P6-11 Comparison of dipole analysis and functional MRI in localization of epileptiform discharges

Objective: To test the relative accuracy of two different strategies for source localization of interictal spikes. Methods: 20 patients with intractable epilepsy underwent surgery for epileptogenic focus resection. Digital EEG (21 channels 10 20 & T1T2) recorded before surgery was analyzed. Spikes were automatically searched and classified according to morphology and topography, visually screened and averaged. A dipole fitting method (BESA) and a distributed source model (eLORETA) for source localization of interictal spikes was applied to the same data. Results: Both BESA and eLORETA located sources only within the resected area (RA) in 12 patients. BESA located multiple sources, most of them inside RA with few outside RA in 2 patients and eLORETA in 4. Both methods found sources predominantly outside RA, with still some within RA in three patients. BESA showed sources only outside RA in 3 patients and eLORETA in only one. The two methods gave discordant results in 5 patients. The best match to the RA was found in patients with mesiotemporal lobe epilepsy. When analyzing the 37 independent spike clusters found across the 20 patients, BESA located the spike within RA in 22 (59.5%), in the close vicinity (same lobe) in 2, in contralateral homologous areas in 4, in another lobe but still near RA in another 5, and elsewhere in 4. Localization with eLORETA gave sources within RA in 25 cases (67%), in the same lobe in two, in contralateral homologous areas in 4, in another lobe near RA in 4 and elsewhere in 2. Discordances in localization between both methods were found for just 5 of the 37 spikes (p < 0.001). Conclusions: Both BESA and LORETA localized most interictal spikes within the resected area. Some localization discordance was found, and the best match with clinical resection decisions was achieved when using both methods together.