Abstracts

Rationale: Topiramate (TPM) is well known for its negative effects on cognitive performance in healthy controls (HC) and in patients with epilepsy. Various studies have demonstrated these behavioral effects either in comparison to placebo or to antiepileptic drugs (AEDs). Only one small fMRI study directly assessed the effect of TPM on language-related activation. The knowledge of the effects of TPM on fMRI activation is of paramount importance in epilepsy patients undergoing presurgical evaluation where fMRI results are increasingly used for surgical staging. Therefore, the goal of this study was to assess the effects of TPM on fMRI BOLD signal.Methods: This study included 8 patients with right temporal lobe epilepsy (RTLE) and 8 with LTLE taking TPM (TPM+) that were matched for age and handedness to 8 R and 8 LTLE not taking TPM (TPM-). HC were matched 2:1 for age and handedness (N=32) to R and LTLE TPM+ patients separately. Non-TLE patients were specifically excluded to provide a relatively uniform sample. The fMRI paradigm involved semantic decision/tone decision task (in scanner behavioral data were collected). FMRI data were processed using CCHIPS including motion correction and spatial normalization. For each subject, voxels activated in the fMRI task were identified using GLM; random-effects analysis was performed to determine significant group activation differences; only activations with an uncorrected p<0.001 were included (Z score >3.091). Spatial smoothing of 4 mm and cluster threshold of 20 voxels were applied. One-sample t-tests were performed within each group to assess task-specific activations. Direct group comparisons between RTLE TPM+ vs. TPM- patients and then RTLE TPM+ patients and HC were made. Similar group comparisons were made for the LTLE patients. Results: Groups were matched for handedness and age. There were no differences in the epilepsy groups for age, age of epilepsy onset/duration and number of current AEDs. The in scanner performance of patients was worse when compared to HC all p<0.043 except the comparison for semantic decision in LTLE+ vs. HC (p=0.094). Overall, all groups showed activation in regions typical for this task. Regression analyses comparing LTLE TPM+ vs. TPM- showed significant differences in that TPM+ patients had shifts of BOLD signal to the right hemisphere when compared to the TPM- patients and their overall activation level was higher. For RTLE patients, the pattern of activation differences was overall similar to LTLE group. Comparisons with the HC showed overall much higher BOLD signal in TLE patients than in HC.Conclusions: While patients with TLE performed worse than HC on behavioral measures, as a group they showed higher BOLD signal than HC. This may suggest higher (and probably less efficient) utilization of cortical resources in epilepsy patients. These differences were more striking in TPM+ vs. TPM- patients including changes in lateralization of language functions. These results suggest that fMRI data in TPM+ patients need to be interpreted with caution as TPM may affect not only cognitive performance but also BOLD signal.