Electroencephalogram Analysis Based On Time Series From Network

The brain can be regarded as a complex network system, and the function unit of the system can be seen as the node of the network. The research on the conversion of the electroencephalogram(EEG) signal to the network has become a hot topic in the study of EEG signals. The study of time series converted from complex networks has been ignored. In this paper, we use the method of the reverse analysis of EEG signal, and the analysis of the time series. The main work of this paper is as follows:Firstly, this paper presents a method to analyze epileptic EEG based on time series from improved k-nearest neighbor network. The single lead Epilepsy EEG signal was converted into network based the improved k-nearest neighbor network. The power spectrum of the time series from network and the original time series are analyzed and compared. The experimental results show that studying power spectrum of time series from network is more easily than power spectrum of time series directly generated from brain data to distinguish between normal people and the patients with epilepsy. In addition, the clustering coefficient of the network is also able to distinguish between normal persons and patients with epilepsy. The research can provide important reference for the study of epilepsy and clinical diagnosis.Secondly, this article analyzes different attention state EEG based on the time series generated from improved k-nearest neighbor network. The experimental results show that this method can effectively distinguish between counting number state EEG signals and closing eyes state EEG signals. This study can provide an important reference for the study of EEG.Thirdly, the thesis puts forward the method of analyzing time series from network based on Kendall non-parametric coordination factor, and the method is used to analyze the EEG epilepsy. In the proposed method, the 16-lead original EEG is constructed into network by using the extended Kendall rank correlation coefficient. Then the network is transformed into time series, and mean value of maximum power spectrum of the time series is analyzed. In this paper, appropriate Kendall non-parametric coordination factor is selected by experiment. The experiment results show that, compared with analyzing mean maximum power spectrum power spectrum of original time series or the characteristics of brain electrical network, the average value of the maximum power spectrum of the time series from the extension of the Kendall network is more effective to distinguish between normal people and patients with epilepsy. The research can contribute to the clinical diagnosis and analysis of epilepsy.