Abstracts

Rationale: Previous studies demonstrated compromise to temporal lobe fiber tracts in patients with chronic MTLE. Additional data demonstrate progressive neocortical thinning within frontal and temporal lobe regions in refractory MTLE. This study investigates the relationship between fiber tract compromise and cortical thinning or volume loss within the specific regions to which each track projects. Methods: Diffusion tensor imaging (DTI) was performed in 26 patients with MTLE (13 left, 13 right) and 36 healthy controls. Fiber tracts were constructed using an automated probabalistic DTI atlas that was derived from manual tracings in healthy controls and MTLEs. Fractional anisotropy (FA) was calculated for four fiber tracts [uncinate fasciculus (UF), arcuate fasciculus (AF), inferior fronto-occipital fasciculus (IFOF), and parahippocampal cingulum (PHC)]. Based on previous neuroanatomical studies of fiber tract projections, 4 medial temporal, 3 lateral temporal, 5 frontal, and 2 cingulate regions were selected for inclusion in the analysis. Hippocampal and amygdala volumes were also examined. Pearson correlations were performed between FA values for each fiber tract and cortical thickness values within regions innervated by that fiber tract. To optimize power, patients with right and left TLE were combined in the analysis. Results: FA values for all four fiber tracts were lower in patients with MTLE relative to controls, suggesting widespread damage to fronto-temporal and temporo-occipital white matter. There were no correlations between cortical thinning of the entire right or left temporal lobe and fiber FA values. Rather, lower FA in the left PHC was related to left entorhinal thinning. Lower FA in the UF was related to cortical thinning in the left temporal pole and left orbitofrontal cortex, as well as volume loss in the left amygdala. Lower FA in the left IFOF was associated with cortical thinning in the left fusiform, lingual, and pars triangularis. Lower FA in the AF was related to cortical thinning in left middle temporal gyrus only. In addition, lower FA values for all four fiber tracts were associated with hippocampal volume loss. Fiber FA values were not related to age of seizure onset, illness duration, or number of anticonvulsant medications. Conclusions: Our results suggest that microstructural compromise to temporal lobe fiber tracts is not associated with global temporal lobe cortical thinning, but rather is associated with a marker of disease-severity (i.e., hippocampal volume loss) as well as cortical thinning within specific cortical and subcortical regions likely innervated by that fiber tract. These relationships were observed primarily within the left hemisphere. The neuropathological substrates of cortical and subcortical damage in MTLE are not established, nor it is clear whether such damage occurs simultaneously or in succession. Longitudinal multimodal imaging is needed to elucidate the dynamic versus static nature of cerebral pathology in MTLE.