Analysis of Electroencephalogram Background Activity in Epileptic Patients and Healthy Subjects Using Dispersion Entropy

Hypothesizing that analysis of electroencephalogram signals using nonlinear and deterministic chaos theory may provide clinicians with information for medical diagnosis and assessment of the applied therapy, in this study we investigated the electroencephalogram background activity in normal and epileptic subjects using a newly defined quantifier adapted from nonlinear dynamics and deterministic chaos theory: dispersion entropy. Dispersion entropy quantifies degree of variability of a time series. Receiver operating characteristic analysis reveals that Dispersion entropy can discriminate between seizure and non-seizure states with very high accuracy. The application of such a technique is justified by ascertaining the presence of nonlinearity in the electroencephalogram time series through the use of surrogate test. The false positive rejection of the null hypothesis is eliminated by employing Welch window before the computation of the Fourier transform and randomizing the phases, in the generation of the surrogate data. Paired t-test revealed significant differences between Dispersion entropy of both the original and surrogated time series, indicating the presence of deterministic chaos in the original electroencephalogram time series.