Ventricular enlargement in schizophrenia: a primary change in the temporal lobe?

Abstract

Background: The anatomical origin of the enlargement of the cerebral ventricles in schizophrenia is obscure. Methods: In this study, the volumes of the hemispheres and lateral ventricles were assessed in MRI scans of 43 formalin-fixed brains (23 from patients and 19 comparison subjects) using a spline ‘snake’ segmentation method. Results: A bilateral ventricular volume increase was found in schizophrenia. Whereas enlargement of the lateral ventricle (mean: 54%) as a whole was related to age of onset and was greater in females than in males, enlargement of the temporal horn (mean: 54%) was not strongly related to age of onset or sex. Lateral ventricle volume was negatively correlated with STG, fusiform and parahippocampal volume in schizophrenia. Hemispheric volumes were unchanged. Conclusions: The differing correlates of the components of ventricular enlargement suggest a degree of selectivity of the disease process with a focus in the temporal lobe.

Introduction

Of 43 MRI studies in which lateral ventricle size was measured in schizophrenia, 77% found enlargement as did 75% of the earlier CT studies (McCarley et al., 1999). A mean increase of 26% has been estimated in MRI studies (Wright et al., 2000). Yet, the meaning for the disease process remains obscure. Lack of bimodality in the distribution of ventricular size within the patient group (Vita et al., 2000) suggests a single disease entity. Absent periventricular gliosis Roberts et al., 1987, Bruton et al., 1990 and the presence of ventricular enlargement at the time of the first episode (DeLisi et al., 1991) are consistent with a neurodevelopmental Weinberger, 1987, Crow et al., 1989 rather than a degenerative origin.

Ventricular enlargement presumably reflects a change in brain structure. There are indications that brain structural change is localised—temporal lobe volume (Dauphinais et al., 1990) and length Crow and Crow, 1996, Highley et al., 1998a have been found reduced relative to the brain as a whole. It has been suggested that temporal horn enlargement in schizophrenia is due to a volume reduction of adjacent limbic structures (Brown et al., 1986), and several MRI studies have found enlarged temporal horns and smaller temporal limbic structures Bogerts et al., 1990, Kawasaki et al., 1993, Shenton et al., 1992. However, only Shenton et al. found a negative correlation between left temporal horn and left parahippocampal gyrus volumes, whereas the other two studies found no correlation and interpreted this as evidence against a simple interaction between limbic tissue loss and temporal horn enlargement. No correlation was also found between amygdala volume and temporal horn volume in brains from the series reported in the present study (Chance et al., 2002). This stands in contrast to the negative correlations between amygdala volume and temporal horn volume reported for epilepsy (a neurodegenerative rather than neurodevelopmental process).

There is a tendency for enlargement of the ventricle to be greater on the left side (e.g. Bruton et al., 1990, DeLisi et al., 1991, Narr et al., 2001) in addition to the tendency for more marked enlargement of the temporal horns. These trends converge in studies that have found the temporal horn in the left hemisphere to be particularly enlarged Bogerts et al., 1985, Crow et al., 1989.

That the brain changes in schizophrenia have a genetic component is suggested by similar findings, e.g. a reduction in cortical grey matter (Cannon et al., 1998) and enlarged ventricles (Lawrie et al., 1999) in the siblings of patients. However, the ventricles are also found to be larger in the ill members of pairs of twins discordant for suffering from schizophrenia Reveley et al., 1982, Suddath et al., 1990. Ventricular enlargement, therefore, and the expression of illness are dependent on both genetic and nongenetic factors (Silverman et al., 1998; Cannon et al., 1989)—perhaps better interpreted as epigenetic factors. Gender is one modulator of phenotype. Sex differences are present in both the expression of illness (i.e. in age of onset of psychosis) and the extent of abnormality of brain structure (Nopoulos et al., 1997) as well as in cerebral asymmetry (Bear et al., 1986).

In this study, MRI scans from a series of postmortem brains were used to assess gross volume of the cerebral hemispheres, lateral ventricles and temporal horns to examine the selectivity of the change in ventricular structure to the temporal horns. An MRI method was chosen because this noninvasive technique allows ventricle volume to be assessed without compromising the structural integrity of the walls of the ventricles that may be distorted by dissection or the introduction of a radio-opaque medium for visualisation.