Abstracts

FMRI has potential to replace IAP in presurgical evaluation of epilepsy surgery candidates. Studies report that a battery of fMRI tasks may be better than a single task in lateralizing and localizing cognitive functions. In this study, we compare 2 fMRI language tasks (verb generation [ndash] VGT and semantic decision/tone decision [ndash] SDTD) and IAP in their ability to lateralize language functions in epilepsy patients undergoing presurgical evaluation. 41 healthy controls received fMRI/VGT; 25 of them also received fMRI/SDTD task. Data on these subjects were used to generate global language region of interest (ROI) comprising Broca and Wernicke ROIs for each of the tasks. These ROIs were later applied to analyze epilepsy subjects. 10/13 epilepsy subjects also underwent IAP. FMRI tasks were presented in block design paradigms along with appropriate resting tasks. Behavioral data were collected for both tasks. Images were acquired over the entire brain on a 3T Bruker/Biospec 30/60 MRI scanner using a T2*-weighted gradient-echo EPI pulse sequence (TR/TE=3000/38ms, FOV=25.6x25.6cm, matrix 64x64 pixels, slice thickness=5mm, flip angle=90[deg]). Laterality Indices (LIs) for each task were calculated for each subject based on individual cross-correlation activation maps using average cross-correlation values. Pearson correlation between fMRI and IAP language lateralization was computed. Independent samples t-test was used to determine difference in means in LI between healthy controls and epilepsy subjects for the fMRI tasks. Activations were observed in all subjects with both tasks. Results for healthy controls show mean LI for VGT of 0.178 SD [plusmn]0.103, and for SDTD 0.261 SD [plusmn]0.12. Both tasks activated left-hemispheric Broca[apos]s and Wernicke[apos]s language areas. Significant mean LI shift towards more bilateral language distribution was noted for SDTD task in epilepsy patients in comparison to healthy controls (SDTD 0.15 vs. 0.26; p=0.04). Moderate or strong correlations were observed between VGT and IAP (r=0.718; p=0.013), SDTD and IAP (r=0.78; p=0.005), and between the two language fMRI tasks (r=0.815; p=0.001). Results indicate high correlations between language tasks and IAP and good convergent validity between fMRI language tasks. Results also confirm atypical (more bilateral) language lateralization in epilepsy patients in comparison to healthy controls which is related either to the effects of seizures on cortical plasticity or brain injury preceding the onset of seizures. Finally, either of these fMRI tasks can be used for language localization in epilepsy patients undergoing presurgical evaluation. (Supported by The Neuroscience Institute, Cincinnati, OH.)