The Phase Synchronization And Coupling Direction Analysis Of Epileptic EEG Signals

Aiming at the need for the analysis of epilepsy, the epileptic EEG data analyzed by using phase synchronization and coupling direction theory in this thesis. Recently, phase synchronization and coupling direction are becoming the main approaches to analyze the epileptic EEG data, since this method can discover the interaction or information transfer among neural networks at the different areas in the brain. But, the application of coupling direction on multivariate EEG data is still in its infancy. There exist three methods on the estimation of coupling direction: the method based on the phase space reconstructions, evolution map approach based on the FFT and the method base on the information theory. In this thesis, above three methods are employed to comparative analysis of the EEG signal generating by epilepsy model and the clinical EEG data. At the same time, a new method on coupling direction is proposed to overcome the drawbacks of above three methods.First, multi-channels epileptic brain model are built in this thesis; then three coupling directional approaches are used to analyze the epileptic EEG data. To calculate the coupling directional index and synchronization strength index based on phase space reconstructions, the parameters of embedding dimension and delay time firstly should be estimated, herein the multiple-autocorrelation and Cao's method are employed to reconstruct a phase space of EEG recordings. In the method based on information theory, a Bayesian estimator is proposed to estimate phase coupling directional index, which greatly reduces the dependence on data length, improve the performance of computing during the real-time and gain a better accuracy of the results.The result of analysis shows that the directional approach can distinguish the coupling direction between two EEG data by analyzing the multi-models and real EEG. The interactions among the different brain areas can be foundthe phase synchronization change in the different periods of EEG is plotted. It is very helpful to understand the mechanism of phase synchronization on epileptic seizure.