Abstracts

Rationale: Patients suffering from non-lesional medial temporal lobe epilepsy can be divided into 2 sub-groups: TLE with mesial temporal sclerosis as defined by MRI (TLE-mts) and those with normal appearing MRI on visual inspection (TLE-no). Recent investigations have demonstrated separate patterns of brain structure abnormalities associated with both forms of TLE, suggesting that TLE-no may not simply be a milder form of TLE-mts but rather represents a distinct different clinical TLE entity. In this study, diffusion tensor imaging (DTI) parameters fractional anisotropy (FA) and mean diffusivity (MD) are used to investigate patterns of white matter alterations associated with both TLE-mts and TLE-no. Methods: 15 TLE-mts (6 left, 9 right onset), 15 TLE-no (9 left, 6 right onset) and 20 controls subjects were studied on a 4T MRI system and 4 EPI based DTI images (TR/TE = 6s/77ms, nominal resolution 2x2x3 mm, 6 encoding gradients, b=800s/mm) were acquired. After averaging of the 4 sets, eddy-current and geometric distortion correction was performed and FA and MD maps were calculated. The images of right TLE patients were right-left flipped so the ipsilateral hemisphere was on the left in all cases. Original and flipped control subject images were also included in the analysis to account for physiological hemispheric differences. The individual FA maps were non-linearly registered to a standard FA template and the derived transformation matrices were applied to the MD maps. The data was analyzed using tract-based spatial statistics (TBSS), from FSL with non-parametric voxel-wise t-tests. Permutation analysis with threshold-free cluster enhancement (TFCE) was used to correct for multiple comparisons. Results: Compared with controls, TLE-mts patients showed significantly reduced FA throughout the white matter of both cerebral hemispheres (ipsi > contralateral). A significant cluster of FA reduction in TLE-no was revealed in the anterior corpus callosum. There were no statistically significant clusters where FA values in patients exceeded those of controls. No statistically significant clusters were found in analysis of MD values. Conclusions: In this study widespread FA reduction, a marker for axonal damage, was demonstrated in TLE-mts, while localized FA reduction was found in TLE-no. In TLE-mts FA reduction was consistent with a previous whole brain analysis study. In TLE-no the FA reduction was confined to the anterior corpus callosum which supports previous work demonstrating structural grey matter damage in the frontal lobes. These results add to the growing literature indicating that brain damage in TLE is widespread not confined to the temporal lobes. The TBSS method did not detect significant MD abnormalities in either TLE group. TBSS has been previously shown to be an optimal method for detecting subtle FA changes; however, analysis is confined to the mean tract skeleton which may not be optimal for MD changes, which may occur in the grey matter.