Elsevier

Biological Psychiatry

Volume 50, Issue 7, 1 October 2001, Pages 531-539
Biological Psychiatry

Selective bilateral hippocampal volume loss in chronic schizophrenia

https://doi.org/10.1016/S0006-3223(01)01121-0Get rights and content

Abstract

Background: The hippocampus is implicated in the pathophysiology of schizophrenia; however, volumetric changes are subtle and have limited diagnostic specificity. It is possible that the shape of the hippocampus may be more characteristic of schizophrenia.

Methods: Forty-five patients with chronic schizophrenia and 139 healthy control subjects were scanned using magnetic resonance imaging. Hippocampi were traced manually, and two-dimensional shape information was analyzed.

Results: Two shape factors were found to be adequate to represent variance in the shape of the hippocampus. One of these factors, representing volume loss behind the head of the hippocampus, provided a degree of discrimination between patients with chronic schizophrenia and healthy control subjects; however, overall hippocampal volume following appropriate adjustment for brain volume showed a similar level of discrimination. Patients with chronic schizophrenia were best characterized using these two measures together, but diagnostic specificity was only moderate.

Conclusions: This study identified that less of the hippocampus was distributed in its posterior two-thirds in patients with chronic schizophrenia, and specifically in the region just posterior to the hippocampal head. Group discrimination on the basis of hippocampal volume and shape measures was moderately good. A full three-dimensional analysis of hippocampal shape, based on large samples, would be a useful extension of the study.

Introduction

The hippocampus has been implicated in the pathophysiology of schizophrenia on clinical (Sachdev 1998), neuropsychological (Saykin et al 1991) and neuroanatomical (Harrison 1999) grounds. High-resolution magnetic resonance imaging (MR) studies have generally supported the notion of hippocampal volume loss in schizophrenia Nelson et al 1998, Velakoulis et al 1999, but the exact nature of the hippocampal abnormality remains uncertain. Neuropathological data suggest reduced neuropil and neuronal size, rather than neuronal loss (Benes et al 1991), although there are studies that fail to find any hippocampal abnormality in postmortem tissue (for review, see Harrison 1999).

It has been suggested that if hippocampal abnormalities predominantly affect neural architecture and connectivity, then analysis of hippocampal shape may be more sensitive than volumetric assessment when examining MR images (Csernansky et al 1998). In some of the early MR studies of hippocampal volume in schizophrenia, reduction was found to be limited to specific areas, with some studies suggesting that anterior regions were more reduced DeLisi et al 1988, Rossi et al 1994, Suddath et al 1990, whereas others demonstrated posterior hippocampal changes Becker et al 1996, Bogerts et al 1990, Bogerts et al 1993. Recently, Maier and colleagues (2000) reported no significant volume loss or shape difference between patients and control subjects.

Methodological problems limit the interpretation of such studies and may explain their discrepancies. Suddath et al (1990), for example, only measured the anterior hippocampus on three contiguous slices, and only compared their subjects on one of those slices. Other studies have measured the anterior hippocampus and amygdala as a single structure Bogerts et al 1990, Maier et al 2000, Rossi et al 1994, Shenton et al 1992, Whitworth et al 1998 or used thick image slices of between 3.1 and 5 mm Becker et al 1996, Bogerts et al 1990, Bogerts et al 1993, DeLisi et al 1988, Rossi et al 1994, which do not allow differentiation of the anterior hippocampus from the amygdala. Higher-resolution MR images produce larger numbers of thinner slices and allow the delineation of hippocampus and amygdala. Using such images, Fukuzako et al 1995, Fukuzako et al 1996 identified reduced hippocampal length, defined as the number of traced slices, in chronic and first-episode schizophrenia, but did not identify the site of their reported hippocampal shortening.

Only one published study has definitively studied hippocampal shape in schizophrenia (Csernansky et al 1998). Fifteen patients and 15 matched control subjects were compared using high-dimensional transformations of a computerized brain template. This indicated abnormalities of hippocampal anatomy in lateral regions of the head of the hippocampus bilaterally. These are areas that are thought to have prominent projections to the prefrontal cortex Barbas and Blatt 1995, Carmichael and Price 1995, a finding that is in keeping with the hypothesis of a disrupted fronto-temporal network in schizophrenia (Weinberger et al 1994). In addition, it supports previous work by Bilder and colleagues (1995), who have found that reduced volume of the anterior hippocampus is associated with neuropsychological deficits on tests believed to be sensitive to frontal lobe function.

One limitation of the study of Csernansky et al (1998) is that only 15 patients and 15 control subjects were studied. Given the large degree of normal nonpathological variation in hippocampal volume, a large control sample would be beneficial. We assessed the hippocampal shape of 45 patients with chronic schizophrenia, and compared them to a group of 139 healthy volunteers. We used a related approach to that of Csernansky et al (1998) to examine the profile or “shape” of the hippocampus as defined by the volumes of contiguous slices. Conventional statistical techniques, such as analysis of variance, have been used to analyze comparable data from other structures, but such tests assume statistical independence of the volumes of adjacent slices. To use all available data, and to avoid the problem of correlation between measures, principal components analysis of slice data were carried out before making shape comparisons between groups. Based on the work cited above, we predicted that patients with schizophrenia would have less of their hippocampal volume distributed anteriorly, when compared to control subjects.

Section snippets

Subjects

Chronic schizophrenia patients were recruited from the Rehabilitation Unit at Royal Park Psychiatric Hospital, Melbourne, Australia. Diagnoses were based on a combination of structured clinical interview and medical record review to provide DSM-III-R (67% of patients) or DSM-IV criteria. All patients were medicated at time of scanning and had at least 2 years of neuroleptic exposure. Healthy volunteers were recruited by approaching ancillary hospital staff and through advertisements. These

Results

The average volume, length, and shape of the left and right hippocampi in control subjects and patients with schizophrenia are shown in Figure 1. The construction of these average hippocampal shapes is explained below, but at this point it is useful to describe the main features of the diagram. First, it can be seen that there is little difference in the length of the hippocampi, either comparing between groups or between left and right hippocampi. A volume difference between groups is

Discussion

This study has shown that, as previously reported, hippocampal volumes were bilaterally smaller in the patients with chronic schizophrenia compared to the control group, but that there were no differences in hippocampal lengths between groups (Velakoulis et al 1999). Using the distribution of hippocampal volume along its long axis, we derived a mathematical model of the hippocampal shape, which enabled statistical comparisons between the two groups. In contrast to previous reports of anterior

Acknowledgements

Dr. Stuart was supported by the National Alliance for Research in Schizophrenia and Depression (U.S.). The work was supported by grants to Assoc. Prof. Pantelis from the National Health and Medical Research Council (Australia), NHMRC Schizophrenia Research Unit, the Australian Communications and Computing Institute, the Jack Brockhoff Foundation, the Ian Potter Foundation, the L.E.W. Carty Trust, the Meredith Woods Trust, and the Percy Baxter Charitable Trust, Melbourne, Australia.

References (38)

  • F.M Benes et al.

    Reduced neuronal size in posterior hippocampus of schizophrenic patients

    Schizophr Bull

    (1991)
  • R.S Briellman et al.

    TLE patients with postictal psychosis microdysplasia and anterior hippocampal preservation

    Neurology

    (2000)
  • S Carmichael et al.

    Limbic connections of the orbital and medial prefrontal cortex in macaque monkeys

    J Comp Neurol

    (1995)
  • R Cattel

    The scree test for the number of factors

    Multivariate Behav Res

    (1966)
  • J.G Csernansky et al.

    Hippocampal morphometry in schizophrenia by high dimensional brain mapping

    Proc Natl Acad Sci USA

    (1998)
  • L.E DeLisi et al.

    Perinatal complications and reduced size of brain limbic structures in familial schizophrenia

    Schizophr Bull

    (1988)
  • H Fukuzako et al.

    Shortening of the hippocampal formation in first-episode schizophrenic patients

    Psychiatry Clin Neurosci

    (1995)
  • R Gorsuch

    Factor Analysis

    (1983)
  • H Harman

    Modern Factor Analysis

    (1967)
  • Cited by (0)

    View full text