Research Of Brain Functional Network Based On Graph Theoretical Approach

The human brain is a nonlinear and non-stationary complex dynamic system formed by alarge number of interconnected neurons. When a lot of cortical pyramidal neurons are excited,electroencephalogram (EEG) can detect the change of scalp potential. EEG is one of thenoninvasive imaging techniques with simple operation and high temporal resolution and manystructured, functional and pathological informations can be acquired，and it is an effectivemethod for brain fuctional analysis, disease diagnosis, prevention and treatment.Acupuncture therapy is the treasure of traditional Chinese medicine. After a long term ofdevelopment and evolution, the acupuncture therapy, which originated from our ancientancestors’ living practice, is forming a unique and complete disease diagnosis and treatmentsystem. Therefore, how to explain the traditional acupuncture mechanism more scientifically hasbecome a pressing problem putting in front of our researchers.In this article, in order to explore the reaction of acupunctured brain, the data is processedand analysed base on correlation matrix and graph theoretical. Firstly, an experiments is designedwhich acupuncture the subjects at ‘Zusanli’ acupuncture point (ST-36) with different frequencies(i.e.50,100,150and200times/min) respectively, and the20-channel EEGs are recordedsimultaneously. The data of EEGs is preprocessed with the band-pass filter and threshold methodto eliminate frequency interference and artifact.Secondly, a method of correlation matrix analysis (CMA) is presented to research thesynchronization of multi-channel EEGs signal. After the normalization of pretreated EEGs, thecorrelation coefficient between each two channels is obtained and a correlation matrix isconstructed. By repeating the retention to obtain a proxy data with the same power spectrum asthe original data’s and randomed correlation between each two channels, then obtain the proxymatrix as the correlation one, and decompose the two matrixes. The synchronization value of themulti-channel EEGs is quantized to synchronization analysis. The results show that thesynchronizations of most subjects’ EEGs and the number of cluster synchronization increasesignificantly after acupuncture.Finally, analysis on the brain functional network is conducted based on graph theoreticalapproach (GTA). GTA has been applied to the solution of practical problems, and it is the mostimportant method of complex network analysis. Then a complete brain functional network withthe correlation matrix constructed in synchronization analysis by choose appropriate threshold.Because of alpha is most evident band when people are sober quiet and eyes closed, the data of EEGs is decomposed with wavelet multi-scale mothod to extract alpha band. Lastly, the brainfunctional network of EEG and alpha band is established and comparative analysis on thecharacteristics of EEG and alpha band networks. The results show that the node degreedistributions of the networks change obviously, it means that the scale-free characteristicenhance after acupuncture. The global efficiency of brain and alpha band nerwork decrease, andthe latter’s local characteristic enhances.The significance of this topic research is according to the different acupuncture state andfrequency of the brain EEGs, to probe the characteristics and regular pattern of human brain, to acertain extant, it verify the means of acupuncture on the human body the positive influence.