Prediction of Epileptic Disease Based on Complex Network

Abstract

The purpose of this study is to observe epilepsy brain network evolution from network perspective and implement of epileptic disease prognosis. Local visibility graph method is on the basis of visibility graph method adding a sliding time window and building a number of sliding time window with the complex network topology. It is in order to observe the time dependence of the network. We divided the electrocorticogram(EEG) time series into three parts. They were the time series during normal period, pre-epilepsy period and seizures occur period. Then build three network topology graphs and observed its evolution process. The results show that the network module structure of the epileptic EEG from normal period to pre-epilepsy period then to seizures occur period disappeared. And it form the arc of the zonal distribution. These characteristics of complex networks provide new ideas for the prediction of epileptic disease.