基于HOC-SVM的运动状态下脑电的特征提取与分类

研究人脑在不同运动状态下的脑电信息,不仅能够揭示出各种运动状态对于大脑活动的影响,也是工程技术人员设计脑-机接口与神经修复系统的关键技术之一。文章根据脑电信号的μ节律变化,首次将表征时间序列摆动特性的高阶过零分析(Higher Order Crossing,HOC)方法运用于运动状态下的脑电信号的特征提取并结合支持向量机(Support Vector Machine,SVM)对输入的高阶过零特征量进行了有效的分类。将该方法提取的特征量与基于统计学的特征量分别用SVM进行分类,结果表明本方的识别率明显高于基于统计学特征量的方法。说明基于HOC-SVM方法在脑电信号的特征提取与分类中有较强的可行性和实用性。

A Better Method of Feature Extraction and Classification of Electroencephalography (EEG) Signals in Motion State with Higher-Order Crossing (HOC) and Support Vector Machine (SVM)

Sections 1 through 3 of the full paper explain the better method mentioned in the title,which we believe is new and better than that of the statistically based feature extraction.Their core consists of:(1) according to changes in the brain μ-rhythm of EEG,we employ for the first time the HOC method to extract the features of EEG signals in motion state;(2) we use the SVM to effectively classify the EEG features extracted with the HOC method;(3) we compare the features extracted with our new method with those extracted with the statistically based method.The comparison results,given in Tables 2 and 3,and their analysis show preliminarily that the pattern recognition rate of our new method is much higher than that of the statistically based feature extraction method.