Cortical grey matter and benzodiazepine receptors in malformations of cortical development. A voxel-based comparison of structural and functional imaging data

Abstract

Using [11C]flumazenil-PET and statistical parametric mapping (SPM), we have shown recently that regions of increased and decreased benzodiazepine receptor density may be seen in patients with localization-related epilepsy due to malformations of cortical development. These abnormalities were seen both within and beyond lesions visually apparent on high-resolution MRI. We have also shown, using an interactive anatomical segmentation technique and volume-of-interest measurements, that subtle and unsuspected abnormalities of cortical grey matter volume were found in the same group of patients on high-resolution MRI, beyond the lesions visually apparent. In 10 patients with localization-related epilepsy and malformations of cortical development, we have now applied the automated and objective technique of SPM to the analysis of high-resolution structural MRI. Each individual patient was compared with 16 normal control subjects. We have then simultaneously compared the structural and functional data obtained for each individual patient with normal control high-resolution MRI and [11C]flumazenil-PET image using a novel technique. This comparison allowed the detection of functional abnormalities that were not accounted for by either visible or unsuspected structural abnormalities, in an automated and statistically rigorous manner. Five patients had abnormalities of cortical grey matter volume detected using SPM; only these five patients had been found abnormal using the previous volume-of-interest technique. Six of the 10 patients showed regions of cerebral cortex with disproportionate flumazenil binding compared with local grey matter volume. This included regions not found to have abnormal flumazenil binding on analysis of the PET data alone. Furthermore, regions found to have abnormal binding on examination of the PET data alone were, in some instances, shown to be accounted for by abnormalities of cortical grey matter volume. We conclude that the analysis of ligand PET data should always include a comparison with structural MRI; such comparisons are greatly facilitated by the novel approach described.