Research On Emotion Recognition Method Based On PLV Brain Network

The human brain plays a major role in the generation of human emotion.Research on emotion recognition based on Electroencephalogram(EEG)has attracted increasing attention from scholars at home and abroad.Effectively mining the functional mechanisms of the brain provides new insights for emotion recognition,which is essential for understanding the connection patterns of human brain tissue and the way emotions are generated from a system-level perspective.The key to emotion analysis is to understand the rules of functional division and collaboration of different brain regions.However,existing emotion recognition research has two shortcomings in this regard: Most researches on the brain focus on the energy and global perspectives,but ignore the functional correlation between different brain regions and the effective connection characteristics of local brain regions;and the information interaction mode of the brain when processing emotional activities is time-varying and highly connected.At present,only the analysis of the brain at the static level cannot effectively reflect the dynamic change rule of the connections between brain regions when emotions are generated.The main content and innovation points of this paper aim to solve the above two problems,as follows:(1)In order to solve the problem that the topological features of brain network at global scale cannot reflect the coupling relationship between brain regions completely,this paper proposes an emotion recognition method based on the phase synchronization relationship between EEG channel information at global and local scales.Phase Locking Value(PLV)is used to measure the phase synchronization degree of EEG from each channel to describe the functional connection relationship of the brain.After constructing the PLV brain network,this paper integrates the two attributes of functional integration and separation to analyze the differences in brain connectivity of different emotions.Furthermore,the modular structure of the brain network is constructed through community detection to extract its more comprehensive local characteristics for connectivity analysis.The research results show that compared with positive emotions,the brain regions under negative emotions have higher phase synchronization,more complex brain connectivity patterns,and more obvious modular structures.In addition,this paper found that the key cortical brain areas related to emotional stimulation act as the brain network hub to dominate emotional processing activities.Through the verification on the DEAP data set,it is found that the analysis framework of this research effectively improves the accuracy of emotion recognition.(2)In order to solve the problem that EEG analysis from a static perspective alone cannot discover the time-variation of information interaction patterns when the brain processes emotional activities,this paper proposes a framework of emotion analysis,which quantifies the phase synchronization relationship under different time windows to reveal the rules of functional dynamic coordination of different brain regions.There are such central regions(hubs)in the brain,which are highly connected and highly central,and are responsible for coordinating the dynamic interaction between brain regions.This paper constructs emotion-related dynamic brain networks,and studies the structure of these network hubs that reflect functional connectivity.This framework proves that network hubs have formed a closely connected organization called "rich-club",which can be used as the core network architecture for identifying different emotions.The results show that the dynamic connection patterns produced by different emotions are different,which is manifested in the different rich-club organization.The discovery of the rich-club structure related to emotions in this study is of great significance for studying the brain mechanisms behind human emotional activities.