Construction And Analysis Method Of Brain Network Based On Wavelet Coherence

Coherence analysis can identify the synchronization relationship between research phenomenon. The auditory cognitive brain network based on wavelet coherence is set up by EEG. It is able to analysis the organization principle of brain function and the law of dynamic evolution in the perspective of the whole brain, and deepen the understanding of human brain cognitive.This paper makes a depth research on the basis nerve electrophysiology, basic characteristics and basic approach of event-related potentials. An auditory perception experiment was designed in order to verify the construction and analysis of brain network based on wavelet coherence. We designed for three different stimuli which were using oddball model. Then we did the acquisition and pre-processing for EEG data. In addition, as the Fourier transform has a problem of insufficient accuracy in the analysis of non-stationary signals coherence, this paper presents a wavelet coherence method, which can show the characteristics of the signals in the time domain and frequency domain. We calculated and analyzed the wavelet coherence by auditory perception experiment, and obtained the synchronization state in various regions of the human brain by different auditory stimuli.To research auditory perception of human brain in a more comprehensive perspective, we used a complex network method based on wavelet coherence. We took the wavelet coherence value as the weight of edge in complex network, and constructed weighted network of wavelet coherence under different thresholds. Then we analyzed the weighted network of wavelet coherence by weighted clustering coefficient and weighted characteristic path length which has more cognitive significance. The result proved that the complex network has an obvious small-world property. And the human brain has diversity in different stimuli. At the same time this paper established the automatic determination method of threshold T and the automatic excavation method of active brain areas. By comparison with other brain network analysis indicators, this paper indicate that the complex network method based on wavelet coherence is effective in neural processing mechanism of sound diversity.Finally, this paper studied a minimum spanning tree method based on wavelet coherence to solve the threshold selection problem in complex network. We analyzed the properties of the minimum spanning tree combine short time window. Then we achieved the visualization of data relationship between brain regions, and got conclusion of cognitive science in dynamic evolution process between different brain regions.