To the Editor: Mary A. Walker, et al. (1) concluded that their stereological study of hippocampal volume and neuron number in schizophrenia provided evidence against a primary pathology of hippocampal structure and against the notion of schizophrenia as a limbic system disorder (2). While the stereological techniques employed allowed Ms. Walker et al. to draw strong inferences about hippocampal volume and cell number in schizophrenia, it is important to add some cautionary notes to their conclusions.

First, it is possible that subtle structural changes of the hippocampus involve primarily the anterior but not the posterior division (3). Ms. Walker et al. did not test for such a regionally selective volume difference. Furthermore, there is intriguing new evidence that only some hippocampal neurons are affected in schizophrenia (4, 5). Pathology in a subpopulation of hippocampal neurons could remain undetected with the study design of Ms. Walker et al.

Second, the fact that the photographic images of brain slices, but not the histological preparations, revealed smaller hippocampal volume in schizophrenia is intriguing. The histologically defined volume estimates included only the cell-containing regions, whereas the volume estimates derived from photographic images included adjacent tissue as well. Ms. Walker et al. interpreted this as evidence for changes of the parahippocampal gyrus to strengthen their argument that isocortical but not allocortical structures of the medial temporal lobe are abnormal in schizophrenia. However, an alternative explanation seems more likely: the tissue adjacent to the cell-containing regions that is reduced in volume in schizophrenia is the hippocampal white matter, which cannot be distinguished easily from hippocampal gray matter on unstained brain slices. A decrease of white but not gray matter volume and normal total number of all subpopulations of hippocampal neurons is exactly the finding of the only other stereological study of the hippocampus in schizophrenia (6).

It seems that the study by Ms. Walker et al. provides further evidence for a subtle abnormality of the hippocampus in schizophrenia, possibly leading to a disconnection of the hippocampus from isocortical modules in the frontal and temporal lobes.