Localization of brain tissue responsible for initiating seizures is an essential procedure that precedes surgical removal of this tissue in patients with intractable partial complex seizures. In the best case, good localization can be achieved with visual analysis of seizure onsets recorded from hours of scalp recorded EEGs. Unfortunately, good localization can often only be achieved by surgical implantation of electrodes on the surface of, or in the brain. We wish to facilitate research aimed at reducing the need for surgical localization by extracting more information from scalp recorded EEGS. To accomplish this, we propose to develop an epilepsy localization workstation on a personal computer. The system will include software tools to (1) record live EEGS, as well as read, display, and edit long records of multichannel EEG data stored on video tape, (2) perform automated and manual artifact detection, (3) detect epileptiform events (4) perform mathematical source localization, and (5) register sources with MRI images. In Phase 1, we will conduct surveys to determine the needs of epileptologists, review the literature pertaining to source localization and the use of realistic models of the head and sources, develop a taxonomy of source models, and design the system. The full system will be implemented in Phase II.
