Mapping Regional Hippocampal Abnormalities in First Episode Schizophrenia

Katherine L. Narr1, Robert M. Bilder2, Philip Szeszko2, Paul M. Thompson1, Seonah Jang1, Sharon Kim1, Kiralee M. Hayashi1, Dina Asunction1, Roger P. Woods1, Arthur W. Toga1
1Laboratory of Neuro Imaging, Brain Mapping Division, Department of Neurology, UCLA School of Medicine
2Department of Psychiatry and Biobehavioral Sciences, UCLA School of Medicine, Los Angeles, CA

Mesolimbic abnormalities are robustly observed in schizophrenia. In vivo imaging studies report hippocampal volume reductions in chronic and first episode patients and in their biological relatives. While earlier studies suggest that hippocampal volume reductions are a trait marker of schizophrenia and genetic vulnerability, similar abnormalities have been observed in other neuropsychiatric populations. Several studies have suggested that anterior hippocampal regions may be particularly susceptible to volume reduction and that these reductions may correlate with frontal lobe dysfunction. Novel brain mapping methods applied to high-resolution imaging data may provide greater diagnostic specificity by identifying hippocampal subregions that are most vulnerable to disease processes in schizophrenia. The goals of this study were thus (1) to confirm the presence of hippocampal volume reductions near the onset of schizophrenia and (2) to apply novel brain mapping methods to isolate regional changes in hippocampal morphometry in schizophrenia relative to healthy individuals.

High-resolution (256x256x124; 1.5 mm slice separation) T1-weighted MR images were obtained from 65 patients in their first episode of schizophrenia (mean age = 24.3+/-8.2 SD, 47m/18f) and 77 normal control subjects (mean age = 29.4+/-4.6, 33m/44f). Image volumes were corrected for magnetic field inhomogeneities and resliced into a standard orientation using a six-parameter rigid-body linear transformation with no scaling. Investigators blind to group status traced the hippocampi on coronal brain slices while following anatomical landmarks in orthogonal viewing planes. Hippocampal volumes and linear measures were retained for use as dependent measures in statistical analyses and total brain volumes were measured for use as covariates. Anatomical mesh modeling methods, which match equivalent hippocampal surface points between subjects in 3D, were used to obtain hippocampal surface averages within diagnostics groups. Distances between surface points and the central core of each hippocampal surface model were calculated to identify regional hippocampal volume decreases in schizophrenia. These distances, measured at homologous hippocampal surface points in each individual, were compared statistically between diagnostic groups and mapped back into 3D.

First episode patients exhibited significant bilateral hippocampal volume decreases compared to controls (right: p<.001; left: p<.02), and right hemisphere reductions in hippocampal heights (p<.001) and widths (p<.02) after correction for individual differences in brain size. Average maps of hippocampal surface anatomy also showed smaller hippocampal models in patients with schizophrenia compared to controls. Local hippocampal volume losses were more evident in the right hemisphere in CA1 and CA2 regions in anterior and midbody regions.

Results confirm that hippocampal volume loss is a robust neuroanatomical correlate of schizophrenia and is present at disease onset. Although global volume reductions appear characteristic, mid- to anterolateral hippocampal regions show more pronounced volume changes in patients with first episode schizophrenia compared to normal.