A multi scale time–frequency analysis on Electroencephalogram signals

Abstract

It is well known, the Electroencephalogram (EEG) signals are non-stationary signals which helps us in understanding the complex brain dynamics, cognitive processes, etc. In this paper, we have analysed the healthy and Epilepsy seizure subjects through continuous wavelet transform using Morlet wavelet function. We identify the key differences between healthy and epilepsy seizure EEG signals’ low frequency periodic modulations and transient time varying patterns. Persistent intermixing of alpha and beta waves is found to be a key characteristic feature of the patients. The frequency intermixing is completely absent in signals from the hippocampal formation of the opposite hemisphere of the brain for the patients without seizure, akin to the healthy subjects. Our study further reveals dominance of frequency broadened gamma waves for seizure patients as compared to the low frequency regular alpha and beta waves for the healthy subjects. The time–frequency localization of wavelet transform clearly shows transfer of power to high frequency beta waves from the low frequency alpha waves in the signals of the epileptogenic zone of the patients. The observed frequency intermixing, reported here, is analogous to the bi-stability behaviour of dynamical systems.