基于等分符号化熵的情感脑电信号分析

Analysis of emotional EEG signal based on equal symbolic entropy

如何提取有效的特征一直是情感研究的一个热点。结合脑电信号非线性方法中排列熵计算效率高、样本熵计算准确率高的优点,提出了等分符号化熵(ESE)算法,并试图验证这种新的特征在情感脑电分析中的有效性。该算法在相空间重构前对信号进行等概率符号化处理,用符号矢量的相等计算重构分量比例。仿真结果显示,ESE算法在logistic映射中计算效率与计算准确度均有良好的表现。将ESE算法用于情感脑电信号的分析,结果表明部分脑区可以有效地区分正负性情感,表明此算法可有效地衡量不同类型情感的特征。

It＇s a hotspot in emotional research that how to extract an effective feature. Combining the advantages of the sample entropy algorithm with high accuracy and the permutation entropy algorithm with high accuracy in the nonlinear method of EEG,this paper proposed equal symbolic entropy（ ESE） and tried to verify the effectiveness of this new feature in emotional EEG analysis. In this algorithm,the signals were reconstructed with equal probability before the reconstruction of phase space,and the proportion of reconstructed components was calculated by using the equal vector of symbols. The simulation results show that the ESE algorithm has a better performance in the computational efficiency and accuracy in logistic map. It was used to analyze the emotion signals and the results show that some brain regions can effectively distinguish between positive and negative emotions. This feature can effectively reflect the different types of emotions.