The Synchronization Research Of Two-channel EEG Signals Based On Coherence Algorithm

Electroencephalogram(EEG)is a tool of diagnosing and researching common clinical,it is one of the technologies of noninvasive study the human brain function activities.The change of EEG signal contains important biometric information.These characteristics are great significance to grasp neurological disease.The purpose of this paper is to develop the consistency method of two-channal EEG signals,and use this method to study EEG signals of diabetes and epilepsy patients.Firstly,after studying the minimum variance distortionless response algorithm,the three kinds of smooth consistency window algorithms are researched,and the simulation results of the three methods should be improved in accuracy.Secondly,the improvement new method the squared magnitude coherence of smoothing of the minimum variance distortionless response algorithm is proposed,and it uses the Cheriet-Belouchrani kernel,which is often used in time-frequency analysis with compact support properties.The simulation experiment shows that the smoothing of the minimum variance distortionless response has higher accuracy and anti-noise performance in the narrow-band signal and broad-band signal.Finally,using the smoothing of minimum variance distortionless response algorithm to study the 31 patients(18 aMCI,13 the control group)with diabetes and epileptic EEG signals.Resting state EEG sequence is broken down into four bands(delta,theta,alpha,beta),in each frequency band of different brain areas to study the square coherence.Through the statistical analysis of the consistency between the two channels on both sides of the brain,the one side of the brain with long distance and short distance to find frequency bands and channel groups with significant features.Diabetes among patients with mild cognitive dysfunction and normal patients have obvious differences in the frontal-right temporal,frontal-occipital,right temporal-occipital,left-right temporal.The analysis results consistent with the pathological and there is a correlation between cognitive function.Experimental results show that the analysis of EEG coherence in the clinical has diagnostic significance,for diabetes patients can be diagnosed earlier aMCI,reaching the purpose of assist physicians.The statistical of epileptic EEG signals shows that there is difference in ictal and interictal periods,which provides a basis for the diagnosis of epilepsy.