Epileptic Seizure Prediction and Identification of Epileptogenic Region using EEG Signal

This paper presents a method to predict an epileptic seizure using synchrony analysis of EEG signals. The topographical map of the brain is divided into different regions and synchrony measures are computed between these different regions to identify the region responsible for epileptic seizure. The molecular and biochemical process of seizure generation during pre-ictal period is essential for seizure to start and this property has been investigated here in EEG signals to predict the onset of epileptic seizure. Particular region of the brain responsible for seizure is identified by pairing the EEG signals from different regions of brain and identifying the changes in synchrony measures corresponding to those regions. Two synchrony measures one from time domain, correlation, and other from frequency domain, coherence, are used in this work to validate the observations. Coherence and correlation increases in pre-ictal state and hence seizure onset can be predicted in advance. The results shows that epileptogenic region of the brain can also be identified.