Research On Epileptic EEG Signal Based On Randomly Distributed Embedding(RDE)

EEG is the embodiment of human neuron discharge.Pathological EEG contains very important information,which can not only cognitively improve human’s ability to understand the mechanism of the brain,but also help doctors to understand the brain in clinical practice.It provides a powerful reference for the diagnosis,treatment and prognosis of diseases and more importantly,lays the foundation for a deeper exploration of the potential of the brain in the future.Based on previous works,under the premise of small data volumes and high data dimensions,this paper applies two new methods to predict and analyze EEG signals.The specific research work is as follows:Firstly,feasibility analysis of Lyapunov exponent based on small data volume and Randomly Distributed Embedding in nonlinear chaotic system.In this paper,the analysis method based on the Lyapunov exponent of small data volume is selected,and excellent analysis results are obtained in the classical Logistic nonlinear model,which proves the superiority of this method for nonlinear dynamic model and chaotic system signal analysis.In this paper,the prediction method based on randomly distributed embedding is selected,and good performance is obtained in the prediction research of R(?)ssler chaotic model,which proves the superiority of this method for the prediction performance of nonlinear dynamic model.Secondly,EEG signal processing and analysis of normal and epileptic based on Lyapunov exponent with small amount of data.In this paper,the above algorithms are able to analyze the EEG signals of the epileptic group and the healthy group.The results show that the Lyapunov exponent of the epileptic patient group is always smaller than that of the healthy group around the optimal embedding dimension and the optimal embedding delay of the phase space reconstruction.The conclusion that the chaotic characteristics in the healthy brain system is greater than that of the brain system with disease,which confirms the previous conclusions.Thirdly,prediction and analysis of healthy and epileptic EEG signals based on Randomly Distribution Embedding.In this chapter,the above algorithm was able to predict the EEG signals of the epileptic group and the healthy group.Prediction of short-term high-dimensional EEG signals in epileptic patients and healthy controls using a Randomly Distribution Embedding framework and compared with other prediction methods.The results show that the Randomly Distribution Embedding has less error in the prediction performance of the future five-step state of EEG signals.Moreover,the five-step predicted RMSE value of random distribution embedding for EEG prediction of epileptic patients can be used as an important basis for significantly distinguishing healthy groups(p<0.05).