Analyzing Spatio-temporal Patterns of Epileptic EEG Signals by Dynamic Mode Decomposition (Mathematical study on Epilepsy)

Abstract

In this work, we apply a dynamic mode decomposition (DMD) to epileptic EEG data to figure out the spatio-temporal patterns in the data. It is found that some patterns with high frequency have direct influence on the epileptic seizure, and instantaneously capture the abnormal patterns of the neuron firing with high precision. We prepare an ictal and interictal EEG data for an epileptic patient, and investigating the dyamics of the DMD modes captured by the proposed algorithm for the windowed EEG signals. From the results we conclude that the DMD modes extracted from the EEG signals can be useful for analyzing and understanding the dynamics of the epileptic EEG data.