The Research Of Multi-channel Eeg Synchronization Algorithm Based On Ordinal Pattern

EEG synchronization is considered to be the important performance of the brain to communicate, interacte, coordinate information between various regions. So the synchronization research on different regions of the brain electrical signal is of great significance to explore the collaboration between brain regions and to deep understand the brain disorder mechanism. Based on ordinal pattern, this paper studied the coupling strength and directionality of EEG signal.First, after studying the ordinal pattern and mutual information, normalized weighted-permutation mutual information(NWPMI) was developed for analysing coupling strength analysis of dual channel signals. Simulation showed NWPMI method had stronger robustness to the noise. Combined with NWPMI and S-estimator method, a global synchronization index named S-estimator based on Weight Normalization Permutation Mutual Information(SWNPMI) was developed. And SWNPMI method was used in the analysis of epileptic EEG signals and diabetes aMCI EEG signals.Second, the simulation analysis on Motif-Synchronization(MS) algorithm was done. Further based on MS and S-estimator, synchronization index, global synchronization index, three kinds of multi-channel EEG synchronization algorithm were proposed, respectively called S-estimator based on Motif-Synchronization(SMS), Synchronization Index based on Motif-Synchronization(SIMS) and Global Synchronization Index based on Motif-Synchronization(GSIMS). They were more simple, less amount of calculation than SWNPMI. The simulation showed they could better reflect the synchronization strength between multi-channel EEG signals. Especially, SMS algorithm was simplest. So SMS algorithm was used to analyze diabetes aMCI EEG signals. Results showed that the two groups existed significant differences. Pearson linear correlation analysis between cognitive function and EEG synchronization was done, and there was a certain correlation between cognitive function and EEG synchronization.Finally, permutation conditional mutual information(PCMI) algorithm was studied. Experiments showed that the coupling directionality estimated by PCMI could more accurately reflect the relationship between coupling coefficient, and PCMI method Had better ability to resist noise. Moreover, the coupling directionality of epileptic EEG signals and diabetes aMCI EEG signals were analyzed by PCMI method, and the results were consistent with the pathology.