Improving diagnosis of some brain disease by analysing chaotic indices of EEG signals

Abstract

This paper investigated the effect of Electroencephalogram signals on diagnosis of the possible brain dysfunctions. It is the most common non-offensive way to analyse brain healthiness on the basis of chaotic theory. Normal person's EEG signals are different from the people who suffer from epilepsy, schizophrenia, post-traumatic stress disorder and Alzheimer's diseases in many respects, namely amplitude, frequency, statistical features and in general dynamic behaviour. The findings indicate EEG signals are non-linear and chaotic. EEG signal was considered as chaotic time series in the research. Diagnosis of diseases was conducted using analysis of chaotic parameters. They include Lyapunov exponent and correlation dimension. For this end, appropriate algorithms were rendered to extract necessity parameters for reconstructing phase space and calculated chaotic indices with relative consideration. The findings showed patients are recognised from healthy persons. It was also possible to distinguish two types of epilepsy, namely grand mal and temporal lobe. The results showed acceptable way of predicating epilepsy. Visual diagnosis of disorders by EEG is a big challenge for neurologist as complexity of the brain dysfunctions. This research can be useful for physicians to diagnose and predict diseases. © 2016 Inderscience Enterprises Ltd.