Quantitative grey matter histological measures do not correlate with grey matter probability values from in vivo MRI in the temporal lobe

Voxel-based morphometry (VBM) is commonly used to study systematic differences in brain morphology from patients with various disorders, usually by comparisons with control subjects. It has often been suggested, however, that VBM is also sensitive to variations in composition in grey matter. The nature of the grey matter changes identified with VBM is still poorly understood. The aim of the current study was to determine whether grey matter histopathological measurements of neuronal tissue or gliosis influenced grey matter probability values that are used for VBM analyses. Grey matter probability values (obtained using both SPM5 and FSL-FAST) were correlated with neuronal density, and field fraction of NeuN and GFAP immunopositivity in a grey matter region of interest in the middle temporal gyrus, in 19 patients undergoing temporal lobe resection for refractory epilepsy. There were no significant correlations between any quantitative neuropathological measure and grey matter probability values in normal-appearing grey matter using either segmentation program. The lack of correlation between grey matter probability values and the cortical neuropathological measures in normal-appearing grey matter suggests that intrinsic cortical changes of the type we have measured do not influence grey matter probability maps used for VBM analyses.