English

应用科学学报 ›› 2006, Vol. 24 ›› Issue (4): 359-362.

• 论文 • 上一篇    下一篇

脑电近似熵分析的思维分类识别

江朝晖1, 高翠云2, 冯焕清1   

  1. 1. 中国科学技术大学电子科学与技术系, 安徽合肥 230026;
    2. 安徽建筑工业学院计算机与信息工程系, 安徽合肥 230022
  • 收稿日期:2005-01-15 修回日期:2005-04-18 出版日期:2006-07-31 发布日期:2006-07-31
  • 作者简介:江朝晖,博士,研究方向:信号采集、信息处理,E-mail:jiangzkh@ustc.edu.cn;冯焕清,教授,博导,研究方向:信号采集、信息处理,E-mail:hqfeng@ustc.edu.cn
  • 基金资助:
    中国科学技术大学校青年基金资助项目(KB2508)

Mental Task Classification Based on Approximate Entropy of EEG

JIANG Zhao-hui1, GAO Cui-yun2, FENG Huan-qing1   

  1. 1. Department of ElectronicScience & Technology, University of Science and Technology of China, Hefei 230026, China;
    2. Department of Computer & Information Engineering, Anhui Institute of Architecture & Industry, Hefei 230022, China
  • Received:2005-01-15 Revised:2005-04-18 Online:2006-07-31 Published:2006-07-31

PDF

118

被引次数

9

摘要： 以近似熵为指标,分别计算6导脑电信号的复杂度,作为神经网络分类器的输入特征向量,对5种思维任务进行分类识别.测试结果表明,针对5次练习的均值而言,平均识别率为73%,说明利用多导脑电的复杂性进行思维辨识是可行的.

关键词: 思维作业, 近似熵, 脑电, 神经网络

Abstract: Mental activity is closely related to the complexity of electroencephalogram (EEG) in different areas of brain. In order to classify five different kinds of mental tasks, approximate entropy (ApEn), an index of complexity, is calculated from six channels EEG and serve as input feature vector to a neural network.Testing results show that the average classification accuracy is 73% obtained from the average of 5 trials, indicating feasibility of classifying different mental tasks based on complexity of multi-channel EEG.The introduced technique is useful in the research of human machine interface.

Key words: electroencephalogram (EEG), mental task, approximate entropy (ApEn), neural network

中图分类号: