Abstracts

Rationale: Whole brain morphometrical MRI studies have shown extra-hippocampal atrophy in patients with medial temporal lobe epilepsy (MTLE). Structural brain compromise in MTLE beyond gray matter atrophy is less understood. It has been hypothesized that extra-hippocampal gray matter atrophy in MTLE is a consequence of hippocampal disconnection rather than the effect of seizures. This study investigated the relationship between white matter loss, hippocampal disconnection and extra-hippocampal gray matter atrophy in MTLE, employing 3 different MRI methods: (1) voxel-based morphometry (VBM), (2) diffusion tensor imaging (DTI) and (3) probabilistic tractography (PT). We hypothesized that patients with MTLE show a relationship between limbic gray matter atrophy and loss of hippocampal connections, suggesting a deafferentation mechanism underlying part of extra-hippocampal atrophy. Methods: We studied 23 patients with MTLE and unilateral hippocampal atrophy (8 left and 15 right) and 34 matched healthy individuals. Subjects were scanned in a Philips 3T MRI (yielding a T1-weighted image and a 15 direction DTI). VBM was performed on T1 images (with SPM5). The mean gray matter volume (GMV) was extracted from the regions of interest (ROI) comprised in the Anatomical Automatic Labeling Atlas (AAL). From DTI images, we derived fractional anisotropy (FA), mean diffusivity (MD) maps, and probabilistic tractography (with origin seeds placed on the hippocampi), using the software FSL. A Manova was used to investigate differences in GMV between controls and patients. Voxel-wise analyses evaluated differences in FA, MD and PT between patients and controls. Z-scores were computed for GMV from AAL ROIs and peri-hippocampal FA and MD (in a ROI encompassing all fibers from PT from the hippocampus), and correlated, to investigate the relationship between hippocampal deafferentation and regional GMV atrophy. Results: Patients with MTLE exhibited a significant widespread reduction in GMV. They also exhibited disruption in white matter connectivity. A reduction of FA was observed in peri-hippocampal and widespread limbic areas in patients with left and right MTLE; and patients with right MTLE exhibited a significant increase in MD in temporal and frontal regions (Figure 1 upper panel). Patients with MTLE showed a significant reduction in hippocampal PT in limbic areas (Figure 1, right panel). There was significant relationship between standardized (Z-scores) loss of hippocampal connections and GMV atrophy, particularly in frontal and temporal limbic areas (representative regions are shown in table 1). Conclusions: These results suggest that white matter connectivity is abnormal in patients with MTLE. There is a relationship between brain atrophy and hippocampal disconnection, indicating that hippocampal deafferentation plays a role in extra-hippocampal GMV atrophy. These abnormalities may account (independent from gray matter atrophy), for the neuropsychological deficits commonly observed in patients with MTLE.