Abstracts

Rationale: Effective source localization via scalp EEG is a highly desirable component for the treatment of epilepsy for a variety of reasons. However, many techniques are limited in the accuracy of their estimations as a result of the fact that they utilize average atlas models as opposed to head models generated using subject specific information. Indeed, a technique of source estimation which utilizes head models derived from individual subject head tissue data, as gathered from CT's, MRI's, etc., which is both accurate and streamlined to the point that it can be used in a clinical setting, has not been developed. This study will attempt to demonstrate that electrical imaging software which utilizes subject-specific head models, yields source localization results which are at least as accurate as those generated by a similar software package which uses an average atlas head model. Methods: Our clinical procedure will include (1) a physician visually identifying interictal epileptiform activity in the form of spikes and discharges, (2) source localization of said activity with both GeoSource 2.0 and (3) GeoSource 3.0, (4) a comparison of the results generated by the two software packages, and (5) validation of the results obtained through variables such as intracranial EEG and surgical outcomes, when applicable. Results: In this preliminary study, 14 patients with refractory epilepsy that exhibited ictal and interictal discharges are examined and compared. The preliminary results demonstrate that patient-specific head models yield source localization results which are at least as accurate as those generated by standard atlas head models. Conclusions: The preliminary results demonstrate that EEG source localization software packages using subject-specific head models are at least as accurate as those which utilize averaged atlas models. Thus, the utilization of software packages take advantage of the opportunity to include patient-specifc MRIs in identifyfing regions of the brain associated with epileptic activity are likely to serve as a clinically important tools in the evaluation and treatment of difficult epilepsy.