Modular based dynamic analysis of EEG signals using non-linear feature

Most of the real systems including a large number of physical, physiological and biochemical signals exhibit non-stationarity or time-varying behavior. Electroencephalogram is brain signal that allows gaining the understanding of the complex inner mechanisms of the brain and abnormal brain waves and plays a vital role in diagnosis of different brain disorders. We have carried out a study for nonlinear feature extracted from epochs of epileptic signal; and classification of EEG signals with feature from various epochs of the signal. The nonlinear properties of the time series are investigated by calculating Hurst exponent values during epileptic seizures, and in the interval between the seizures. During uncontrolled electrical discharges, the long-range correlation effects do appear in EEG signals in all cases, as the Hurst exponent values were above 0.5. It is observed that the proposed method can provide better performance, is much efficient and faster as compared to the time-frequency based techniques while classifying and discriminating seizure activities.