Prediction of Memory Disorders Using Non-Linear Analysis of EEG

Abstract

The EEG signal represents the electrical activity of brain. The bio-signals are highly subjective that appears random in time scale. This paper describes the non-linear analysis of time series data of EEG signals of normal subjects. The EEG data is obtained from physionet data base. The dataset consists of  EEG signal acquired during mental task done by the normal people. The conditional volatility of the EEG signals is investigated using the 1D Gaussian GARCH Model. This GARCH model will be one of screening tool for identifying EEG-Mental Disorders in future.