Jr, Iacono WG, Lykken DT. 1993. Behavioral genetics of congni-
tive ability: a life-span perspective. In Nature, Nurture and Psychology, ed. R Plomin, GE
McClearn, pp. 59–76. Washington, DC: Am. Psychol. Assoc.
Mega MS, Dinov ID, Lee L, Masterman DM, O’Connor SM, et al. 2000. Orbital and dor-
solateral frontal perfusion defect associated with behavioral response to cholinesterase
inhibitor therapy in Alzheimer’s Disease. J. Neuropsychiatr. Clin. Neurosci. 12:209–18
Morley KI, Montgomery GW. 2001. The genetics of cognitive processes: candidate genes in
humans and animals. Behav. Genet. 31(6):511–31
Mummery CJ, Patterson K, Price CJ, Asburner J, Frackowiack RS, Hodges JR. 2000. A voxel-
based morphometry study of semantic dementia: relationship between temporal lobe at-
rophy and semantic memory. Ann. Neurol. 47(1):36–45
Narr KL, Cannon TD, Woods RP, Thompson PM, Kim S, et al. 2002. Genetic contributions
to altered callosal morphology in schizophrenia. J. Neurosci. 22(9):3720–29
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

Narr KL, Thompson PM, Sharma T, Moussai J, Zoumalon C, et al. 2001. Three-dimensional
mapping of gyral shape and cortical surface asymmetries in schizophrenia: gender effects.
by University of Sussex Library on 11/10/06. For personal use only.

Am. J. Psychiatry 158(2):244–55

Narr KL, Thompson PM, Szeszko P, Robinson D, Jang S, et al. 2003. Regional specificity of
hippocampal volume reductions in first-episode schizophrenia. NeuroImage 21(4):1563–
75
Neale MC. 1997. Mx: Statistical Modeling. Richmond: Va. Commonwealth Univ., Dep. Psy-
chiatry
Neisser U, ed. 1998. The Rising Curve: Long-Term Gains in IQ and Related Measures. Washington,
DC: Am. Psychol. Assoc.
Oddy WH, Kendall GE, Blair E, De Klerk NH, Stanley FJ, et al. 2003. Breast feeding and
cognitive development in childhood: a prospective birth cohort study. Paediatr. Perinat.
Epidemiol. 17:81–90
Oppenheim JS, Skerry JE, Tramo MJ, Gazzaniga MS. 1989. Magnetic resonance imaging
morphology of the corpus callosum in monozygotic twins. Ann. Neurol. 26:100–4
Pfefferbaum A, Sullivan EV, Carmelli D. 2001. Genetic regulation of regional microstructure
of the corpus callosum in late life. Neuroreport 12(8):1677–81
Pfefferbaum A, Sullivan EV, Swan GE, Carmelli D. 2000. Brain structure in men remains
highly heritable in the seventh and eighth decades of life. Neurobiol. Aging 21:63–74
Pinker S. 2002. The Blank Slate. New York: Viking
Plomin R. 1999. Genetics of childhood disorders. III. Genetics and intelligence. J. Am. Acad.
Child Adolesc. Psychiatry 38(6):786–88
Plomin R, DeFries JC, Loehlin JC. 1977. Genotype-environment interaction and correlation
in the analysis of human behavior. Psychol. Bull. 84:309–22
Plomin R, Kosslyn SM. 2001. Genes, brain and cognition. Nat. Neurosci. 4:1153–54
Plomin R, Petrill SA. 1997. Genetics and intelligence: What’s new? Intelligence 24:53–78
Posthuma D, De Geus EJ, Baare WF, Hulshoff Pol HE, Kahn RS, Boomsma DI. 2002. The
association between brain volume and intelligence is of genetic origin. Nat. Neurosci. 5:83–
84
Prabhakaran V, Rypma B, Gabrieli JDE. 2001. Neural substrates of mathematical reasoning: a
functional magnetic resonance imaging study of neocortical activation during performance
of the Necessary Arithmetic Operations Test. Neuropsychology 15:115–27
Prabhakaran V, Smith JAL, Desmond JE, Glover GH, Gabrieli JDE. 1997. Neuronal substrates
of fluid reasoning: an fMRI study of neocortical activation during performance of the
Raven’s Progressive Matrices Test. Cogn. Psychol. 33:43–63

[Link] • Genetics of Brain Structure and Intelligence 21

 AR245-NE28-01 ARI 19 May 2005 10:1

Rasser PE, Johnston PJ, Lagopoulos J, Ward PB, Schall U, et al. 2005. Analysis of fMRI BOLD
activation during the Tower of London Task using Gyral Patter and Intensity Averaging
Models of Cerebral Cortex. NeuroImage. In press
Rex A, Pfeifer L, Fink H. 2001. Determination of NADH in frozen rat brain sections by
laser-induced fluorescence. Biol. Chem. 382(12):1727–32
Ridley M. 2003. Nature via Nurture: Genes, Experience, and What Makes Us Human. New York:
Harper Collins
Rombouts SA, Barkhof F, Witter MP, Schettens P. 2000. Unbiased whole-brain analysis of
gray matter loss in Alzheimer’s disease. Neurosci. Lett. 285(3):231–33
Roses AD. 1996. Apolipoprotein E and Alzheimer’s disease. A rapidly expanding field with
medical and epidemiological consequences. Ann. N. Y. Acad. Sci. 802:50–57
Rowe DC, Jacobson KC. 1999. Mainstream: research in behavioral genetics. In Behavioral
Genetics: The Clash of Culture and Biology, ed. RA Carson, MA Rothstein, pp. 12–34. Balti-
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

more/London: Johns Hopkins Univ. Press

Salovey P, Mayer JD, Caruso D. 2002. The positive psychology of emotional intelligence.
In The Handbook of Positive Psychology, ed. CR Snyder, SJ Lopez, pp. 159–71. New York:
by University of Sussex Library on 11/10/06. For personal use only.

Oxford Univ. Press

Schmader T, Johns M. 2004. Converging evidence that stereotype threat reduces working
memory capacity. J. Pers. Soc. Psychol. 85:440–52
Sesardic N. 2002. Heritability and indirect causation. Proc. Philos. Sci. Assoc. 70:1002–14
Small GW, Ercoli LM, Silverman DH, Huang SC, Komo S, et al. 2000. Cerebral metabolic
and cognitive decline in person at genetic risk for Alzheimer’s disease. Proc. Natl. Acad.
Sci. USA 97(11):6037–42
Sowell ER, Mattson SN, Thompson PM, Jernigan TL, Riley EP, Toga AW. 2001. Mapping
callosal morphology and cognitive correlates: effects of heavy prenatal alcohol exposure.
Neurology 57:235–44
Sowell ER, Peterson BS, Thompson PM, Welcome SE, Henkenius AL, Toga AW. 2003.
Mapping cortical change across the human life span. Nat. Neurosci. 6(3):309–15
Sowell ER, Thompson PM, Holmes CJ, Batth R, Travner DA, et al. 1999. Localizing age-
related changes in brain structure between childhood and adolescence using statistical
parametric mapping. NeuroImage 9:587–97
Spearman C. 1927. The Nature of “Intelligence” and the Principles of Cognition. London, UK:
Macmillan
Steele CM, Aronson J. 1995. Stereotype threat and the intellectual test performance of African
Americans. J. Pers. Soc. Psychol. 69:797–811
Sternberg RJ. 1999. The theory of successful intelligence. Rev. Gen. Psychol. 3:292–16
Suddarth RI, Christison GW, Torrey EF, Casanova MF, Weinberger DR. 1990. Anatomical
abnormalities in the brains of monozygotic twins discordant for schizophrenia. N. Engl.
J. Med. 322(12):789–94
Sullivan EV, Pfefferbaum A, Swan GE, Carmelli D. 2001. Heritability of hippocampal size in el-
derly twin men: equivalent influence from genes and environment. Hippocampus 11(6):754–
62
Swan GE, Carmelli D, Reed T, Harshfield GA, Fabsitz RR, Eslinger PJ. 1990. Heritability of
cognitive performance in ageing twins. The National Heart, Lung, and Blood Institute
Twin Study. Arch. Neurol. 47(3):259–62
Thompson P, Hayashi KM, de Zubicaray G, Janke AL, Rose SE, et al. 2003. Dynamics of gray
matter loss in Alzheimer’s disease. J. Neurosci. 23(3):994–1005
Thompson P, Hayashi KM, de Zubicaray G, Janke AL, Rose SE, et al. 2004. Mapping hip-
pocampal and ventricular change in Alzheimer’s disease. NeuroImage 22(4):1754–66

22 Toga · Thompson
 AR245-NE28-01 ARI 19 May 2005 10:1

Thompson P, Toga AW. 1996. A surface-based technique for warping three-dimensional im-
ages of the brain. IEEE Trans. Med. Imag. 15(4):402–17
Thompson PM, Cannon TD, Narr KL, van Erp T, Poutanen V-P, et al. 2001a. Genetic
influences on brain structure. Nat. Neurosci. 4(12):1253–58
Thompson PM, Cannon TD, Toga AW. 2002. Mapping genetic influences on human brain
structure. Ann. Med. 24:523–36
Thompson PM, MacDonald D, Mega MS, Holmes CJ, Evans AC, Toga AW. 1997. Detection
and mapping of abnormal brain structure with a probabilistic atlas of cortical surfaces. J.
Comp. Assist. Tomogr. 21(4):567–81
Thompson PM, Toga AW. 2005. Cortical disease and cortical localization. Nat. Encycl. Life Sci.
In press
Thompson PM, Vidal C, Giedd JN, Gochman P, Blumenthal J, et al. 2001b. Mapping adoles-
cent brain change reveals dynamic wave of accelerated gray matter loss in very early-onset
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

schizophrenia. Proc. Natl. Acad. Sci. USA 98(20):11650–55

Thompson PM, Woods RP, Mega MS, Toga AW. 2000. Mathematical/computational chal-
by University of Sussex Library on 11/10/06. For personal use only.

lenges in creating deformable and probabilistic atlases of the human brain. Hum. Brain
Mapp. 9:81–92
Tramo MJ, Loftus WC, Stukel TA, Green RL, Weaver JB, Gazzaniga MS. 1998. Brain size,
head size, and intelligence quotient in monozygotic twins. Neurology 50(5):1246–52
Turkheimer E, Haley A, Waldron M, D’Onofrio B, Gottesman II. 2003. Socioeconomic status
modifies heritability of IQ in young children. Psychol. Sci. 14:623–28
Vogel F, Motulsky AG. 1997. Human Genetics. New York: Springer-Verlag
Weinberger DR, Delisi LE, Neophytides AN, Wyatt RJ. 1981. Familial aspects of CT scan
abnormalities in chronic schizophrenic patients. Psychiatry Res. 4:65–71
Winterer G, Goldman D. 2003. Genetics of human prefrontal function. Brain Res. Brain Res.
Rev. 43(1):134–63
Wright IC, McGuire PK, Poline JB, Travere JM, Murray RM, et al. 1995. A voxel-based
method for the statistical analysis of gray and white matter density applied to schizophrenia.
NeuroImage 2(4):244–52
Wright IC, Sham P, Murray RM, Weinberger DR, Bullmore ET. 2002. Genetic contributions
to regional human brain variability estimated in a twin study by path analysis of parcellated
grey matter maps. NeuroImage 17:256–71
Wright IC, Sharma T, Ellison ZR, McGuire PK, Fiston KJ, et al. 1999. Supra-regional brain
systems and the neuropathology of schizophrenia. Cereb. Cortex 9(4):366–78
Zeineh MM, Engel SA, Thompson PM, Bookheimer SY. 2001. Unfolding the human hip-
pocampus with high resolution structural and functional MRI. Anat. Rec. 265(2):111–20
Zeineh MM, Engel SA, Thompson PM, Bookheimer SY. 2003. Dynamics of the hippocampus
during encoding and retrieval of face-name pairs. Science 299(5606):577–80
Zhou Y, Cheshire A, Howell CA, Ryan DH, Harris RB. 1999. Neuroautoantibody immunore-
activity in relation to aging and stress in apolipoprotein E-deficient mice. Brain Res. Bull.
49(3):173–79

[Link] • Genetics of Brain Structure and Intelligence 23

Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]
by University of Sussex Library on 11/10/06. For personal use only.
AR245-NE28-01
ARI
19 May 2005
10:1
Contents ARI 13 May 2005 13:21

Annual Review of
Neuroscience

Volume 28, 2005

Contents

Genetics of Brain Structure and Intelligence

Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

Arthur W. Toga and Paul M. Thompson p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 1

The Actin Cytoskeleton: Integrating Form and Function at the Synapse
by University of Sussex Library on 11/10/06. For personal use only.

Christian Dillon and Yukiko Goda p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p25

Molecular Pathophysiology of Parkinson’s Disease
Darren J. Moore, Andrew B. West, Valina L. Dawson, and Ted M. Dawson p p p p p p p p p p p p p57
Large-Scale Genomic Approaches to Brain Development and Circuitry
Mary E. Hatten and Nathaniel Heintz p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p89
Autism: A Window Onto the Development of the Social
and the Analytic Brain
Simon Baron-Cohen and Matthew K. Belmonte p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 109
Axon Retraction and Degeneration in Development and Disease
Liqun Luo and Dennis D.M. O’Leary p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 127
Structure and Function of Visual Area MT
Richard T. Born and David C. Bradley p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 157
Growth and Survival Signals Controlling Sympathetic Nervous System
Development
Natalia O. Glebova and David D. Ginty p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 191
Adult Neurogenesis in the Mammalian Central Nervous System
Guo-li Ming and Hongjun Song p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 223
Mechanisms of Vertebrate Synaptogenesis
Clarissa L. Waites, Ann Marie Craig, and Craig C. Garner p p p p p p p p p p p p p p p p p p p p p p p p p p p p 251
Olfactory Memory Formation in Drosophila: From Molecular
to Systems Neuroscience
Ronald L. Davis p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 275
The Circuitry of V1 and V2: Integration of Color, Form, and Motion
Lawrence C. Sincich and Jonathan C. Horton p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 303

v
Contents ARI 13 May 2005 13:21

Molecular Gradients and Development of Retinotopic Maps

Todd McLaughlin and Dennis D.M. O’Leary p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 327
Neural Network Dynamics
Tim P. Vogels, Kanaka Rajan, and L.F. Abbott p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 357
The Plastic Human Brain Cortex
Alvaro Pascual-Leone, Amir Amedi, Felipe Fregni, and Lotfi B. Merabet p p p p p p p p p p p p p p 377
An Integrative Theory of Locus Coeruleus-Norepinephrine Function:
Adaptive Gain and Optimal Performance
Gary Aston-Jones and Jonathan D. Cohen p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 403
Neuronal Substrates of Complex Behaviors in C. elegans
Mario de Bono and Andres Villu Maricq p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 451
Annu. Rev. Neurosci. 2005.28:1-23. Downloaded from [Link]

Dendritic Computation
Michael London and Michael Häusser p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 503
by University of Sussex Library on 11/10/06. For personal use only.

Optical Imaging and Control of Genetically Designated Neurons

in Functioning Circuits
Gero Miesenböck and Ioannis G. Kevrekidis p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 533

INDEXES

Subject Index p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 565

Cumulative Index of Contributing Authors, Volumes 19–28 p p p p p p p p p p p p p p p p p p p p p p p p p p p 577
Cumulative Index of Chapter Titles, Volumes 19–28 p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p 582

ERRATA

An online log of corrections to Annual Review of Neuroscience chapters

may be found at [Link]

vi Contents