A cross-study meta-analysis and three-dimensional comparison of cell counting in the anterior cingulate cortex of schizophrenic and bipolar brain

Using a two-dimensional cell counting approach, a 1991 study in the anterior cingulate cortex (ACCx) detected a reduction in the density of nonpyramidal neurons in layers II–VI of schizophrenic subjects. Schizophrenics without superimposed mood disturbances showed a 16% decrease in layer II, while schizoaffectives showed a 30% decrease, suggesting that a decreased density of nonpyramidal neurons in layer II of ACCx might vary more strongly with affective disorder than with schizophrenia. Two follow-up studies from this laboratory, one a replication of that reported in 1991 and the other an analysis of tyrosine hydroxylase immunoreactive fibers, were undertaken in ACCx of normal controls and schizophrenics. These three data sets have been combined and a meta-analysis of the density of pyramidal, nonpyramidal and glial cells was performed to explore whether changes in the density of interneurons in ACCx may be a reliable finding in the major psychoses. Not all groups have reported this finding, but several had employed a different cell counting technique (i.e. three dimensional optical dissector), which could help to explain the discrepant findings in schizophrenia and affective disorder. The data from each of three different studies (now designated as studies A, B and C, respectively) have been internally normalized, combined into a single dataset and analyzed using nonparametric statistics. Tissue blocks from a subset of cases in study B (six controls, six schizophrenics and six bipolars) were embedded in celloidin and counted using an “unbiased” three dimensional counting method (study D). The data from studies A and B indicate that the density of nonpyramidal neurons in layer II of ACCx in the schizoaffective and bipolar samples was significantly decreased. In the schizophrenics, the nonpyramidal neurons were also decreased, but only by 15%. All three groups also showed a decrease of pyramidal neurons in layers IV, V and VI, but this difference was significant only in layer IV of the schizophrenics. When data from study C were added, the differences in pyramidal and nonpyramidal neurons were less striking. For study D, the pattern of findings are strikingly similar to those obtained in studies A, B and C, indicating that both 2D and 3D cell counting methodologies are capable of detecting the same differences. Taken together, these results indicate that the earlier finding of a decreased density of nonpyramidal neurons in ACCx of schizophrenics is consist