Abstracts

Rationale: Patients with left-sided temporal lobe epilepsy (TLE) and presurgically intact verbal memory are at highest risk of postsurgical verbal memory deteriorations. We investigated whether left-sided TLE patients with and without severe postsurgical verbal memory loss differed according to fMRI activation patterns associated with a simple fMRI word generation task. We studied whether presurgical language activations were correlated with postsurgical changes in verbal memory. Methods: We studied 16 left-sided TLE patients with hippocampus sclerosis with presurgically unimpaired verbal memory (percentiles >16 on the direct and delayed recall trials of the Verbal Learning and Memory Test, VLMT). Seven patients had postsurgical verbal memory decline (>1 SD reduction on the delayed recall trial of the VLMT), whereas nine did not have memory alterations. Prior to mesiotemporal resection all patients were investigated with a fMRI language task. In this blocked fMRI task patients had to generate as many words as possible starting with a given letter in each of the 10 activation conditions. Results: The whole patient group exhibited left-sided activations within the lateral prefrontal cortex, mainly in the inferior frontal gyrus (p<.001 corrected). Those patients with postsurgical memory decline had stronger activations within the left posterior middle temporal gyrus (x=-45, y=-54, z=0, k=30, p=.02 corrected) compared to the patients with preserved verbal memory. Patients without postsurgical memory decline did not have stronger activation within any brain region. We did not observe any correlation between prefrontal lobe activation and postsurgical memory outcome. However, a regression analysis revealed a negative association between postsurgical memory alterations and presurgical language activation within a small cluster located in the left posterior superior temporal gyrus (x=-39, y=-45, z=12, k=13, p=.06). Conclusions: We did not find patterns of stronger activation in those patients without postsurgical memory decline compared to those with memory loss. This result indicates that reorganization capacity in the context of verbal memory cannot directly be demonstrated with an fMRI language task. Nevertheless, our results suggest that activation within the left lateral posterior temporal lobe is associated with verbal memory outcome. Those patients with activation increases within the left posterior temporal lobe seem to be at a higher risk for postsurgical memory decline. Further studies with larger patient groups are needed to clarify whether or not an individual prediction of verbal memory outcome in patients with left-sided TLE can be made by language associated fMRI activations.