Application of EEG Signal Recognition Method Based on Duffing Equation in Psychological Stress Analysis

Abstract

Based on the study of the feature extraction algorithm based on the multiple empirical mode decomposition of the Duffing equation, this paper proposes a corresponding improved algorithm, completes the identification and analysis of the psychological pressure dimension space under the audiovisual induction method, and designs two typical psychological types of music and pictures. Based on the stress induction experiment, an audiovisual-induced psychological stress recognition system based on EEG (electroencephalogram) signals was built. Aiming at the problem that the spatial uniform sampling method cannot well reflect the dynamic characteristics of the multivariate EEG signal, based on the Duffing equation, a nonuniform sampling algorithm that adaptively selects the projection direction is proposed. At present, the use of the Duffing equation to detect weak unknown signals is to select a set of fixed parameters. Analysis of these two aspects to determine the parameters of the system is based on the parameter analysis of the Duffing equation oscillator. Due to the sensitivity of the Duffing equation to the initial value, the choice of parameters has a great influence on the detection effect. In response to this situation, the relationship between the parameters and initial values of the Duffing equation is analyzed. From the relationship between the parameters and the initial values, the influence of different parameters on the detection effect is analyzed to verify the superiority of the current equation parameters. First, the multichannel EEG signal is nonuniformly sampled multiempirical modal decomposition, and an effective intrinsic modal function is selected to extract the mental stress EEG characteristics. Experimental results show that the EEG signal recognition algorithm based on the Duffing equation effectively extracts EEG signal features and improves the classification accuracy of mental stress EEG signals.