Seizure Prediction System--Clinical Tool for Epilepsy

Epilepsy is a neurological disorder, second only to stroke, that affects 1% of the world's population. The overall aim of this project is to develop a commercial epileptic seizure warning and prediction (ESWP) software package for use in electroencephalographic (EEC) monitoring systems from multiple manufacturers. Because of its innovation and the immediate need for such a system in a clinical and research setting, its implementation and testing in one clinical center will be the focus of Phase I, and its validation in several clinical centers and settings will be addressed in Phase II. The proposed ESWP software will use the STLmax algorithm for seizure prediction (lasemidis et al., 2003), developed on the basis of a nonlinear dynamical analysis of EEG, as well as implement and use other potential seizure prediction algorithms. Initially tested on continuous off-line EEG recordings of 0.76 to 5.84 days in duration, from 5 patients with refractory temporal lobe epilepsy, STLmax resulted into 82% sensitivity of seizure prediction, a false prediction rate of 0.16 per hour, and a warning on average 71.7 minutes prior to a predicted seizure. The following overall goals will be accomplished in Phase I of this project: (1) Test data analysis algorithms implemented in ESWP system using access to EEG data acquired with one of the existing commercial EEG instruments, analyze these data in real time and on-line to issue epileptic seizure warnings 30-90 minutes prior to a seizure onset. (2) Test and refine a clinically relevant graphical user interface (GUI), which will control ESWP and allow the medical personnel (technicians, nurses, physicians) to use the seizure warnings to improve the diagnosis, prognosis, and treatment of epileptic patients. The prospective Phase II of this project will accomplish the following aims: (1) Refine and extend the ESWP software package to work with other major existing commercial EEG monitoring systems en route to commercialization and licensing to EEG system manufacturers, (2) Perform extensive clinical studies to test the performance of the ESWP software in mutiple clinical centers. Successful implementation of the proposed system would find several clinical and research applications for the treatment of epilepsy, e.g. improve patient safety by alerting the existing nursing staff long prior to a seizure, time the doses of anticonvulsants (drugs or electromagnetic stimuli) to improve their efficacy in the control of seizures or time ictal diagnostic procedures in radiological studies for epileptogenic foci localization.