Abstract

Because healthy monozygotic twins share an identical genetic complement, they provide a unique opportunity to explore the genetic and environmental determinants of brain development. The purpose of this study was to evaluate the similarities between measures of cerebral and subcortical volumes and surface morphology in monozygotic twins compared with a matched control sample. Combinations of automated and manual methods were used to evaluate total brain volume, gray matter, white matter, ventricles and volumes of the frontal, temporal, parietal and occipital lobes. An artificial neural network algorithm was used to measure the cerebellum, thalamus, caudate and putamen. Measures of surface morphology included an index of gyral and sulcal curvature, surface area and cortical depth. The cerebral volume regions, including the gray matter, white matter and lobar volumes, were highly correlated within monozygotic twin pairs, with nearly all correlation coefficients >0.90. The cerebellum was also highly correlated (r = 0.99). Reasonably high correlations were found for the cortical depth (r = 0.84), caudate (r = 0.84), thalamus (r = 0.75) and putamen (r = 0.75). The surface measures, however, demonstrated the least correlation within twin pairs and thus are more prone to environmental influences. The high to moderate correlations between MZ twins compared with the matched controls highlights the role of heredity in both prenatal and postnatal neurodevelopment.

Introduction

Monozygotic (MZ) twins are nature's form of cloning. This scenario creates two individuals raised in similar environments with essentially identical genes. Yet even when twins are separated at birth and raised in different homes, attributes such as intelligence, temperament, and personality characteristics are often very similar (Bouchard et al., 1990). Having an identical genetic complement creates a myriad of physical and psychological features that are consistently similar, yet not identical (Machin, 1996). It is not uncommon for MZ twins to have such similar facial features that only those who are well acquainted with them are able to tell them apart.

These striking similarities in the facial features of MZ twins tell us that gene–gene interactions are almost the sole determinant in the development of normal craniofacial relationships. Thus, if we know that craniofacial development and brain development are intimately intertwined, it is plausible to also hypothesize that genes would have a strong role in brain development. The development of the face and the development of the brain, under both normal and pathologic conditions, are intimately related (Sperber, 1992; Kjaer, 1995). In fact, the phrase ‘The face reflects the brain’ is well known in the field of dysmorphology (Winter, 1996). Syndromes of craniofacial maldevelopment are frequently accompanied by cognitive deficits, indicating that the genes affecting craniofacial development are intertwined with those in brain development. For example, the longer the trinucleotide repeat in Fragile X syndrome, the greater the level of mental retardation and the more prevalent the dysmorphic facial features (Rousseau et al., 1994; Jones, 1997).

Genes that regulate the underlying architecture of the brain are expressed mainly during uterine life (Leckman and Lombroso, 1998; Lombroso, 1998; Naegele and Lombroso, 1998; Rakic and Lombroso, 1998; Rubenstein, 1998; Vaccarino and Lombroso, 1998; Rubenstein and Rakic, 1999), although modeling continues throughout the life span (Yakovlev and Lecours, 1967; Jernigan et al., 1991b; Giedd et al., 1996; Hockfield and Lombroso, 1998a,b; Peters et al., 1998; Sowell et al., 1999a,b; Bartzokis et al., 2001). Since gene expression is not devoid of environmental influences, specific features of brain development can be modulated via non-genetic factors (Wiesel, 1982; Kempermann et al., 1997; Knudsen, 1998). It is likely that environmental and random factors account for the observation that twins are ‘similar’ rather than ‘identical’. Extreme examples of MZ twin dissimilarities do exist for a wide array of neurodevelopmental disorders, ranging from holoprosencephaly (Schinzel et al., 1979; Machin et al., 1985) to schizophrenia (Pollin et al., 1966; Stabenau and Pollin, 1967; Weinberger et al., 1992). Given the complex nature of cerebral development, it would not be surprising to find dissimilarities even between healthy MZ twins.

Midline structures of the brain measured by magnetic resonance imaging (MRI) are reported to be highly correlated between healthy monozygotic twins (Oppenheim et al., 1989; Biondi et al., 1998; Tramo et al., 1998). For example, the correlation coefficient for the distance of the line connecting the anterior and posterior commissures (the AC–PC line) in MZ twins is 0.70 (Biondi et al., 1998). The corpus callosum length (Oppenheim et al., 1989; Biondi et al., 1998) and area (Oppenheim et al., 1989) are also highly correlated, ranging from 0.73 to 0.90 and 0.90 to 0.99, respectively. However, not all midline structures have such high correlations in MZ twins. For example, the length of the perimeter of the corpus callosum has an interclass correlation coefficient (ICC) of 0.62.

Studies have demonstrated that comparisons of brain volume consistently exceed the high correlations found in the midline structures (Bartley et al., 1997; Carmelli et al., 1998; Tramo et al., 1998; Pennington et al., 2000). Volumes of both the right and left hemispheres have ICCs of ~0.95 (Bartley et al., 1997). Three studies included dizygotic twin pairs with a full twin study and reported heritability scores of >90% for total brain volume (Bartley et al., 1997; Carmelli et al., 1998; Pennington et al., 2000). Not all cerebral structures in MZ twins show high rates of similarities. For example, Steinmetz et al. (Steinmetz et al., 1995) demonstrated that MZ twins discordant for handedness exhibited differing degrees of asymmetry of the planum temporale. Epigenetic factors are probable explanations, not only for asymmetries of the planum temporale, but also for other surface measures of the brain, which show significantly lower correlations than the volumetric data.

Whereas high correlations exist for brain volumes and midline structures, more variability between twins has been described in the patterns of gyral and sulcal development (Bartley et al., 1997; Biondi et al., 1998; Lohmann et al., 1999). Previous MRI studies of the surface of the cerebral cortex, however, have used relatively incomplete and/or rudimentary measures. Evaluating sulcal depth, Lohmann et al. (Lohmann et al., 1999) found that the deeper primary sulci, those formed earlier in development, displayed more similarities than the more shallow, tertiary sulci. The strongest evidence for surface pattern dissimilarities between MZ twins came from a full twin study by Bartley et al. (Bartley et al., 1997). This study, which used a cross-correlation algorithm to compare structural brain images, demonstrated that although influenced by genes, the majority of the morphologic variance between MZ twins was a result of random environmental effects. Tramo et al. (Tramo et al., 1995) measured the surface area in ten pairs of MZ twins and found both hemispheres to be highly influenced by genetic factors, with a greater genetic influence on the left hemisphere.

None of the twin studies to date have examined a comprehensive group of brain measures reflecting the full spectrum of developmental processes that begin during gestation and are completed during early adulthood. This study compares a large array of brain structures and surface measures in MZ twins and attempts to expand the potential of MRI technology to examine the role of genetic influences on brain development. Pairs of unrelated controls matched on age, gender, height and paternal education were compared with the twin sample to evaluate the effect of these demographic variables on the brain correlations. Measures include total brain volume, total gray matter (GM) and white matter (WM) volumes, the GM/WM volumes for each of the four cerebral lobes (frontal, parietal, temporal and occipital), and surface and ventricular CSF volumes. Volumes of subcortical structures (thalamus, caudate and putamen), which have not been described previously in MZ twins, are measured by an automated (non-manual) technique that uses an artificial neural network (ANN).

In addition to these comprehensive volume measurements, the surface measures were obtained to examine the ongoing process of gyrification, which continues into the early twenties (Zilles et al., 1988; Armstrong et al., 1995) and which is likely to be more influenced by environmental factors. Surface measures used in this study are novel three-dimensional indices, and are fully automated and well validated (Magnotta et al., 1999a). They include the total brain surface area, gyral and sulcal curvature indices, cortical depth, and a measure of surface complexity. Correlation coefficients for cortical thickness, curvature of the sulci and gyri, and individual subcortical structures have not previously been reported in twins.

Materials and Methods

The sample population consisted of 12 healthy right-handed MZ twin pairs and 12 pairs of controls (see Table 1). The control pairs were matched by age, sex, height and paternal education. Handedness was assessed using the Edinburgh Handedness Inventory (Oldfield, 1971). The mean ages for the twin and control groups were 24.5 ± 7.2 years (mean ± SD) and 24.4 ± 7.2 years, respectively. Both groups were evenly matched, with six male and six female pairs. The mean duration of education was 14.2 years for both the patient and control groups. All were enrolled at the Mental Health Clinical Research Center at the University of Iowa Hospital and Clinics after the nature of the study was fully explained and informed consent obtained. Exclusion criteria included a positive history of medical, neurological or psychiatric illness, including alcohol and substance abuse, and age <18 or >40 years. Confirmation of zygosity was assessed by the identical matching of five different blood type antigens, six serum proteins and four red cell enzymes.

MRI Data Acquisition and Processing

Images were obtained on a 1.5 T GE Signa MRI scanner. Three different MRI sequences were used for each subject. The T1-weighted spoiled grass sequence was acquired with the following parameters: slice thickness = 1.5 mm, slice number = 124, TE = 5 ms, TR = 24 ms, flip angle = 40, NEX = 2, FOV = 26 cm, matrix = 256 × 192. The PD- and T2-weighted images were obtained with the following parameters: slice thickness = 3.0 or 4.0 mm, TE = 36 ms for PD or 96 ms for T2, TR = 3000 ms, NEX = 1, FOV = 26 cm, matrix = 256 × 192, with an echo train length = 8. All scans were rated for overall quality and for movement artifacts using a 0–4 scale (0 = very poor, 4 = excellent). All but one scan received quality ratings between 3 and 4. The image processing was performed on a Silicon Graphics workstation using the locally developed BRAINS software package (Andreasen et al., 1992, 1993; Cohen et al., 1992). The images were initially realigned and resampled. A Talairach coordinate system (Talairach and Tournoux, 1988) was warped to the image. Extracranial tissue was removed using edge detection techniques and manual tracing. The pixels representing the measures of GM, WM and cerebral spinal fluid (CSF) were identified using a multispectral discriminant analysis-based segmentation algorithm applied to the three image sequences described above (T1, T2, PD) (Harris et al., 1997, 1999). The segmentation algorithm provides both a discrete classification of tissue into GM, WM and CSF and a continuous classifier that contains partial voluming information. The measures reported in this study represent the results of the discrete classification of specific lobes and their constituent GM, WM and CSF volumes.

Volumetric Measures of Structure and Tissue Composition

Within the Talairach space, boxes were assigned to identify specific brain regions. The intracranial volume was subdivided into brain tissue and CSF. The CSF was broken down into surface CSF and internal CSF (ventricles and cisterns). The subcortical regions and cerebellum were removed and the cortical tissue was subdivided into the four lobes utilizing pre-assigned Talairach boxes. Utilizing the discrete segmentation procedure, each of the brain subdivisions was further divided into GM and WM. An automated algorithm was used to quantify the volumes of GM, WM and CSF within each lobe. The use of stereotactic techniques have been shown to be accurate for measures of cerebral and lobar volumes (Collins et al., 1994; Andreasen et al., 1996).

Measures of Surface Anatomy

A thorough description of the algorithm used to quantify the surface morphology has been described by Magnotta et al. (Magnotta et al., 1999a). The continuous segmented image is used to identify the region of ‘pure’ cortical GM (i.e. the GM corrected for partial voluming). The parametric center of this ‘pure’ cortical GM was calculated and a trianglebased isosurface was created. The resulting 3-D iso-surface spanned the brain at the approximate spatial center of the cortical GM. The surface area was calculated as the sum of triangular areas covering this surface of the brain.

Both a sulcal and gyral curvature index was calculated by determining the vector angle normal to each triangle surface compared with neighboring vector angles up to four triangle surfaces away. Convex values (i.e. positive) represent gyri and concave values (i.e. negative) represent sulci. Higher absolute values of curvature reflect ‘tighter’ curvature, whereas lower values represent a more ‘broader’ curve.

Cortical depth is calculated as the distance of the vector normal to the triangular surface to the internal GM/WM interface. Each triangle in the surface was assigned four surface normals, one on each corner and one in the center. The cortical depth for each triangle is defined as vector with the minimum distance to the 50% GM and 50% WM region. Since the triangle iso-surface lies at the parametric center of the GM, the cortical depth is approximately half of the actual cortical depth. Values were multiplied by two to obtain a measure of the actual cortical depth.

An index of surface complexity is calculated by the ratio of the total cortical surface area to the overall brain volume to the 2/3 power. These measurements have been shown to have high scan–rescan reproducibility (R2 = 0.99 for cerebral volume; R2 = 0.97 for surface area; R2 = 0.82 for surface complexity; and R2 = 0.88 for cortical thickness) (Magnotta et al., 1999a).

Artificial Neural Network (ANN) Measurements of the Cerebellum and Subcortical Structures

The volumes of the cerebellum, thalamus, caudate and putamen were calculated utilizing an ANN. The ANN is a parallel-processing network of subunits that are hierarchically layered and are designed to learn from experience. A thorough description of the fully connected feed-forward, three-layer ANN used in this study has been reported by Magnotta et al. (Magnotta et al., 1999b). During the training phase, the ANN is fed a series of images with proven inter- and intrarater reliability for a specific region of interest (ROI), e.g. the putamen. The ANN consists of an input layer, a hidden layer and an output layer. Using an iterative algorithm, the ANN uses the ‘gold standard’ ROIs to re-weight each input layer until the learning approaches an optimum match.

Once trained, a bounding box was manually defined surrounding the cerebellum and each subcortical ROI to be used as the search space for the ANN. The cerebellum was not segmented due to poor face validity, an inevitable consequence of measuring GM in a finely foliated structure with 1.5 mm slice thickness. Each ROI was then manually inspected by a trained technician blind to the twins status to rule out any gross abnormalites. The intraclass correlation coefficients and measures of overlap for the ANN compared with manual tracing techniques are shown in Table 2. A graphic representation of the structures identified by the ANN is shown in Figure 1.

Statistical Analysis

The twin pairs were randomly assigned as either twin A or twin B, and Pearson correlations were performed for each variable. The same analysis strategy was also calculated for the constructed control sample. A comparison of the two independent Pearson correlations from the twin and control samples was compared using a Fisher's r-to-Z. A comparison of the twin versus control samples for each of the volume and surface measurements was performed using two-tailed tests (Stuart and Ord, 1987).

Results

The control sample was matched as closely as possible to the twin sample on age, height and paternal education. As shown in Table 1, the correlation for these three variables are between 0.95 for height (compared with a correlation of 0.96 for the twin sample) and 1.0 for paternal education level. Although the inter-pair variability exceeded the intra-pair variability in the constructed control pairs for educational level attained, maternal education level and parental socioeconomic status, thereby reducing the correlation coefficients, the standard deviations for these measures are relatively narrow. Thus, the control pairs are relatively closely matched on all demographic measures, and the twin group and control group are also closely matched.

Tissue Composition and Cortical Subregions

The results of the Pearson's correlation coefficient for the brain volumetric data are shown in Tables 3–5.

High correlations exist for the left hemisphere (r = 0.98), right hemisphere (r = 0.98) and total brain volume (r = 0.99) for the MZ twin pair. Furthermore, when the brain is progressively divided into subregions and tissue types, the correlations remain high. The total cerebral gray matter (GM) and white matter (WM) volumes have correlations of 0.99 and 0.98, respectively. These patterns remain consistent when the brain is further parcellated into its four lobes (Table 3). The only two lobar volumes that have correlations below 0.90 are the frontal WM (0.83) and the occipital GM (0.69).

The correlations for the constructed control pairs are very low. The only exception is in the cerebral WM, which demonstrates a moderate correlation of 0.40 (Table 3). The lobar WM volumes demonstrate moderate correlations for the temporal (0.41) and occipital lobes (0.51) (Table 4). Even so, the confidence intervals for all the measures of the control pairs include a zero correlation in their range.

The cerebellar volumes were also highly correlated between twins (r = 0.99), whereas the subcortical structures were slightly less so (Table 5). The correlations between twins was highest for the caudate (r = 0.84), although this was very similar to the 0.75 correlation for both the putamen and the thalamus. Since these values are approximately the same as the intraclass reliability measurements for the ANN, the slightly reduced correlation of these structures may be secondary to measurement error. The volumes of the subcortical structures and the cerebellum between the matched control pairs were low (Tables 3 and 5), except for ventricular volume (r = 0.52).

The ventricular volumes were relatively highly correlated (r = 0.85) in the MZ twins, yet this failed to achieve significance when compared with the age and gender matched control population. The volume of the surface CSF surrounding the brain also displayed a high correlation (r = 0.91) between twins and was very low in the control pairs (r = 0.11).

Surface Morphology

Whereas the volumetric data for the cortical, lobar and tissue measures showed high correlations in the twins, the brain surface morphology failed to maintain this consistent pattern (see Table 6). Measures of surface area (r = 0.69), gyral curvature (r = 0.63), sulcal curvature (r = 0.58) and surface complexity (r = 0.49) showed only modest correlations. The sole surface measure that maintained a relatively high correlation was the cortical depth (r = 0.84).

Discussion

Although healthy MZ twins share both identical genetic complements and very similar environments, they are not ‘identical’ (Machin, 1996). The inter-twin differences lie on a continuum, with some features under tighter genetic control, while others have a greater potential for modulation by non-genetic influences. The present study supports previous research demonstrating that grossly defined brain volumes between MZ twins are highly correlated (Bartley et al., 1997; Carmelli et al., 1998; Tramo et al., 1998), whereas measures of surface morphology display more variability (Steinmetz et al., 1995; Bartley et al., 1997; Lohmann et al., 1999). Newly reported findings in this study are the high correlations in the MZ twins for the WM and GM for each of the four cerebral lobes, the cerebellum and subcortical structures. This study also demonstrates that the depth of the cerebral cortex is also highly correlated between monozygotic twins (r = 0.84).

There is a complex progression of neurodevelopment from the early formation of the neural tube to the continued reshaping of the adult brain. The most rapid period of growth occurs in utero, and involves neuronal differentiation, migration and connections via axons and dendrites with subcortical and cortical structures (daddyCaviness, 1982; Rakic, 1988; Lombroso, 1998; Vaccarino and Lombroso, 1998). The peak of synaptogenesis occurs within the first year following birth (Huttenlocher, 1990), and there is a progressive decrement in GM during adolescence and adulthood (Gur et al., 1991; Jernigan et al., 1991b; Giedd et al., 1996; Sowell et al., 1999a,b; Bartzokis et al., 2001). Myelination, which also demonstrates substantial early growth, has been shown to increase during adolescence and into early adulthood (Yakovlev and Lecours, 1967). Since the twins were studied as adults, it is unclear where along the developmental trajectory the correlations in surface patterns and volumetric data deviated.

The developmental period during which small deviations between twins would be amplified over time occurs during early development. According to the radial unit hypothesis of cortical development, the neurons that are formed via mitosis at the ventricular zone migrate along radial glial guide cells to their ultimate location on the cortical plate (Rakic, 1988). Each successive generation of neurons migrates in an inside-out pattern, with later generations migrating through previously developed cells before reaching their ultimate position (Sidman and Rakic, 1973). The depth of the gray mantle in the cortex is derived from the six layers that develop in this inside-out pattern. Whereas the depth of the cortex has increased only slightly during evolution, surface area has shown considerable expansion (Welker, 1990). The high correlation related to cortical depth found in this study may be representative of a stronger genetic modulation in the ontogeny of the six-layered cortical mantle.

Whereas the development of cortical depth is dependent on the number of neurons produced in each unit plate, the surface area of the brain is dependent on the number of contributing radial units along the ventricular zone (Rakic, 1988, 1995). These proliferative units are formed during the first 6 weeks of gestation and are aligned in a radial orientation along the ventricular zone (Sidman and Rakic, 1973; Rakic, 1974). The greater the number of radial units, the greater the number of lined projections to the cortical plate and the greater the surface area. During the first 6 weeks of gestation, the proliferative units divide symmetrically, with each progenitor cell producing two progenitor cells with each mitotic cycle (Rakic, 1988). Since each round of mitosis results in an exponential increase in the number of progenitor cells, it is postulated that small changes affecting the duration of symmetric growth have resulted in the robust ontogenetic growth in surface area (Rakic, 1995). Thus small epigenetic influences during the phase of symmetric cell division could have fairly pronounced effects on the surface area, whereas a similar influence during asymmetric cell division may have little or no effect on the cortical depth.

During the migration of neurons from the ventricular zone to the cortical plate, efferent projections originating in the thalamus connect with the migrating neurons (Rakic, 1988). This developmental process allows for direct neuronal connections between cortical GM and the thalamus. Such a direct relationship between cortical GM and the thalamus would presume a volumetric relationship between the two structures. A Pearson's correlation comparing the total thalamic volume and cortical GM for all subjects was 0.76 (P < 0.0001). Although in the expected direction (i.e. larger brains would be expected to have both larger thalami and cortical GM), this finding is reduced, but still significant when controlling for both height (r = 0.58; P < 0.0001) and total brain tissue (r = 0.37; P = 0.01). This relationship between the thalamus and cortical gray matter meshes well with the present neurodevelopmental and neuroanatomic understanding of thalamic/cortical GM connectivity.

The dramatic expansion of the surface area during evolution is contained within the cranial vault through the folding or gyrification of the brain surface. During the second trimester of pregnancy the brain is a lissencephalic structure (Retzius 1891; Welker, 1990; Naidich et al., 1994; Armstrong et al., 1995). The brain undergoes extensive gyrification during the third trimester, and the primary and secondary fissures are readily observed at the time of birth. Following birth, the ratio of the volume of the cerebral cortex to the cortical surface area remains relatively constant throughout development (Dareste, 1862; Armstrong et al., 1995). Since the brain nearly triples its volume from birth to adulthood, the gyri and sulci continue to develop as well, maintaining this constant ratio.

Since the correlation in surface morphology between twins was only moderate, it is also possible that the discordant surface patterns develop during the postnatal period. Support for this comes from Lohmann et al. (Lohmann et al., 1999), who demonstrated that the deeper and earlier-developed sulci of the brain are more highly correlated than the superficial or tertiary sulci. In fact, the deeper structures of the central sulcus are similar even among non-related individuals (White et al., 1997). Lohmann et al. (Lohmann et al., 1999) found that the tertiary sulci, which develop mainly after birth, appear to be more affected by non-genetic influences. Bartley et al. (Bartley et al., 1997) demonstrated that the development of cortical patterns is determined primarily by random environmental factors. These findings are supported in the present study by the lower correlations for both the gyral (r = 0.63) and sulcal (r = 0.58) curvature indices, the surface complexity index (r = 0.49) and the measure of surface area (r = 0.69). It is plausible that the greater non-shared environmental influences that are present for postnatal twins, coupled with the pronounced cortical plasticity of early development, effect changes in the patterns of cortical surface morphology.

Two studies have attempted to visually identify and match MZ twin brains based on their surface morphology (Bartley et al., 1997; Biondi et al., 1998). This task would be trivial if one were matching the faces of MZ twins. Whereas the brain is highly tuned to identifying facial features (i.e. the similarities between twins), the task of sorting brains based on brain morphology proved more difficult. Although raters improved their performance with time, there were mixed results in their ability to adequately identify twin brains. Bartley et al. (Bartley et al., 1997) reported only a 50% success rate in matching MZ twin brains, whereas Biondi et al. (Biondi et al., 1998) used a larger number of surface renderings and reported a 90% success rate. Perhaps the similarities of the primary and secondary sulci, as described by Lohmann et al. (Lohmann et al., 1999), allow for enough similarities for an observer to identify MZ twin brains with reasonable accuracy if there are enough views.

Whereas the surface measures were less correlated, brain volumes were highly correlated. Three studies have evaluated heritability scores of brain volume utilizing both monozygotic and dizygotic twin pairs (Bartley et al., 1997; Carmelli et al., 1998; Pennington et al., 2000). These studies are consistent and demonstrate heritabilities for brain volume ranging from 91 to 97%. Pennington et al. (Pennington et al., 2000) parcellated the brain into specific regions of interest and performed a factor analysis. These analyses resulted in a cortical and subcortical factor that differed in measures of heritability. A lower heritability was reported for the cortical factor (56%) than for the subcortical factor (70%). A majority of the twins in their study, however, were discordant for learning disabilities.

The subcortical regions measured in the present study were slightly less correlated in twins than the cortical volumes. The correlations ranged from 0.75 for the thalamus and the putamen to 0.84 for the caudate. This may be explained by methodological limitations. The introduction of variability in measuring the smaller subcortical structures could greatly reduce the strength of the correlation, especially with a sample size of only 12 twin pairs. Whereas the reliability to accurately define the outline of the cerebral cortex and cerebellum is on the order of 0.98, the ANN was able to accurately define the subcortical ROIs with ICCs on the order of 0.80. This added variability to both the twin and control pairs could potentially result in a reduction of the overall correlation. Nevertheless, the relatively high correlation of the subcortical and cortical structures in MZ twins, coupled with the high thalamic/cortical GM correlation, reflects the substantial genetic contributions to brain development.

With the exception of WM and ventricular volume, the correlations in the age, sex, height and paternal education matched control sample were low. Since there are reports of age related decreases in GM (Gur et al., 1991; Jernigan et al., 1991a; Bartzokis et al., 2001) and subcortical volume (Raz et al., 1995; Gunning-Dixon et al., 1998), it is possible that a reduction in correlations for the controls was related to either the small sample size or inter-subject variability being on the same order of magnitude as the age, sex and stature affects. The correlation for the total cerebral WM was 0.40, and the lobar WM volumes are mild to moderately correlated (Table 4). These may reflect the gender differences in cerebral WM that have been reported previously (Passe et al., 1997). Furthermore, the high correlation found for ventricular volume in unrelated pairs may also reflect age and gender related differences (Haug, 1977; Grant et al., 1987).

High correlations exist in the twin sample for measures of total brain volume, cerebral brain volume, cortical GM, total cerebral GM and total cerebral WM (Table 2). Furthermore, when the brain is divided into its four lobes, the lobar volumes remain highly correlated for both GM and WM (Table 3). The only cortical variable with an ICC of <0.80 was the occipital lobe GM (r = 0.69). Although potentially a type II error, it is also possible that there is greater developmental plasticity in this region. The visual cortex may undergo differential patterns of development based on discordant visual signals between the twins (Wiesel, 1982). For example, individuals who are born blind will recruit the visual cortex when processing sensorimotor information (Cohen et al., 1997). Although this is an extreme example of cortical plasticity, a large number of small, discrete differences in sensory input to the brain may augment discordant developmental patterns between twins.

White matter volume also was highly correlated between twins (r = 0.98). This probably reflects a strong genetic influence on cortical myelination. Since the vast majority of the twins (10/12) were under the age of 29, and since myelination has been shown to continue to as late as 32 years of age (Yakovlev and Lecours, 1967), the dynamic changes of myelination also appear to follow similar developmental trajectories. The vast majority of myelination, however, occurs prior to 4 years of age (Yakovlev and Lecours, 1967). Although the internal boundaries of the WM are arbitrarily pre-defined by Talairach boxes, the measures of frontal, temporal, parietal and occipital WM are highly correlated between MZ twins. Only the frontal WM (r = 0.83) has a Pearson's correlation coefficient of <0.90, and this value lies well within the confidence intervals for the other WM volumes. Considering the later development of white matter tracts, it is also possible that frontal white matter is more susceptible to nongenetic influences during development. The utilization of diffusion tensor imaging techniques to twin populations may be beneficial in addressing this question.

There are a number of weaknesses in the present study. Specifically, we cannot dissect the roles of genetic and shared environmental factors or non-shared environmental factors. We were not able to use a dizygotic twin sample, and thus were not able to report measures of heritability. Since the twins had some brain regions that were very highly correlated and other regions that were less so, certain inferences can be made regarding genetic and non-genetic contribution to brain development. However, we cannot rule out that the moderate correlation in surface measures was attributed to shared environmental factors. We were able to form a closely matched sample of control pairs to add an additional comparison. However, this population lacks both shared genetic and environmental factors. Finally, the number of subjects is relatively small (n = 12). Even so, the lower limits of the confidence intervals for most measures were generally quite high.

In conclusion, the study supports findings that are in accord with current theories of the ontogeny of the cerebral development, and supports previous research demonstrating that brain size and volume are highly correlated between MZ twin pairs. In addition, we have now added new data to the twin literature, demonstrating that these similarities are also present for GM, WM, the lobar volumes of the cerebrum, the cerebellum and subcortical structures. We also demonstrate that the depth of the cortical GM is highly correlated in monozygotic twins, which has not been previously reported. Furthermore, this study also demonstrates that surface morphology is less highly correlated between MZ twins, presumably indicating that the development of sulcal and gyral characteristics are under greater nongenetic influence than other brain measures. Many different types of nongenetic influences may contribute to the plasticity of the cortical surface characteristics, such as educational experiences, physical activity or social interactions. Furthermore, probabilistic events during the complex process of neurodevelopment [e.g. connections between neurons (Muller et al., 1997)], differences in gene expression either by chance or modulated by early-immediate genes (Abraham et al., 1993; Worley et al., 1993), or variability in cell–cell interactions (Fletcher et al., 1991) may also contribute to variability between MZ twins. The role of such influences is an important subject for future investigations. Finally, exploring the surface patterns in twins longitudinally starting very early in life may provide evidence about the role of postnatal environmental influences on brain morphology.

Table 1

Demographics of twin and matched control pairs

 Twins Controls 
 Mean (SD) r Mean (SD) r 
Age (years)  24.5 (7.0) N/A  24.3 (5.9)  0.98 
Height (cm) 168 (10.2) 0.96 176 (9.7)  0.96 
Educational level (years)  14.1 (1.4) 0.96  14.7 (1.4) –0.30 
Father's educational level (years)  13.5 (3.1) N/A  12.3 (2.0)  1.0 
Mother's educational level (years)  13.5 (1.7) N/A  12.5 (1.1) –0.04 
Parental SES  2.96 (0.20) N/A  2.83 (0.48) –0.013 
 Twins Controls 
 Mean (SD) r Mean (SD) r 
Age (years)  24.5 (7.0) N/A  24.3 (5.9)  0.98 
Height (cm) 168 (10.2) 0.96 176 (9.7)  0.96 
Educational level (years)  14.1 (1.4) 0.96  14.7 (1.4) –0.30 
Father's educational level (years)  13.5 (3.1) N/A  12.3 (2.0)  1.0 
Mother's educational level (years)  13.5 (1.7) N/A  12.5 (1.1) –0.04 
Parental SES  2.96 (0.20) N/A  2.83 (0.48) –0.013 
Table 2

ICC and overlap measures for brain volumes

Structure Two technicians Artificial neural network and one technician 
 n ICC Overlap n ICC Overlap 
Two independent technicians are compared versus one technician and the artificial neural network. 
Whole brain 15 0.99 0.94 15 0.95 0.92 
Cerebellum 0.91 0.88 10 0.98 0.91 
Thalamus       
    Left 15 0.90 0.83 10 0.92 0.82 
    Right 15 0.87 0.82 10 0.90 0.83 
Caudate       
    Left 10 0.69 0.73 10 0.93 0.80 
    Right 10 0.88 0.76 10 0.86 0.80 
Putamen       
    Left 27 0.79 0.79 0.88 0.77 
    Right 27 0.69 0.78 0.71 0.74 
Structure Two technicians Artificial neural network and one technician 
 n ICC Overlap n ICC Overlap 
Two independent technicians are compared versus one technician and the artificial neural network. 
Whole brain 15 0.99 0.94 15 0.95 0.92 
Cerebellum 0.91 0.88 10 0.98 0.91 
Thalamus       
    Left 15 0.90 0.83 10 0.92 0.82 
    Right 15 0.87 0.82 10 0.90 0.83 
Caudate       
    Left 10 0.69 0.73 10 0.93 0.80 
    Right 10 0.88 0.76 10 0.86 0.80 
Putamen       
    Left 27 0.79 0.79 0.88 0.77 
    Right 27 0.69 0.78 0.71 0.74 
Table 3

Pearson correlations of brain size and tissue composition

Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Total brain tissue 1327.9 (146.8) 0.99 1355.2 (103.2) –0.03 0.0001 
Cerebrum 1154.1 (134.9) 0.99 1175.2 (98.6) –0.02 0.0001 
Cerebral GM  689.8 (73.5) 0.98  704.4 (63.1) –0.15 0.0001 
Cerebral WM  464.3 (71.1) 0.98  470.9 (50.1)  0.40 0.0001 
Cortical GM  634.0 (67.5) 0.99  648.3 (60.4) –0.14 0.0001 
Cerebellum  131.4 (14.8) 0.99  135.1 (9.2)  0.20 0.0001 
Surface CSF  39.1 (21.8) 0.91  52.8 (29.5)  0.11 0.001 
Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Total brain tissue 1327.9 (146.8) 0.99 1355.2 (103.2) –0.03 0.0001 
Cerebrum 1154.1 (134.9) 0.99 1175.2 (98.6) –0.02 0.0001 
Cerebral GM  689.8 (73.5) 0.98  704.4 (63.1) –0.15 0.0001 
Cerebral WM  464.3 (71.1) 0.98  470.9 (50.1)  0.40 0.0001 
Cortical GM  634.0 (67.5) 0.99  648.3 (60.4) –0.14 0.0001 
Cerebellum  131.4 (14.8) 0.99  135.1 (9.2)  0.20 0.0001 
Surface CSF  39.1 (21.8) 0.91  52.8 (29.5)  0.11 0.001 
Table 4

Volume measurements of GM and WM correlations for the four brain lobes

Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Frontal lobe      
    WM 173.1 (25.5) 0.83 179.0 (21.9)  0.28 0.05 
    GM 260.1 (33.6) 0.92 273.5 (28.3) –0.22 0.0001 
Parietal lobe      
    WM 115.8 (19.2) 0.97 116.0 (14.4)  0.29 0.0001 
    GM 140.8 (15.7) 0.93 140.4 (15.4) –0.20 0.0001 
Temporal lobe      
    WM  73.0 (10.6) 0.96  75.4 (9.6)  0.41 0.0005 
    GM 152.4 (13.3) 0.92 156.2 (14.5)  0.05 0.0001 
Occipital lobe      
    WM  61.4 (17.0) 0.97  56.6 (9.1)  0.51 0.0005 
    GM  70.5 (9.6) 0.69  68.1 (10.5) –0.44 0.001 
Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Frontal lobe      
    WM 173.1 (25.5) 0.83 179.0 (21.9)  0.28 0.05 
    GM 260.1 (33.6) 0.92 273.5 (28.3) –0.22 0.0001 
Parietal lobe      
    WM 115.8 (19.2) 0.97 116.0 (14.4)  0.29 0.0001 
    GM 140.8 (15.7) 0.93 140.4 (15.4) –0.20 0.0001 
Temporal lobe      
    WM  73.0 (10.6) 0.96  75.4 (9.6)  0.41 0.0005 
    GM 152.4 (13.3) 0.92 156.2 (14.5)  0.05 0.0001 
Occipital lobe      
    WM  61.4 (17.0) 0.97  56.6 (9.1)  0.51 0.0005 
    GM  70.5 (9.6) 0.69  68.1 (10.5) –0.44 0.001 
Table 5

Volume measurements of subcortical regions

Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Ventricular volume 12.6 (5.0) 0.85 12.8 (6.9)  0.52 0.1 
Caudate  5.69 (0.66) 0.84  6.08 (0.73) –0.17 0.001 
Putamen  9.99 (0.90) 0.75 10.15 (1.14)  0.29 0.11 
Thalamus 12.91 (1.28) 0.75 13.74 (1.48)  0.0 0.02 
Brain region Twins Controls P
 Mean (SD) r Mean (SD) r  
The P value is comparing the Pearson correlations between twins and controls. 
Ventricular volume 12.6 (5.0) 0.85 12.8 (6.9)  0.52 0.1 
Caudate  5.69 (0.66) 0.84  6.08 (0.73) –0.17 0.001 
Putamen  9.99 (0.90) 0.75 10.15 (1.14)  0.29 0.11 
Thalamus 12.91 (1.28) 0.75 13.74 (1.48)  0.0 0.02 
Table 6

Pearson's correlation coefficients for the surface measures

Brain region Twins (rControls (rP
The P value is comparing the Pearson correlations between twins and controls. 
aThe mean surface area for the twins group was 1.79 × 105 cm2 (SD = 2.68 104) and was 1.82 × 105 cm2 (SD = 1.47 104) for the control group. 
bThe mean cortical depth for the twin group was 4.28 mm (SD = 0.40) and was 4.26 mm (SD = 0.44) for the control group. 
Surface areaa 0.69 –0.22 0.008 
Gyral curvature 0.63  0.14 0.15 
Sulcal curvature 0.58 –0.05 0.08 
Cortical depthb 0.84  0.21 0.01 
Surface complexity 0.49 –0.16 0.09 
Brain region Twins (rControls (rP
The P value is comparing the Pearson correlations between twins and controls. 
aThe mean surface area for the twins group was 1.79 × 105 cm2 (SD = 2.68 104) and was 1.82 × 105 cm2 (SD = 1.47 104) for the control group. 
bThe mean cortical depth for the twin group was 4.28 mm (SD = 0.40) and was 4.26 mm (SD = 0.44) for the control group. 
Surface areaa 0.69 –0.22 0.008 
Gyral curvature 0.63  0.14 0.15 
Sulcal curvature 0.58 –0.05 0.08 
Cortical depthb 0.84  0.21 0.01 
Surface complexity 0.49 –0.16 0.09 
Figure 1.

Visual representation of the artificial neural net cutouts for the cerebellum (turquoise), thalamus (blue), caudate (gray), putamen (red) and globus pallidus (light green). Since the inter-rater reliability for the globus pallidus at present is poor, it was excluded from the analysis.

Figure 1.

Visual representation of the artificial neural net cutouts for the cerebellum (turquoise), thalamus (blue), caudate (gray), putamen (red) and globus pallidus (light green). Since the inter-rater reliability for the globus pallidus at present is poor, it was excluded from the analysis.

References

Abraham WC, Mason SE, Demmer J, Williams JM, Richardson CL, Tate WP, Lawlor PA, Dragunow M (
1993
) Correlations between immediate early gene induction and the persistence of long-term potentiation.
Neuroscience
 
56
:
717
–727.
Andreasen NC, Cohen G, Harris G, Cizadlo T, Parkkinen J, Rezai K, Swayze VW (
1992
) Image processing for the study of brain structure and function: problems and programs.
J Neuropsychiatry Clin Neurosci
 
4
:
125
–133.
Andreasen NC, Cizadlo T, Harris G, Swayze VW, O'Leary DS, Cohen G, Ehrhardt J, Yuh WT (
1993
) Voxel processing techniques for the antemortem study of neuroanatomy and neuropathology using magnetic resonance imaging.
J Neuropsychiatry Clin Neurosci
 
5
:
121
–130.
Andreasen NC, Rajarethinam R, Cizadlo T, Arndt S, Swayze VW, Flashman LA, O'Leary DS, Ehrhardt JC, Yuh WT (
1996
) Automatic atlas-based volume estimation of human brain regions from MR images.
J Comput Assist Tomogr
 
20
:
98
–106.
Armstrong E, Schleicher A, Omran H, Curtis M, Zilles K (
1995
) The ontogeny of human gyrification.
Cereb Cortex
 
5
:
56
–63.
Bartley AJ, Jones DW, Weinberger DR (
1997
) Genetic variability of human brain size and cortical gyral patterns.
Brain
 
120
:
257
–269.
Bartzokis G, Beckson M, Lu PH, Nuechterlein KH, Edwards N, Mintz J (
2001
) Age-related changes in frontal and temporal lobe volumes in men: a magnetic resonance imaging study.
Arch Gen Psychiatry
 
58
:
461
–465.
Biondi A, Nogueira H, Dormont D, Duyme M, Hasboun D, Zouaoui A, Chantome M, Marsault C (
1998
) Are the brains of monozygotic twins similar? A three-dimensional MR study.
AJNR Am J Neuroradiol
 
19
:
1361
–1367.
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
–228.
Carmelli D, DeCarli C, Swan GE, Jack LM, Reed T, Wolf PA, Miller BL (
1998
) Evidence for genetic variance in white matter hyperintensity volume in normal elderly male twins.
Stroke
 
29
:
1177
–1181.
Caviness VS Jr (
1982
) Early events of neocortical assembly: experimental studies and human pathology.
Int J Neurol
 
17
:
102
–109.
Cohen G, Andreasen NC, Alliger R, Arndt S, Kuan J, Yuh WT, Ehrhardt J (
1992
) Segmentation techniques for the classification of brain tissue using magnetic resonance imaging.
Psychiatry Res
 
45
:
33
–51.
Cohen LG, Celnik P, Pascual-Leone A, Corwell B, Falz L, Dambrosia J, Honda M, Sadato N, Gerloff C, Catala MD, Hallett M (
1997
) Functional relevance of cross-modal plasticity in blind humans.
Nature
 
389
:
180
–183.
Collins DL, Neelin P, Peters TM, Evans AC (
1994
) Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space.
J Comput Assist Tomogr
 
18
:
192
–205.
Dareste MC (
1862
) Sur les rapports de la masse encéphalique avec le développment de l'intelligence.
Bull Soc Anthropol Paris
 
3
:
26
–54.
Fletcher TL, Cameron P, De Camilli P, Banker G (
1991
) The distribution of synapsin I and synaptophysin in hippocampal neurons developing in culture.
J Neurosci
 
11
:
1617
–1626.
Giedd JN, Snell JW, Lange N, Rajapakse JC, Casey BJ, Kozuch PL, Vaituzis AC, Vauss YC, Hamburger SD, Kaysen D, Rapoport JL (
1996
)Quantitative magnetic resonance imaging of human brain development: ages 4–18.
Cereb Cortex
 
6
:
551
–560.
Grant R, Condon B, Lawrence A, Hadley DM, Patterson J, Bone I, Teasdale GM (
1987
) Human cranial CSF volumes measured by MRI: sex and age influences.
Magn Reson Imaging
 
5
:
465
–468.
Gunning-Dixon FM, Head D, McQuain J, Acker JD, Raz N (
1998
) Differential aging of the human striatum: a prospective MR imaging study.
AJNR Am J Neuroradiol
 
19
:
1501
–1507.
Gur RC, Mozley PD, Resnick SM, Gottlieb GL, Kohn M, Zimmerman R, Herman G, Atlas S, Grossman R, Berretta D, et al. (
1991
) Gender differences in age effect on brain atrophy measured by magnetic resonance imaging.
Proc Natl Acad Sci USA
 
88
:
2845
–2849.
Harris G, Andreasen NC, Cizadlo T, Bailey J, Bockholt J, Travis K, Arndt S (
1997
) Improving tissue segmentation in magnetic resonance imaging: multiple pulse sequences and automated training class selection.
NeuroImage
 
5
:
402
.
Harris G, Andreasen NC, Cizadlo T, Bailey JM, Bockholt HJ, Magnotta VA, Arndt S (
1999
) Improving tissue classification in MRI: a three-dimensional multispectral discriminant analysis method with automated training class selection.
J Comput Assist Tomogr
 
23
:
144
–154.
Haug G (
1977
) Age and sex dependence of the size of normal ventricles on computed tomography.
Neuroradiology
 
14
:
201
–204.
Hockfield S, Lombroso PJ (
1998
) Development of the cerebral cortex: IX. Cortical development and experience: I.
J Am Acad Child Adolesc Psychiatry
 
37
:
992
–993.
Hockfield S, Lombroso PJ (
1998
) Development of the cerebral cortex: X. Cortical development and experience: II.
J Am Acad Child Adolesc Psychiatry
 
37
:
1103
–1105.
Huttenlocher PR (
1990
) Morphometric study of human cerebral cortex development.
Neuropsychologia
 
28
:
517
–527.
Jernigan TL, Archibald SL, Berhow MT, Sowell ER, Foster DS, Hesselink JR (
1991
) Cerebral structure on MRI, Part I: Localization of age-related changes.
Biol Psychiatry
 
29
:
55
–67.
Jernigan TL, Trauner DA, Hesselink JR, Tallal PA (
1991
) Maturation of human cerebrum observed in vivo during adolescence.
Brain
 
114
:
2037
–2049.
Jones KL (1997) Smith's recognizable patterns of human malformation, 5th edn. Philadelphia, PA: W.B. Saunders.
Kempermann G, Kuhn HG, Gage FH (
1997
) More hippocampal neurons in adult mice living in an enriched environment.
Nature
 
386
:
493
–495.
Kjaer I (
1995
) Human prenatal craniofacial development related to brain development under normal and pathologic conditions.
Acta Odontol Scand
 
53
:
135
–143.
Knudsen EI (
1998
) Capacity for plasticity in the adult owl auditory system expanded by juvenile experience.
Science
 
279
:
1531
–1533.
Leckman JF, Lombroso PJ (
1998
) Development of the cerebral cortex: IV. Transcription factors.
J Am Acad Child Adolesc Psychiatry
 
37
:
451
–452.
Lohmann G, von Cramon DY, Steinmetz H (
1999
) Sulcal variability of twins.
Cereb Cortex
 
9
:
754
–763.
Lombroso PJ (
1998
) Development of the cerebral cortex: VI. growth factors: I.
J Am Acad Child Adolesc Psychiatry
 
37
:
674
–675.
Machin GA (
1996
) Some causes of genotypic and phenotypic discordance in monozygotic twin pairs.
Am J Med Genet
 
61
:
216
–228.
Machin GA, Sperber GH, Wootliffe J (
1985
) Monozygotic twin aborted fetuses discordant for holoprosencephaly/synotia.
Teratology
 
31
:
203
–215.
Magnotta VA, Andreasen NC, Schultz SK, Harris G, Cizadlo T, Heckel D, Nopoulos P, Flaum M (
1999
) Quantitative in vivo measurement of gyrification in the human brain: changes associated with aging.
Cereb Cortex
 
9
:
151
–160.
Magnotta VA, Heckel D, Andreasen NC, Cizadlo T, Corson PW, Ehrhardt JC, Yuh WT (
1999
) Measurement of brain structures with artificial neural networks: two- and three-dimensional applications.
Radiology
 
211
:
781
–790.
Muller TH, Swandulla D, Zeilhofer HU (
1997
) Synaptic connectivity in cultured hypothalamic neuronal networks.
J Neurophysiol
 
77
:
3218
–3225.
Naegele J, Lombroso PJ (
1998
) Development of the cerebral cortex: VIII. Apoptosis: neuronal Hari-Kari.
J Am Acad Child Adolesc Psychiatry
 
37
:
890
–892.
Naidich TP, Grant JL, Altman N, Zimmerman RA, Birchansky SB, Braffman B, Daniel JL (
1994
) The developing cerebral surface. Preliminary report on the patterns of sulcal and gyral maturation–anatomy, ultrasound, and magnetic resonance imaging.
Neuroimaging Clin N Am
 
4
:
201
–240.
Oldfield RC (
1971
) The assessment and analysis of handedness: The Edinburgh Inventory.
Neuropsychologia
 
9
:
97
–113.
Oppenheim JS, Skerry JE, Tramo MJ, Gazzaniga MS (
1989
) Magnetic resonance imaging morphology of the corpus callosum in monozygotic twins.
Ann Neurol
 
26
:
100
–104.
Passe TJ, Rajagopalan P, Tupler LA, Byrum CE, MacFall JR, Krishnan KR (
1997
) Age and sex effects on brain morphology.
Prog Neuropsychopharmacol Biol Psychiatry
 
21
:
1231
–1237.
Pennington BF, Filipek PA, Lefly D, Chhabildas N, Kennedy DN, Simon JH, Filley CM, Galaburda A, DeFries JC (
2000
) A twin MRI study of size variations in human brain.
J Cogn Neurosci
 
12
:
223
–232.
Peters A, Morrison JH, Rosene DL, Hyman BT (
1998
) Feature article: are neurons lost from the primate cerebral cortex during normal aging?
Cereb Cortex
 
8
:
295
–300.
Pollin W, Stabenau JR, Mosher L, Tupin J (
1966
) Life history differences in identical twins discordant for schizophrenia.
Am J Orthopsychiatry
 
36
:
492
–509.
Rakic P (
1974
) Neurons in rhesus monkey visual cortex: systematic relation between time of origin and eventual disposition.
Science
 
183
:
425
–427.
Rakic P (
1988
) Specification of cerebral cortical areas.
Science
 
241
:
170
–176.
Rakic P (
1995
) A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution.
Trends Neurosci
 
18
:
383
–388.
Rakic P, Lombroso PJ (
1998
) Development of the cerebral cortex: I. Forming the cortical structure.
J Am Acad Child Adolesc Psychiatry
 
37
:
116
–117.
Raz N, Torres IJ, Acker JD (
1995
) Age, gender, and hemispheric differences in human striatum: a quantitative review and new data from in vivo MRI morphometry.
Neurobiol Learn Mem
 
63
:
133
–142.
Retzius A (
1891
) Ueber den Bau der Oberflachenschicht der Grosshirnrinde beim Menschen und bei den Saugethieren.
Verh Biol Ver
 
3
:
90
–103.
Rousseau F, Heitz D, Tarleton J, MacPherson J, Malmgren H, Dahl N, Barnicoat A, Mathew C, Mornet E, Tejada I, et al. (
1994
) A multicenter study on genotype–phenotype correlations in the fragile X syndrome, using direct diagnosis with probe StB12.3: the first 2, 253 cases.
Am J Hum Genet
 
55
:
225
–237.
Rubenstein JL (
1998
) Development of the cerebral cortex: V. Transcription factors and brain development.
J Am Acad Child Adolesc Psychiatry
 
37
:
561
–562.
Rubenstein JL, Rakic P (
1999
) Genetic control of cortical development.
Cereb Cortex
 
9
:
521
–523.
Schinzel AA, Smith DW, Miller JR (
1979
) Monozygotic twinning and structural defects.
J Pediatr
 
95
:
921
–930.
Sidman RL, Rakic P (
1973
) Neuronal migration, with special reference to developing human brain: a review.
Brain Res
 
62
:
1
–35.
Sowell ER, Thompson PM, Holmes CJ, Batth R, Jernigan TL, Toga AW (
1999
) Localizing age-related changes in brain structure between childhood and adolescence using statistical parametric mapping.
Neuroimage
 
9
:
587
–597.
Sowell ER, Thompson PM, Holmes CJ, Jernigan TL, Toga AW (
1999
) In vivo evidence for post-adolescent brain maturation in frontal and striatal regions.
Nat Neurosci
 
2
:
859
–861.
Sperber GH (
1992
) First year of life: prenatal craniofacial development.
Cleft Palate Craniofac J
 
29
:
109
–111.
Stabenau JR, Pollin W (
1967
) Early characteristics of monozygotic twins discordant for schizophrenia.
Arch Gen Psychiatry
 
17
:
723
–734.
Steinmetz H, Herzog A, Schlaug G, Huang Y, Jancke L (
1995
) Brain (A) symmetry in monozygotic twins.
Cereb Cortex
 
5
:
296
–300.
Stuart A, Ord JK (1987) Kendall's advanced theory of statistics, 5 edn. New York: Oxford University Press.
Talairach J, Tournoux P (1988) Co-planar stereotactic atlas of the human brain: 3-dimensional proportional system: an approach to cerebral imaging. Stuttgart: Thieme Verlag.
Tramo MJ, Loftus WC, Thomas CE, Green RL, Mott LA, Gazzaniga MS (
1995
) Surface area of human cerebral cortex and its gross morphological subdivisions: in vivo measurements in monozygotic twins suggest differential hemisphere effects of genetic factors.
J Cogn Neurosci
 
7
:
292
–301.
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
:
1246
–1252.
Vaccarino F, Lombroso PJ (
1998
) Development of the cerebral cortex: VII. Growth factors: II.
J Am Acad Child Adolesc Psychiatry
 
37
:
789
–790.
Weinberger DR, Zigun JR, Bartley AJ, Jones DW, Torrey EF (
1992
) Anatomical abnormalities in the brains of monozygotic twins discordant and concordant for schizophrenia.
Clin Neuropharmacol
 
15
:
122A
–123A.
Welker W (1990) Why does cerebral cortex fissure and fold. In: Cerebral cortex (Jones EG, Peters A, eds), pp. 3–136. New York: Plenum Press.
White LE, Andrews TJ, Hulette C, Richards A, Groelle M, Paydarfar J, Purves D (
1997
) Structure of the human sensorimotor system: I. Morphology and cytoarchitecture of the central sulcus.
Cereb Cortex
 
7
:
18
–30.
Wiesel TN (
1982
) Postnatal development of the visual cortex and the influence of environment.
Nature
 
299
:
583
–591.
Winter RM (
1996
) What's in a face?
Nat Genet
 
12
:
124
–129.
Worley PF, Bhat RV, Baraban JM, Erickson CA, McNaughton BL, Barnes CA (
1993
) Thresholds for synaptic activation of transcription factors in hippocampus: correlation with long-term enhancement.
J Neurosci
 
13
:
4776
–4786.
Yakovlev PI, Lecours AR (1967) The myelogenetic cycles of regional maturation of the brain. In: Regional development of the brain in early life (Minkowski A, ed.), pp. 3–70. Oxford: Blackwell.
Zilles K, Armstrong E, Schleicher A, Kretschmann HJ (
1988
) The human pattern of gyrification in the cerebral cortex.
Anat Embryol
 
179
:
173
–179.