计算机工程与应用 ›› 2019, Vol. 55 ›› Issue (22): 239-244.DOI: 10.3778/j.issn.1002-8331.1808-0052

• 工程与应用 • 上一篇    下一篇

基于功能连接信息熵的精神分裂症EEG分类研究

李佩珍,王彬,牛焱,田程,相洁   

  1. 太原理工大学 信息与计算机学院,太原 030024
  • 出版日期:2019-11-15 发布日期:2019-11-13

Research on EEG Classification of Schizophrenia Based on Information Entropy of Functional Connection

LI Peizhen, WANG Bin, NIU Yan, TIAN Cheng, XIANG Jie   

  1. College of Information and Computer, Taiyuan University of Technology, Taiyuan 030024, China
  • Online:2019-11-15 Published:2019-11-13

摘要: 为了提高精神分裂症的有效诊断,利用网络功能连接信息熵的方法对51例精神分裂症患者和56例年龄匹配的正常人的脑电信号(Electroencephalogram,EEG)进行了分类。通过采用分频技术、相位同步分析方法、信息熵方法、支持向量机(Support Vector Machine,SVM)分类方法,大幅提高了分类准确率(98.13%),实现了对精神分裂症的有效诊断。该分类方法主要涉及两阶段:利用分频技术和相位同步分析方法,获得各频段的脑电信号在各个时间点的功能连接矩阵;基于整个时间域上的功能连接计算各频段的信息熵,并将其分别作为功能脑网络的分类特征训练SVM分类器,进而对两组被试分类。分类结果表明,该方法大幅提高了精神分裂症检测的准确率。

关键词: 精神分裂症, 脑电信号(EEG), 相位同步, 功能连接信息熵, SVM分类, 功能脑网络

Abstract: In order to improve the effective diagnosis of schizophrenia, this paper uses the method of network functional connection information entropy to classify the Electroencephalogram(EEG) of 51 patients with schizophrenia and 56 age-matched normal subjects. This paper achieves an effective diagnosis of schizophrenia by using the methods of frequency division, phase synchronization analysis, information entropy and Support Vector Machine(SVM), and greatly improves the classification accuracy. The classification method mainly involves two stages. First, the frequency division technique and the phase synchronization analysis method are used to obtain the functional connection matrix of the EEG signals in each frequency band at each time point. Second, the information entropy of each frequency band is calculated based on the functional connections over the entire time domain. And, the information entropy of functional connectivity is used as the classification feature of the functional brain network to train the SVM classifier, then the two groups of subjects are classified. The classification results show that the method proposed in this paper greatly improves the detection accuracy of schizophrenia.

Key words: schizophrenia, Electroencephalogram(EEG), phase synchronization, information entropy of functional connection, SVM classification, functional brain network