The nature and extent of callosal morphological alterations inschizophrenia remain unresolved. A parametric surface modelingapproach using magnetic resonance (MR) images was employed.This provided spatially accurate representations of midsagittalcallosal surfaces in schizophrenic patients (n = 25; 15 males) andnormal controls (n = 28; 15 males). Areas of functionally relevantcallosal channels and measures reflecting callosal shape werevisualized and compared across groups. To register neuroanatomicallandmarks surrounding the corpus callosum, each three-dimensionalMR volume was scaled according to Talairach AC–PC distance, andraw distances included as covariates in multivariate analyses.Results revealed: (i) a marked vertical displacement of the corpuscallosum in patients (P < 0.01); (ii) increases in curvature ofsuperior and inferior callosal surfaces (P < 0.001); and (iii) signif-icant increases in maximum widths in anterior and posterior regionsin male patients compared to male controls; as well as (iv) increasedpatterns of callosal variability in female patients but no effects ofdiagnosis between female groups. These findings demonstrate aclear index of structural neuropathology in male schizophrenicpatients. Displacement and curvature increases were highly cor-related with structural differences in surrounding neuroanatomicalregions, including increased volume of the lateral ventricles (P <0.01).

IntroductionThe corpus callosum plays an integral role in relaying sensory,

motor and cognitive information from homologous regions in

the two cerebral hemispheres. Given that cognitive impairments

and both bilateral and unilateral neuroanatomic structural

abnormalities are characteristic in schizophrenia, the mor-

phology of the corpus callosum has stimulated much research

(Nasrallah et al., 1986; Uematsu et al., 1988; Rossi et al., 1989;

Woodruff et al., 1993, 1997; Jacobsen et al., 1997). Most

investigations have used ‘region of interest’ analyses to assess

differences in midsagittal area, callosal length and width

(Woodruff et al., 1995), while few have investigated group

differences in callosal shape or in three-dimensional location

(Casanova et al., 1990; DeQuardo et al., 1996). Volumetric

methods are less sensitive to across-subject variability and less

likely capture subtle differences in anatomy between groups.

Probabilistic brain atlases using magnetic resonance (MR)

images retain quantitative information on interindividual neuro-

anatomical variability. In these approaches three-dimensional

brain volumes are registered (realigned) in a standardized

coordinate system to allow cross-group comparisons while

preserving subtle intragroup variability patterns (Mazziotta et

al., 1995; Thompson et al., 1997, 1999a,b).

Several observations motivate the study of the corpus

callosum in schizophrenia (Thompson et al., 1999a): (i) callosal

pruning and myelination as well as interhemispheric coherence

continue to develop into early adulthood, a factor that may be

relevant to age of onset in schizophrenia (Njiokiktjien et al.,

1994); (ii) impairments in callosal transfer have been reported in

patients (David et al., 1994), implicating alterations in callosal

connectivity; (iii) structural alterations in asymmetric peri-

sylvian regions linked by the callosum have been reported

(Kikinis et al., 1994; Barta et al., 1997); (iv) callosal myelination

begins prenatally and is susceptible to malnutrition, asphyxia

and toxins of infectious origin; also, these same events are linked

with aberrant neurodevelopmental events in schizophrenia

(Njiokiktjien et al., 1994; Bishop and Wahlsten, 1997; Hoek et

al., 1998); (v) the corpus callosum forms the roof of the superior

horns of the lateral ventricles, which are enlarged in schizo-

phrenic patients (Sharma et al., 1997, 1998). In spite of evidence

linking callosal abnormality to schizophrenia, imaging studies

assessing alterations in callosal morphometry as well as in other

cortical and subcortical structures have produced surprisingly

mixed results (Woodruff et al., 1995). Interpretation of these

results is further obscured by the failure of many studies to

control for brain size in patient and normal populations. Since

callosal size and brain size are correlated (Rauch and Jinkins,

1994; Jäncke et al., 1997), and brain size is reported to be

decreased in MR studies of schizophrenic patients (Zipursky et

al., 1992; Lawrie and Abukmeil, 1998), it is clear headsize

corrections are necessary.

A surface deformation-based approach was employed in this

study to allow quantification of standard callosal parameters

such as area, length and width in addition to other parameters

indexing shape (Thompson et al., 1996a,b, 1997). This approach

was aimed at providing information about structural alterations

in surrounding brain regions in addition to assessing callosal

abnormalities. After some initial scaling to bring callosal surfaces

into register and to correct for differences in overall brain

dimensions, curves and points from the callosal surfaces of

subjects in each group were matched and variability information

from each individual surface warp retained. The scaling ratios

used to align the corpus callosum across subjects can be factored

out either by converting measures back into native space or by

using them as covariates in a multivariate analysis, thereby

ensuring that callosal measures are not affected by brain volume.

These techniques allow a comprehensive analysis of callosal

structure in schizophrenic patients and normal controls and

provide the ability to quantify and visualize any structural

differences as they relate to diagnosis and sex in a three-

dimensional coordinate system in standardized and native

scanner space. Furthermore, callosal shape renderings from

Alzheimer’s patients (n = 10 males) acquired using the same

methods in an earlier study (Thompson et al., 1998) were

compared with the callosal shape renderings of male schizo-

phrenic patients and controls to evaluate whether callosal shape

profiles in schizophrenia are disease specific.

Cerebral Cortex Jan 2000;10:40–49; 1047–3211/00/$4.00

Mapping Morphology of the CorpusCallosum in Schizophrenia

Katherine L. Narr, Paul M. Thompson, Tonmoy Sharma1,

Jacob Moussai, Andrew F. Cannestra and Arthur W. Toga

Laboratory of Neuro Imaging, Department of Neurology,

Division of Brain Mapping, UCLA School of Medicine, Los

Angeles, CA, USA and 1Section of Cognitive

Psychopharmacology, Institute of Psychiatry, Denmark Hill,

London, UK

© Oxford University Press 2000

Hypotheses

We set out to detect a modest decrease in midsagittal area in

schizophrenic patients, looking specifically for area differences

implicating involvement of specific callosal channels as defined

by Witelson et al. (Witelson et al., 1989) and Clarke et al. (Clarke

et al., 1994). Furthermore, we were interested in assessing

alterations in callosal shape measured by curvature and dis-

placement in three-dimensional space in patients versus

controls, which would provide further hypotheses concerning

structural differences in surrounding anatomical regions.

Specifically, relationships between callosal displacement and

lateral and third ventricular enlargement were assessed.

Materials and Methods

Subjects

Twenty-five chronic schizophrenic patients (15 male/10 female; mean

age 31.1 ± 5.6 years) and 28 normal controls (15 male/13 female; mean

age 30.5 ± 8.7 years) were included in the study. Subjects were scanned at

the Institute of Psychiatry, London, UK. Groups did not differ significantly

in age, years of education, height or parental socio-economic class.

Socio-economic status was derived from UK census data (Office of

Population Censuses and Surveys, 1991), using details of ‘best-ever’

parental occupation. There were two left-handed subjects in each group

as defined by the Annett Handedness Scale (Annett, 1970).

All schizophrenic patients met DSM-III-R (American Psychiatric

Association, 1987) criteria and were receiving regular antipsychotic

medication. Control subjects were screened for any personal or family

history of psychiatric illness. Exclusion criteria for both patients and

controls included head trauma, drug abuse and hereditary neurological

disorders. All subjects gave informed written consent for their

participation in this study with ethical permission obtained from the

Bethlem and Maudsley Ethical Committee (Research).

Image Analysis Procedures

High-resolution three-dimensional SPGR MR images (256 × 256 matrix;

20 cm FOV) were acquired using a GE Signa 1.5 T scanner for each

subject as a series of 124 contiguous 1.5 mm coronal slices. Images were

resliced in the sagittal plane at 0.5 mm. To correct for head position and

brain length and to allow for morphometric comparison of the corpus

callosum between subjects, each brain volume was scaled and placed into

the standardized Talairach and Tournoux stereotaxic coordinate system

(Talairach and Tournoux, 1988). In the scaling procedure x, y and z pixel

coordinates were chosen for the anterior commisure (AC), posterior

commisure (PC) and midline. The scan is digitally realigned to place the

AC at the origin (0,0,0) and the PC at (–23.5,0,0) with the inter-

hemispheric fissure aligned vertically at midline. This process brings

midline structures into register according to Talairach AC–PC distance

and provides sufficient alignment to compare corpus callosum para-

meters, and derive precise statistical information about positional and

geometric variability (Thompson et al. 1996a,b, 1997) (Figs 1–5).

The corpus callosum was contoured by technicians blind to diagnosis

in each magnified midsagittal slice by following tissue boundaries using a

mouse-driven cursor. Midsagittal sections were defined by identifying the

interhemispheric fissure in native contiguous 1.5 mm coronal slices and

again in the sagittal plane when volumes were resliced at 0.5 mm after

AC–PC scaling. Coronal midlines were chosen independently by two

investigators and confirmed by the presence of the falx cerebri and

septum pellucidum in the sagittal planes. The reslicing procedure

ensured that callosi were contoured at midline within the limits of our

high-resolution MR data. Reliability of contouring was evaluated by

repeatedly outlining the corpus callosum (six times) in a randomly

chosen brain (contouring r.m.s. error < 1 mm).

Callosal surfaces were reconstructed and within- and between-group

variations in callosal shape parameters quantified using a surface-based

anatomical modeling approach (Thompson et al., 1996a,b, 1997). Brief ly,

algorithms generate a parametric grid after uniformly redigitizing points

from the callosal outline of each three-dimensional data set to form a

regular rectangular mesh that is stretched over the callosal surface

contour. To obtain average parametric meshes and measures of variability

in the shape of the callosal surfaces within each group, the parametric

grids from each surface are matched at corresponding points and the

mean and variation between these sets of points are calculated. The

surface mesh algorithms thus create average representations of callosal

surfaces in three-dimensional space and index the amount of variability

present within each group (Figs 1–4; Fig. 6). To visualize and quantify

regional displacement of average callosal surfaces between groups,

algorithms were used whereby corresponding grid points were matched

and the discrepancy in distance calculated as three-dimensional dis-

placement vectors between points (Fig. 5). Callosum meshes were

divided into superior and inferior pieces to generate surface mesh varia-

bility and displacement maps. Measures of length, width, curvature and

all three-dimensional extents of the corpus callosum were obtained from

the uniformly redigitized grid points in pixel coordinates translated into

millimeters.

Finally, midsagittal callosal renderings where divided into five vertical

partitions representing the (i) anterior third, (ii) anterior body, (iii)

posterior body, (iv) isthmus and (v) splenium (Fig. 7) (Witelson et al.,

1989; Clarke and Zaidel, 1994). Midsagittal areas were obtained in mm2

for the entire callosum and for each midsagittal segment.

Brain and Ventricle Volume

Total brain volume and volumes of the lateral and third ventricles were

acquired by segmenting each three-dimensional image set into gray and

white matter, cerebrospinal f luid (CSF) and background. Brief ly, image

data (16 bit) was processed using a number of steps that included: (i)

correction for signal intensity inhomogeneities; (ii) tissue classification by

choosing representative signal intensity values for the different tissue

types from 120 predefined regions across the entire brain volume (Sowell

et al., 1999); (iii) removal of the skull, cerebellum and extracortical CSF;

and (iv) calculation of brain volume in native space. Reliability of

selection for representative tissue type intensity values was evaluated by

classifying 10 different brains and computing intensity errors (reliability

> 94% for all tissue types). Volume of the lateral and third ventricles were

measured by outlining the regions of interest in consecutive tissue

segmented coronal slices by following anatomical landmarks and

restricted to boundaries between CSF and other tissue types (Fig. 8).

Alzheimer’s Data

To compare the shape profiles of midsagittal callosal surface renderings

between male schizophrenic patients (n = 15), adult controls (n = 15) and

Alzheimer’s patients (n = 10), Alzheimer’s data was included from a

study assessing cortical variability in normal aging and Alzheimer’s

disease (Thompson et al., 1998). Image analysis procedures employed

in this study were identical to those used here and callosal shape para-

meters were measured in the same stereotaxic coordinate system. The

rationale for including this data was to evaluate whether Alzheimer’s

patients exhibit similar differences in callosal shape profiles given that

Alzheimer’s patients show ventricular enlargement and decreases in brain

size that are similarly reported in schizophrenia (Lawrie and Abukmeil,

1998). Callosal renderings from all male groups are overlaid for com-

parison in Figure 9.

Statistical Analyses

Differences in discrete callosal areas and morphometric shape parameters

obtained from the mesh pieces were compared within and between

groups using multivariate analyses of covariance (MANCOVAs) and, when

appropriate, followed by univariate analyses. Multivariate test statistics

control for type I error by adopting a strict criterion for significance. In

order to control for type I error, tests of individual dependent variables

were only performed if the corresponding multivariate tests were found

significant. In these analyses Sex and Diagnosis were used as independent

variables, while curvature and superior and inferior extents of top and

bottom callosal surfaces, length, width and area measures were used as

dependent variables. Talairach scaling ratios used to normalize AC–PC

distance and brain volume were assessed as potential covariates. Results

of these analyses, summarized in Table 1, are described below.

ResultsSeveral striking trends emerged in the callosal morphometric

Cerebral Cortex Jan 2000, V 10 N 1 41

analyses between groups: (i) different patterns of variability

were observed in each group with female schizophrenic patients

showing the highest variability overall; (ii) significant vertical

displacement of the corpus callosum was observed in male

patients compared to controls; (iii) male patients exhibited a

significant increase in curvature for superior and inferior

callosal surfaces; and (iv) increases in callosal thickness were

identified in anterior and posterior callosal regions in male

patients versus male controls. Finally, callosal shape measures

were significantly related to lateral ventricle enlargement.

Table 1Summary of significant results

Dependent variable Effect Mean ± SEM Probability

Raw brain volume (univariate effects) Sex F: 1161.5 ± 18.1 M: 1270.9 ± 21.5 <0.001w/ covariate Sex <0.01AC–PC distance (univariate effects) Sex M: 26.83 ± 0.24 F: 25.68 ± 0.29 <0.006Midsagittal area (multivariate effects) AC–PC covariate <0.001Shape parameters (multivariate effects) Diagnosis <0.053

Diagnosis × Shape variable <0.001Diagnosis × Sex × Shape <0.008AC–PC <0.001AC–PC by Shape <0.001

Length (univariate effects) AC–PC Distance <0.001Top superior boundary a AC-PC distance <0.001

Diagnosis NC: 24.84 ± 0.62 SZ: 26.78 ± 0.58 <0.009Diagnosis × Sex <0.005Male patients > male controls NC M: 23.27 ± 0.56 SZ M: 27.38 ± 0.67 <0.001Diagnosis × Sex w/o Covariate <0.05

Bottom superior boundarya AC-PC distance <0.001Diagnosis NC: 19.67 ± 0.57 SZ: 21.44 ± 0.55 <0.016Diagnosis × Sex <0.008Male patients > male controls NC M: 18.29 ± 0.52 SZ M: 22.03 ± 0.73 <0.001

Inferior rostrum Diagnosis NC: 2.78 ± 0.28 SZ: 2.84 ± 0.34 <0.033Top curvaturea Diagnosis × Sex <0.001

Male patients > male controls NC M: 1.29 ± 0.01 SZ M: 1.33 ± 0.01 <0.001Bottom curvaturea Diagnosis × Sex <0.004

Male patients > male controls NC M: 1.50 ± 0.02 SZ M: 1.63 ± 0.02 <0.001Width (multivariate effects) AC–PC distance <0.001

Diagnosis × Sex <0.004Diagnosis × Width <0.009

Anterior segment (univariate effects) Diagnosis NC: 19.88 ± 0.72 SZ: 21.94 ± 0.82 <0.006Diagnosis × Sex <0.016Male patients > male controls NC M: 18.63 ± 0.76 SZ M: 22.71 ± 0.73 <0.05

Splenium Male patients > male controls NC M: 15.05 ± 0.16 SZ M: 16.70 ± 0.56 <0.001Posterior midbody Male patients > male controls NC M: 5.97 ± 0.25 SZ M: 7.07 ± 0.26 <0.006Ventricle volume (multivariate effects) Diagnosis × Sex <0.01w/ brain volume covariate Diagnosis × Sex × Volume <0.02

Brain volume covariate <0.05Lateral ventriclesa (univariate effects) Diagnosis × Sex <0.01

Male patients > male controls NC M: 10.70 ± 1.60 SZ M: 20.17 ± 2.65 <0.001Third ventricle Diagnosis × Sex <0.03w/o covariate Male patients > male controls NC M: 1.86 ± 0.13 SZ M: 2.58 ± 0.24 <0.01

SZ, schizophrenic patients; NC; normal controls; M, males; F, females. Multivariate analyses were performed with Sex and Diagnosis as factors, and included callosal areas, shape parameters or ventricularvolumes as dependent variables using AC–PC distance or raw brain volume as covariates. Univariate analyses were only performed if the corresponding multivariate test was significant.

Measures of volumes are reported in cm3, and distance measures are in mm.aSignificant correlations between lateral ventricle volume and callosal shape parameters (P < 0.01).

Figure 1. Variability maps of the corpus callosum in female schizophrenic patients and female controls. Average surface meshes reveal unique patterns of variability within eachgroup. The color bar encodes the root mean square magnitude (in millimeters) of the displacement vectors required to create the average callosum meshes. Female patients (n =10)possess higher variability, especially in posterior regions, compared to female controls (n =13).

Figure 2. Variability maps of the callosum in male patients and male controls. The corpus callosum appears more bowed in the vertical axis of the brain in male schizophrenic patientsreflecting a significant displacement in the inferior–superior direction as well as significant differences in callosal curvature across groups (see Results). Overall, variability is higher inmale patients (n = 15) versus male controls (n = 15).

Figure 3. Variability maps superimposing male control and female control callosal surface averages from Figures 1 and 2.

Figure 4. Variability maps superimposing male patient and female patient callosal surface averages from Figures 1 and 2.

Figure 5. Displacement map of the corpus callosum in male schizophrenic patients and male controls: Average surface meshes representing callosal segments are subtractedacross groups in three-dimensional stereotaxic space to reflect intergroup displacement in discrete callosal regions. The color bar encodes the magnitude of the three-dimensionaldisplacement vectors in millimeters. The most pronounced between group displacement can be visualized in anterior and anterior-midbody regions (5 mm). These morphologicaldifferences are compatible with bowing of the callosum in patients.

42 Callosal Morphology in Schizophrenia • Narr et al.

Cerebral Cortex Jan 2000, V 10 N 1 43

Variability and Displacement Maps

Variability maps revealed distinct differences in the patterns of

local variability and shape in each group (Figs 1–4). The average

callosal surface mesh appeared considerably more bowed in the

male schizophrenic group (Fig. 2). Overall, female schizophrenic

patients displayed higher intragroup variability in callosal

parameters than male patients or controls (∼4.5 mm) (Figs 1 and

4). Male patients exhibited higher callosal variability compared

with male controls (Fig. 2). Maximal displacement of callosal

variability averages was visualized in anterior, superior and

inferior midbody regions between male patients and controls,

approaching 5 mm (Fig. 5).

Covariates

Raw AC–PC distances (anterior–posterior axis) and brain volume

were evaluated as potential covariates to remove the effects of

head size and of AC–PC scaling from the data. Preliminary

analyses found AC–PC scaling to be highly correlated with

midsagittal callosal area and shape parameters after normal-

ization (average r = 0.48, P < 0.005). Brain volume, however, did

not correlate significantly with any of the AC–PC scaled

dependent variables (correlations did not exceed 0.15 for any

variable, P > 0.3); thus only raw AC–PC distance was used as a

covariate in subsequent analyses of callosal parameters.

To determine whether the four groups differed in average

values on the AC–PC covariate or brain volume, the AC–PC

scaling measure and brain volumes in native space were used as

dependent variables in separate 2 (Sex) × 2 (Diagnosis) ANOVAs.

There was a significant effect of Sex on AC–PC distance [F(1,49)

= 8.42, P < 0.006], with males (mean ± SEM = 26.84 ± 0.24) having

larger raw AC–PC distances compared to females (mean = 25.72

± 0.29). There were no Diagnosis or Sex × Diagnosis interactions

for brain volume. Brain volume, however, was significantly

dependent upon Sex (males larger, P < 0.001).

Midsagittal Area

A 2 (Sex) × 2 (Diagnosis) MANCOVA analysis confirmed signif-

icance of the AC–PC covariate [F(1,48) = 29.68, P < 0.001]. There

were no further main effects. When AC–PC was excluded as a

covariate the Sex effect approached significance [F(1,50) = 3.63,

P = 0.062], indicating the covariate corrected for the difference

in raw AC–PC distances across Sex.

Shape Parameters

A similar MANCOVA was performed to determine whether

group differences existed in parameters characterizing the

shape of the corpus callosum, including top and bottom curvat-

ure, apices of the superior and inferior callosal surfaces, the most

inferior part of the rostrum and callosal length as dependent

variables. This analysis revealed a marginal multivariate effect of

Diagnosis [F(1,48) = 3.925, P = 0.053] and a marginal multi-

variate Diagnosis × Sex interaction for all shape parameters (P =

0.096). The Diagnosis × Sex interaction, however, differed

significantly for the shape variables [F(6,288) = 2.968, P < 0.008].

The effect of Diagnosis also differed for the dependent measures

[F(6,288) = 62.489, P < 0.001]. These results justified conducting

univariate F-tests to examine Diagnosis × Sex interactions

separately for each variable listed above. Finally, the AC–PC

multivariate effect was highly significant [F(1,48) = 25.34, P <

0.001], but the covariate interacted with the different shape

parameters [F(6,288) = 13.386, P < 0.001].

Thus univariate ANOVAs were first run to determine which of

the shape parameters were significantly affected by the co-

variate. Of these, length [F(1,48) = 22.79, P < 0.001] and the most

superior limits of the top [F(1,48) = 22.24, P < 0.001] and bottom

callosal surfaces, [F(1,48) = 15.76, P < 0.001], were significantly

affected by AC–PC distance. Since the covariate did not account

for a significant portion of the variance for curvature or inferior

surface measures, use of raw AC–PC distance as a covariate was

considered inappropriate for these dependent variables and they

were reanalyzed using ANOVAs without the covariate, again

assessing Diagnosis × Group effects.

Effects of Diagnosis

There were significant effects of Diagnosis for the top and

bottom callosal surface meshes (most superior points) [F(1,48) =

7.34, P < 0.001 and F(1,48) = 6.25, P < 0.016 respectively], and

for the rostrum of the corpus callosum [F(1,48) = 4.84, P < 0.033]

with callosal location displaced in the vertical axis (y-domain) in

three-dimensional space in patients versus controls for all three

measures. Diagnosis × Sex interactions were apparent for the

curvature of top [F(1,49) = 19.49, P < 0.001] and bottom [F(1,49)

= 9.30, P < 0.004] callosal surface meshes. The superior

boundaries of bottom and top callosal surfaces showed Diagnosis

× Sex interactions even after covariate correction [F(1,48) =

7.69, P < 0.008 and F(1,48) = 8.85, P < 0.005 respectively]. In all

cases male patients had higher means ref lecting callosal dis-

placement in midbody regions. Finally, in univariate tests

assessing the assumption of equality for the effect of the

covariate on the four groups, there was a significant Diagnosis ×

Sex × Slope interaction for the superior extent of the bottom

callosal surface. This measure was reanalyzed without the

covariate, but again produced a significant Diagnosis × Sex

interaction [F(1,49) = 4.00, P < 0.05].

Figure 6. Parametric mesh construction. Digital points representing boundaries of the corpus callosum are redigitized to render each respective set of callosal points uniform. Analgorithm generates a parametric grid by creating a rectangular mesh over the uniform points from each callosal surface. Points from each two-dimensional surface mesh are pulledinto correspondence to represent an average callosal surface and retain intragroup variability profiles.

Figure 7. Corpus callosum partitioning protocol. Many partitioning protocols have been applied to generate comprehensive analyses of callosal area. The partitioning schemeemployed in this study divided midsagittal callosal area into anterior third, anterior midbody, posterior midbody, isthmus and splenium (adapted from Witelson et al., 1989; Clarke etal., 1994). Measures of maximal callosal width were obtained for each subdivision.

Figure 8. Brain volume and segmentation of the ventricles. To acquire brain volumes and volumes of the lateral and third ventricles, raw image data was processed in a number ofsteps. Different tissue types were classified according to signal intensity values and brain volumes were calculated. Volumes of the lateral and third ventricles were obtained by fillingin the regions of interest in native space according to neuroanatomic landmarks and CSF boundaries.

Figure 9. Color shape averages. Different shape profiles of the corpus callosum are seen in male Alzheimer’s patients (AD: green), male schizophrenic patients (SZ: blue) and maleadult controls (NC: red). The callosal renderings are taken from variability maps that represent the mean shape profiles within each group after AC–PC distance normalization instereotaxic space. The following composite was aimed at demonstrating that Alzheimer’s patients possess a mean shape profile of the corpus callosum (Thompson et al., 1998) thatis different from that of schizophrenic patients in spite of ventricular enlargement.

44 Callosal Morphology in Schizophrenia • Narr et al.

Cerebral Cortex Jan 2000, V 10 N 1 45

Diagnosis × Sex Interactions

Finally, of measures that revealed significant Diagnosis × Sex

interactions, post hoc tests of simple effects showed differences

in shape parameters between male patients and controls only. In

male groups there were significant differences in curvature for

bottom and top callosal surfaces [F(1,28) = 30.862 and 13.959

respectively, P < 0.001] with surface curvature substantially

increased in male patients. Significant differences in superior

points of top and bottom callosal surfaces [F(1,27) = 26.72, P <

0.001 and F(1,27) = 19.351, P < 0.001] also ref lected vertical

displacement of the callosum in three-dimensional space in male

patients.

Width

Measures of callosal thickness were obtained by taking the

maximum vertical width from each callosal subdivision used in

the area analysis (Fig. 7). The MANCOVA again confirmed the

significance of the AC–PC covariate [F(1,48) = 31.96, P < 0.001].

A significant multivariate Diagnosis × Sex interaction was found

[F(1,48) = 9.20, P < 0.004]. The multivariate effect of Diagnosis

differed for the dependent measures [F(4,192) = 3.47, P < 0.009].

Univariate ANOVAs revealed a significant difference across

Diagnosis for the width of the anterior callosal segment [F(1,48)

= 8.17, P < 0.006], with the width increased in schizophrenic

patients even after covarying for AC–PC distance. The Sex ×

Diagnosis interaction was significant for the splenial and anterior

callosal regions [F(1,48) = 4.58 and 6.19, P < 0.03 and P < 0.016

respectively]. A univariate F-test of simple effects in male

patients and controls showed significant group differences for

the anterior callosal segment, posterior midbody and splenium

[F(1,27) = 4.5, 8.78 and 14.44, P < 0.05, 0.006 and 0.001 respect-

ively]. In these regions male patients exhibited an increase in

callosal thickness compared to controls. There were no

significant differences in callosal widths between female groups.

Ventricular Volumes

In a 2 (Sex) × 2 (Diagnosis) MANCOVA using brain volume as

a covariate and raw lateral and third ventricle volume as

dependent variables there was a trend towards an effect of

Diagnosis [F(1,48) = 3.29, P < 0.08], but a significant Diagnosis ×

Sex interaction [F(1,48) = 4.07, P < 0.01] that also interacted

significantly with the dependent variables [F(1,48) = 5.91, P <

0.02]. There was significant effect of the covariate, but again this

differed for the dependent variables [F(1,48) = 4.06, P < 0.05].

Univariate analyses showed brain volume interacted with lateral

ventricle volume [F(1,48) = 4.09, P < 0.05], but not with third

ventricle volume. There remained, however, a significant Sex ×

Diagnosis interaction for third ventricle volume without the

covariate [F(1,48) = 4.86, P < 0.03]; covarying for brain volume,

this Sex × Diagnosis interaction was present for lateral ventricle

volumes [F(1,28) = 6.80, P < 0.01]. Tests of simple effects

revealed male patients had larger lateral ventricle volumes

[F(1,27) = 10.43, P < 0.001] and third ventricle volumes [F (1,27)

= 6.80, P < 0.01] than male controls.

Correlations

Correlations were performed between callosal shape parameters

(including curvature and vertical extents of the superior and

inferior callosal surfaces), and the volume of the lateral and third

ventricles. Significant two-tailed probability values were found

for correlations between lateral ventricle volumes and callosal

parameters, ref lecting displacement from the above analyses

(top and bottom curvature and superior limits, P < 0.01, r = 0.36,

0.52, 0.64, 0.70 respectively). Ventricular volume increases are

therefore related to increases in callosal surface curvature and

displacement. To ensure that significant correlations between

the shape measures and lateral ventricle volume were not due to

effects of AC–PC scaling or brain volume, partial correlations

between each measure, controlling separately for the effect of

AC–PC distance and for brain volume, were carried out. In all

cases the correlations remained significant upon removal of

variance associated with either AC–PC distance or brain volume.

There were no significant correlations between third ventricle

volumes and the callosal shape parameters.

ConclusionClear differences in patterns of variation and callosal shape were

observed in this morphometric analysis of the corpus callosum

between patients and controls (Figs 1–5). These differences

were clearly modulated by sex, suggesting different structural

patterns in male and female patients. Statistical tests confirmed

group differences in callosal shape parameters in three-

dimensional stereotaxic space, especially between male patients

and controls. Significant correlations ref lecting a relationship

between ventricular enlargement and callosal displacement

were evident.

Area and Width

Studies assessing standard callosal parameters in schizophrenia

have produced mixed results. MR studies report smaller (Rossi et

al., 1989; Woodruff et al., 1993), larger (Nasrallah et al., 1986;

Jacobsen et al., 1997), or no differences in callosal areas in

patients (Uematsu and Kaiya, 1988; Casanova et al., 1990; Raine

et al., 1990; Gunther et al., 1991; Woodruff et al., 1997). Studies

measuring callosal thickness have similarly reported increases

(Nasrallah et al., 1986), decreases (DeQuardo et al., 1996) and

no difference in callosal width across groups. Since the corpus

callosum connects functionally related regions, efforts have

been made to divide this structure into different areas to isolate

malformation of specific callosal channels (Thompson et al.,

1999a). Roughly equivalent estimates of the anterior corpus cal-

losum have been reported as thicker or larger in area (Nasrallah

et al., 1986) in patients, a finding compatible with ours where

maximum anterior callosal widths were greater in patients

compared to controls although were not different in area.

Gender Differences

Differences in callosal area and thickness have been found to

interact with sex. Hoff et al. (Hoff et al., 1994), for example,

reported that first-episode female patients had smaller callosal

area than female controls, partially replicating data (Hauser et

al., 1989) where chronic female schizophrenic patients were

shown to have smaller anterior callosal widths. In contrast,

Nasrallah (Nasrallah et al., 1986) found increased thickness of

anterior and middle callosal areas in females in some

concordance with Raine et al. (Raine et al., 1990) who found

anterior and posterior callosal widths smaller in male versus

female patients with the opposite effects in controls. A trend

towards a reversed sex difference in anterior and posterior

callosal size was also reported by Colombo et al. (Colombo et al.,

1993). The Diagnosis × Sex interaction in this study revealed

significant increases in widths of anterior and posterior callosal

regions in male patients, but no difference or reversal in females.

These effects appear unrelated to brain size given no effects of

diagnosis were evident and sex differences in brain volume

did not ref lect callosal width differences in female groups.

46 Callosal Morphology in Schizophrenia • Narr et al.

Moreover, our analyses show that AC–PC correction sufficiently

controls for relative differences in callosal sizes across sex.

Interestingly Highley et al. have observed an increase in callosal

fiber density in male patients versus controls and the opposite in

female groups in all regions but the posterior midbody and

splenium (Highley et al., 1999). It is possible that these changes

in fiber density could affect callosal thickness, especially for

small diameter fibers (Aboitiz et al., 1992). It is more likely,

however, that increased widths here ref lect bowing of the

callosum in male patients as anterior and posterior portions

appear more extended in three-dimensional space (Fig. 2).

Finally, many studies of callosal morphology have used low-

resolution scans that cannot represent true midline, different

partitioning protocols, and have sometimes neglected to use

head size corrections, all reasons making it difficult to compare

results across studies.

Shape

Overall, clear vertical displacements of the corpus callosum

were observed in the superior–inferior axis of three-dimensional

space in male schizophrenic patients, especially in midbody

callosal regions (Fig. 5). These results partially replicate one of

the few studies measuring callosal shape in addition to standard

morphometric parameters in schizophrenia. Specifically,

Casanova et al. used a statistical analysis of the coefficients of a

Fourier expansion series to demonstrate differences in callosal

shape in twins discordant for schizophrenia (Casanova et al.,

1990). A bowing of the corpus callosum was reported in the

superior–inferior dimension in affected twins compared to

unaffected co-twins, thought to ref lect enlargement of the

midbody of the lateral ventricle superior horns and/or third

ventricle enlargement. Furthermore using a two-dimensional

skeletonization technique, Frumin et al. report shape differ-

ences in the posterior corpus callosum (increased curvature) in

chronic schizophrenic patients that was not apparent in bipolar

patients or normal controls (Frumin et al., 1998). In our sample,

we found significant increases in lateral and third ventricular

volumes in male patients with increases in lateral ventricular

volume significantly relating to increases in callosal displace-

ment and curvature. It appears that ventricular enlargement

results in vertical displacements of the superior horn in male

patients that have been shown in stereotaxic space (Moussai et

al., 1998). In addition, a study using the same shape analyses in

pediatric temporal lobe epilepsy, autism, attention deficit

hyperactivity disorder and juvenile onset schizophrenia detected

similar displacements and bowing of the corpus callosum in the

schizophrenia group only (R. Blanton et al., personal com-

munication) even though ventricular enlargement was not

specific. To further establish whether bowing of the corpus

callosum is specific to male schizophrenic patients, we

compared callosal surface renderings from our male patients and

controls with callosal averages from Alzheimer’s patients in the

same stereotaxic space. The male Alzheimer’s patients exhibited

ventricular enlargement as well as a 24.5% decrease in posterior

midbody callosal regions (Thompson et al., 1998) (Fig. 9). While

the corpus callosum is bowed in both Alzheimer’s and schizo-

phrenic patients, distinct patterns of shape are found in each

group, indicating that the pattern of callosal displacement in

schizophrenia may indeed be disease specific.

Gender

There has been an ongoing controversy concerning sex differ-

ences in callosal morphology with an emphasis on splenial area

in normal populations (De Lacoste-Utamsing et al., 1982;

Holloway and De Lacoste, 1996; Bishop and Wahlsten, 1997;

Davatzikos and Resnick, 1998). Furthermore, callosal size has

been shown to correlate with small brain size, but not with

large brain size (Rauch and Jinkins, 1994; Jäncke et al., 1997).

In either case, gender and/or head size clearly inf luence

normal callosal morphometry, making gender interactions with

morphometric abnormalities in schizophrenic populations of

considerable interest.

There are a number of reasons why differences in callosal

shape may be more apparent in male populations. There appear

to be distinct gender differences in schizophrenic populations.

For example, negative symptoms are more prominent in male

patients and males have an earlier age of onset, a poorer quality

of life and a worse course of illness (DeLisi et al., 1989;

Waddington, 1993; Gur et al., 1996). Furthermore, not only do

clinical features in schizophrenia appear sexually dimorphic but

there appear to be relationships between neuroanatomical

volumes and specific clinical dimensions across the sexes

(Cowell et al., 1996). The curvature of the corpus callosum has

also been noted to increase with age in male populations at a

higher rate than in females (Rajapakse et al., 1996).

Clinical and psychopathological heterogeneity in schizo-

phrenic patients may also account for inconsistencies in results

when assessing structural morphology, including morphology of

the callosum (Colombo et al., 1993). For example, patients with

negative symptoms show smaller callosal sizes (Gunther et al.,

1991; Woodruff et al., 1993), thicker callosa (Coger and

Serafetinides, 1990) or no difference relative to other groups.

Similarly, early-onset schizophrenic patients have been shown to

have larger total, anterior and posterior callosal areas compared

to controls (Bigelow et al., 1983; Coger and Serafetinides, 1990;

Jacobsen et al., 1997). Clearly more information is needed to

establish the relationship of these factors to callosal morph-

ometry in schizophrenia.

Variability Maps

Results mentioned above and those from other studies have

indicated an enormous range of morphometric findings in the

corpus callosum in schizophrenic populations. Different types

of head size correction or failure to include these corrections

may contribute to discrepancies in results (Jäncke et al., 1997).

Failure to replicate findings across studies may also result from

patient heterogeneity and differences in callosal structure

between the sexes across diagnosis. In our study, variability and

displacement maps (Figs 1–5) show clear differences between

populations in patterns of callosal variation. Overall, the vari-

ability and displacement maps indicate a marked upward shift of

the corpus callosum in three-dimensional space in schizophrenic

patients that bears a relationship to lateral but not third ventricle

enlargement. Many studies have reported no inf luence of age in

their findings of callosal pathology (Thompson et al., 1998), but

few have looked at shape across age. Even though the effects of

aging and callosal development are still under investigation,

these results and those from other studies have controlled for age

to some degree. Finally, in our sample, groups were matched for

handedness. Handedness appears to be a predictor of neuro-

anatomical asymmetry and thus bears a relationship to callosal

morphology (Clarke and Zaidel, 1994).

In sum, the methods employed in this study revealed unique

differences in callosal shape and patterns of variability between

schizophrenic patients and normal controls with clear gender

differences. Further investigations are in progress, relating

Cerebral Cortex Jan 2000, V 10 N 1 47

callosal parameters to other neuroanatomical regions shown to

possess structural alterations in schizophrenic patients, such

as asymmetric perisylvian cortices and prefrontal cortices.

Furthermore, structural alterations in surrounding regions such

as thalamus and cingulate cortices may also be related to callosal

displacements. Finally, it is clear that gender inf luences callosal

morphology in schizophrenia. Larger sample sizes and homo-

geneous patient populations matched closely with control

subjects as well as correlations with symptom complexes are

required. This is necessary as it appears callosal morphology in

schizophrenic patients is tempered by a number of clinical

variables, including symptomatology, disease course and age of

onset in addition to sex, handedness and age.

NotesWe especially thank Tonmoy Sharma and his colleagues at the Institute of

Psychiatry for having the foresight to collect this data and we are grateful

to Janice M. Rayman for contributing her statistical expertise.This work

was supported by a Human Brain Project grant to the International

Consortium for Brain Mapping, funded jointly by NIMH and NIDA (P20

MH/DA52176), by a P41 Resource Grant from the NCRR (13642), by

research grants from the National Library of Medicine (LM/MH05639),

the National Science Foundation (BIR 93–22434), the NCRR (RR05056)

and NINCDS (NS38253), and research grants from the Howard Hughes

Medical Institute, United States Information Agency, and the US–UK

Fulbright Commission, London (to P.T.).

Address correspondence to Dr Arthur W. Toga, Laboratory of Neuro

Imaging, Department of Neurology, Division of Brain Mapping, UCLA

School of Medicine, 710 Westwood Plaza, Los Angeles, CA 90095–1769,

USA. Email: toga@loni.ucla.edu.

ReferencesAboitiz F, Scheibel AB, Fisher RS, Zaidel E (1992) Fiber composition of the

human corpus callosum. Brain Res 598:143–153.

American Psychiatric Association (1987) Diagnostic and statistical manual

for mental disorders (3rd edn, revised). Washington, DC: American

Psychiatric Association Press.

Annett M (1970) A classification of hand preference by association. Br J

Psychol 61:303–321.

Barta PE, Pearlson GD, Brill LB, Royall R (1997) Planum temporale

asymmetry reversal in schizophrenia: replication and relationship to

gray matter abnormalities. Am J Psychiat 154:661–667.

Bigelow LB, Nasrallah HA, Rausher FP (1983) Corpus callosum thickness

in chronic schizophrenia. Arch Gen Psychiat 142:284–287.

Bishop KM, Wahlsten, D (1997) Sex differences in the human corpus

callosum: myth or reality? Neurosci Biobehav Rev 21:581–601.

Casanova MF, Sanders RD, Goldberg TE, Bigelow LB, Christison G, Torrey

EF, Weinberger DR (1990) Morphometry of the corpus callosum in

monozygotic twins discordant for schizophrenia: a magnetic

resonance imaging study. J Neurol Neurosurg Psychiat 53:416–421.

Clarke JM, Zaidel E (1994) Anatomical–behavioral relationships: corpus

callosum morphometry and hemispheric specialization. Behav Brain

Res 64:185–202.

Colombo C, Bonfanti, Livian S, Abbruzzese M, Scarone S (1993) Size of

the corpus callosum and auditory comprehension in schizophrenics

and normal controls. Schizophr Res 11:63–70.

Coger RW, Serafetinides EA (1990) Schizophrenia, corpus callosum, and

interhemispheric transfer: a review. Psychiat Res 34:163–184.

Cowell PE, Kostianovsky DJ, Gur RC, Turetsky BI, Gur RE (1996) Sex

differences in neuroanatomical and clinical correlations in schizo-

phrenia. Am J Psychiat 153:799–805.

David AS (1994) Schizophrenia and the corpus callosum: developmental,

structural and functional relationships. Behav Brain Res 64:203–211.

Davatzikos C, Resnick SM (1998) Sex differences in anatomic measures of

interhemispheric connectivity: Correlations with cognition in women

but not men. Cereb Cortex 8:635–640.

De Lacoste-Utamsing MC, Holloway RL (1982) Sexual dimorphism in the

human corpus callosum. Science 216:1431–1432.

DeLisi LE, Dauphinais ID, Hauser P (1989) Gender differences in the

brain: are they relevant to the pathogenesis of schizophrenia. Comp

Psychiat 30:197–208.

DeQuardo JR, Bookstein FL, Green WD, Brunberg JA, Tandon R (1996)

Spatial relationships of neuroanatomic landmarks in schizophrenia.

Psychiat Res 61:81–95.

Frumin M, Golland P, McCarley RW, Hirayasu Y, Salisbury DF, Kikinis R,

Shenton M (1998) Shape differences in the corpus callosum in first

episode schizophrenia and affective disorder. ACNP meeting

(abstract).

Gunther W, Petsch R, Steinberg R, Moser E, Streck P, Heller H, Kurtz G,

Hippius H (1991) Brain dysfunction during motor activation and

corpus callosum alterations in schizophrenia measured by cerebral

blood f low and magnetic resonance imaging. Biol Psychiat

29:535–555.

Gur RE, Petty RG, Turetsky BI, Gur RC (1996) Schizophrenia throughout

life: sex differences in severity and profile of symptoms. Schizophr

Res 21: 1–12.

Hauser P, Dauphinais ID, Berrettini W, DeLisi LE, Gelernter J, Post RM

(1989) Corpus callosum dimensions measured by magnetic resonance

imaging in bipolar affective disorder and schizophrenia. Biol Psychiat

26:659–668.

Highley JR, Esiri MM, McDonald B, Cortina-Borja M, Herron BM, Crow TJ

(1999) The size and fibre composition of the corpus callosum with

respect to gender and schizophrenia: a post-mortem study. Brain

122:99–110.

Hoek HW, Brown AS, Susser E (1998) The Dutch famine and schizo-

phrenia spectrum disorders. Soc Psychiat Psychiat Epidemiol 33:

373–379.

Hoff AL, Neal C, Kushner M, DeLisi LE (1994) Gender differences in

corpus callosum size in first-episode schizophrenics. Biol Psychiat

35:913–919.

Holloway RL, De Lacoste MC (1996) Sexual dimorphism in the corpus

callosum: an extension and replication study. Hum Neurobiol

5:87–91.

Jacobsen LK, Giedd JN, Rajapakse JC, Hamberger SD, Vaituzis AC, Frazier

JA, Lenane MC, Rapoport JL (1997) Quantitative magnetic resonance

imaging of the corpus callosum in childhood onset schizophrenia.

Psychiat Res 68:77–86.

Jäncke L, Staiger JF, Schlaug G, Huang Y, Steinmetz H (1997) The

relationship between corpus callosum size and forebrain volume.

Cereb Cortex 7:48–56.

Kikinis R, Shenton ME, Gerig G, Hokama H, Haimson J, O’Donnell BF,

Wible CG, McCarley RW, Jolesz FA (1994) Temporal lobe sulco-gyral

pattern anomalies in schizophrenia: an in vivo MR three-dimensional

surface rendering study. Neurosci Lett 182:7–12.

Lawrie SM, Abukmeil SS (1998) Brain abnormality in schizophrenia. Br J

Psychiat 172:110–120.

Office of Population Censuses and Surveys (1991) Classification of

Occupations. Standard Occupational Classification (SOS). London:

HMSO.

Mazziotta JC, Toga AW, Evans AC, Fox P, Lancaster J (1995) A probabilistic

atlas of the human brain: theory and rationale for its development.

NeuroImage 2:89–101.

Moussai J, Sharma T, Anvar BA, Narr KL, Cannestra AF, Thompson PM,

Toga AW (1998) 3-Dimensional analysis of lateral ventricles in

schizophrenia. 4th International Conference on Functional Mapping

of the Human Brain. NeuroImage 7:S505.

Nasrallah HA, Andreasen NC, Coffman JA, Olson SC, Dunn VD, Ehrhardt

JC, Champman SM (1986) A controlled magnetic resonance imaging

study of corpus callosum thickness in schizophrenia. Biol Psychiat

21:3:274–282.

Njiokiktjien C, de Donneville L, Vaal J (1994) Callosal size in children with

learning disabilities. Behav Brain Res 64:213–218.

Raine A, Harrison GN, Reynolds GP, Sheard C, Cooper JE, Meddley I

(1990) Structural and functional characteristics of the corpus

callosum in schizophrenics, psychiatric controls, and in normals.

Arch Gen Psychiat 471060–1064.

Rajapakse JC, Geidd JN, Rumsey JM, Vaituzis AC, Hamburger SD,

Rapoport JL (1996) Regional MRI measurements of the corpus

callosum: a methodological and developmental study. Brain Devel

18:379–388

Rauch R A, Jinkins JR (1994) Analysis of cross-sectional area

measurements of the corpus callosum adjusted for brain size in male

and female subjects from childhood to adulthood. Behav Brain Res

64:65–78.

48 Callosal Morphology in Schizophrenia • Narr et al.

Rossi A, Stratta P, Gallucci M, Passariello R, Casacchia M (1989)

Quantification of corpus callosum and ventricles in schizophrenia

with nuclear magnetic resonance imaging: a pilot study. Am J Psychiat

146:99–101.

Sharma T, du Boulay G, Lewis S, Sigmundsson T, Gurling H, Murray R

(1997). The Maudsley Family Study. I: Structural brain changes on

magnetic resonance imaging in familial schizophrenia. Progr

Neuro-Psychopharmacol Biol Psychiat 21:1297–1315.

Sharma T, Lancaster E, Lee D, Lewis S, Sigmundsson T, Takei N, Gurling

H, Barta P, Pearlson G, Murray R (1998). Brain changes in

schizophrenia volumetric MRI study of families multiply affected

with schizophrenia — the Maudsley Family study 5. Br J Psychiat

173:132–138.

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.

Talairach J, Tournoux P (1988) Co-planar stereotaxic atlas of the human

brain. New York: Thieme.

Thompson P, Toga AW (1996a) A surface-based technique for warping

three-dimensional images of the brain. IEEE Trans Med Imag 15:

402–417.

Thompson PM, Schwartz C, Toga AW (1996b) High-resolution random

mesh algorithms for creating a probabilistic 3D surface atlas of the

human brain. NeuroImage 3:19–34.

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 Comput Assist Tomogr

21:567–581.

Thompson PM, Moussai J, Zohoori S, Goldkorn A, Khan AA, Mega MS,

Small GW, Toga AW (1998) Cortical variability and asymmetry in

normal aging and Alzheimer’s disease. Cereb Cortex 8:492–509.

Thompson PM, Narr KL, Blanton RE, Toga AW (1999a) In: Mapping

structural alterations of the corpus callosum during brain develop-

ment and disease (Iacoboni M, Zaidel E, eds). Cambridge, MA: MIT

Press (in press) [review].

Thompson PM, Woods RP, Mega MS, Toga AW (1999) Mathematical/

computational challenges in creating deformable and probabilistic

brain atlases. Hum Brain Map 9 (in press).

Uematsu M, Kaiya H (1988) The morphology of the corpus callosum in

schizophrenia: a MRI study. Schizophr Res 1:391–398.

Waddington JL (1993) Neurodynamics of abnormalities in cerebral

metabolism and structure in schizophrenia. Schizophr Bull 19:55–58.

Witelson SF (1989) Hand and sex differences in the isthmus and genu of

the human corpus callosum: a postmortem morphological study.

Brain 112:799–835.

Woodruff PWR, Pearlson GD, Geer MJ, Barta PE, Chilcoat HD (1993) A

computerized imaging study of corpus callosum morphology in

schizophrenia. Psychol Med 23:45–56.

Woodruff PWR, McManus IC, David AS (1995) Meta-analysis of corpus

callosum size in schizophrenia. J Neurol Neurosurg Psychiat

58:457–461.

Woodruff PWR, Phillips ML, Rushe T, Wright RC, Murray RM, David AS

(1997) Corpus callosum size and inter-hemispheric function in

schizophrenia. Schizophr Res 23:189–196.

Woodruff PWR, Wright IC, Shurique N, Russouw H, Rushe T, Howard RJ,

Graves M, Bullmore ET, Murray RM (1997b) Structural brain abnor-

malities in male schizophrenics ref lect fronto-temporal dissociation.

Psychol Med 27:1257–1266.

Zaidel DW, Esiri MM, Harrison PJ (1997) Size, shape, and orientation of

neurons in the left and right hippocampus: investigation of normal

asymmetries and alterations in schizophrenia. Am J Psychiat

154:812–818.

Zipursky RB, Lim KO, Sullivan EV, Brown BW, Pfefferbaum A (1992)

Widespread cerebral gray matter volume deficits in schizophrenia.

Arch Gen Psychiat 49:195–205.

Cerebral Cortex Jan 2000, V 10 N 1 49