Abstract
Around 1-2 % of the world's population suffers from epilepsy and 20 % of that vulnerable mass can't be cured through surgery or medicine. Epileptic seizures often occur unpredictably and may cause serious damage to the patient in adverse situations, for e.g. getting a seizure while driving or crossing a road can be fatal. In such a scenario, a reliable mechanism to predict the onset of seizure beforehand is much desirable. In this study, A reliable real-time technique for prediction of epileptic seizures is presented. Mean Phase Coherence (MPC), as a measure of phase synchronization, is used to predict impending seizures in a multi-channel EEG data. It was found that during the pre-ictal stages, MPC values and thus phase synchronization between various channels was found to drop significantly below the level in a non-ictal EEG signal. The range of the prediction horizon for seizures varied from 4-10 mins and prediction of impending seizure attack is possible for 8 out of 10 test cases.