Estimating complexity from EEG background activity of epileptic patients

Abstract

We analyze the EEG background activities of five epileptic patients and compare them with the EEGs of five control subjects by means of methods from nonlinear dynamical system theory. We divided the EEG time-series data into frames to select those that fulfilled the stationarity condition, which is necessary in order to apply our methods. Then, we reconstructed the chaotic dynamic attractors expressed in the phase space and calculated their correlation dimensions, which are a measure of the complexity. Furthermore, we recorded an EEG during a petit mal epileptic seizure, whose attractor is a 3-D plot and the correlation dimension will also be presented.