Research On Tactile Emotional Eeg Based On Fusion Feature

Haptic is one of the five senses of the human body,which plays an important role in emotional and interpersonal communication.Haptic can enhance the meaning of other forms of communication,such as vision,hearing or language,and provides a very powerful way to stimulate and regulate human emotion,so it is particularly important to explore the relationship between touch and emotion.With the development of haptic perception,sensor technology and computer technology,EEG as a kind of electrophysiological signal,can directly reflect people’s emotional state.Therefore,in this paper,haptic and EEG signals are combined to study the effect of haptic on emotion.The main contents of this paper are as follows:(1)The emotional experimental paradigm is designed to induce emotion.In this paper,not only the haptic jacket based on Arduino is developed,but also three haptic vibration modes corresponding to happiness,fear and sadness are designed,which can effectively express the target emotion of the subjects.In addition,this paper combines the vibration effects of different haptic jackets with the emotional flashpoints of the film to form haptic-visual-auditory emotional stimulation.(2)A haptic emotion analysis method based on time-frequency is proposed.In this paper,haptic emotional EEG and non-haptic emotional EEG are analyzed by time spectrum analysis and emotional brain region analysis,and the characteristics of power spectral density,differential entropy and wavelet entropy of four frequency bands are extracted.In order to solve the problem of redundancy in the high feature dimension of EEG,the t-SNE algorithm is proposed to reduce the feature dimension.Finally,the SVM algorithm is used to identify haptic emotional EEG and non-haptic emotional EEG.The results show that the classification performance of haptic emotional EEG is better.(3)A brain network analysis method based on PLV is proposed.In this paper,a brain functional network based on PLV is constructed for haptic emotional EEG and haptic emotional EEG,and phase synchronization analysis,network attribute analysis and key connection analysis are carried out.The results show that the network synchronization region and intensity of haptic emotional EEG are generally higher than those of the brain network without haptic state.Five kinds of brain network attributes such as node degree,clustering coefficient,average path length,global efficiency and local efficiency are used as features for emotion recognition,which verifies that PLV brain network in haptic state has more advantages in emotion recognition.(4)In this paper,a haptic emotion recognition system based on fusion feature and Stacking integrated model is proposed.We combine the time-frequency domain characteristics of EEG signals with the brain network characteristics to form the timefrequency-space feature.In addition,an integrated classification model based on Stacking algorithm is designed for emotion recognition.The results show that the classification effect of the proposed fusion feature is better than that of single feature,and the classification effect of Stacking integrated model is better than that of base classifiers.