Abstracts

Rationale: EEG and histopathology suggest that structural and functional abnormalities in TLE-no are less severe and more widespread than in TLE-MTS. Cortical thickness measurements might be better suited to detect such subtle abnormalities than conventional volumetric techniques which are often negative in TLE-no. The aim of this study was to seek and characterize patterns of cortical thinning in TLE-MTS and TLE-no. Methods: T1 weighted whole brain images and high resolution T2-weighted hippocampal images were acquired on a 4T magnet in 68 subjects (35 controls, 13 TLE-MTS, 20 TLE-no). FreeSurfer was used for cortical thickness measurements. To allow for the combination of left and right TLE in the analysis, the MRI of right TLE were side-flipped. Control MRI were also side-flipped and original and side-flipped data was used in the analysis to account for physiological side differences. Group comparisons were done using the statistical routine of FreeSurfer (FDR p = 0.05). Hippocampal subfield volumes were obtained using a manual marking scheme on high resolution T2 images (1). To allow for focus identification and description of thinning patterns in single subjects subfield volumes and mean thickness of each cortical label of the FreeSurfer numerical output were converted into z-scores. Results: TLE-MTS showed the most prominent cortical thinning in ipsilateral medial temporal-occipital regions. In TLE-no the thinning was less well lateralized and most prominent in the temporal-inferior, temporal-lateral and inferior frontal region (cf Fig 1). The hippocampus contained the region with the lowest z-score in 10 TLE-MTS (77%), in the remaining 3 the lowest z-score was in the contralateral hemisphere (temporal, cingulate and occipital). In contrast, only 1 TLE-no (7%) had the lowest z-score in the ipsilateral hippocampus. In the remaining TLE-no, it was in the temporal lobe (5 contralateral), in the frontal lobe (1 ipsi-, 1 contralateral), in the parietal lobe (1 ipsi-, 2 contralateral) and in the cingulate (2 contralateral), i.e., in 80% of the TLE-no the region with the most pronounced tissue damage was not concordant with the EEG focus identification. On average, 2.7 lobes in TLE-MTS (range 0-9) showed regions with severe thinning (z-score≤-2) and 2.5 lobes (range 0-7) in TLE-no. Despite the findings in the group analysis, neither TLE-MTS nor TLE-no showed a consistent pattern of thinning in the individual analysis. Conclusions: To our knowledge, this is the first study showing extensive cortical thinning in TLE-no. The pattern of thinning was different from the pattern in TLE-MTS which was consistent with structural changes described in TLE-MTS by other techniques. Together with the fact that the hippocampus was mostly intact in TLE-no, this further supports the hypothesis that seizures do not originate from the hippocampus but from other cortical regions and that cortical thinning in this group represents a mixture of primarily and secondarily affected regions. Reference: 1. Neurobiol Aging 2007 28:719-26