摘要
目的本文设计了一种听觉刺激方法,即利用产生的左右手想象听觉脑电信号推断脑卒中患者的意图,控制手部康复机器人,以满足脑卒中患者的康复需求。方法实验对象为2名受试者,各进行80次实验。根据脑电信号的非平稳非线性特性,利用Hilbert-Huang变换法分析想象脑电信号的频率特性,以信号能量值为特征向量,利用支持向量机(support vector machine,SVM)算法识别左右手的动作模式。结果在与大脑运动感觉区域相关的C3、C4电极上,听觉脑电信号的事件相关去同步与事件相关同步现象明显,识别效果达80%以上。结论该听觉刺激方法具有较好的识别效果,可用于患有视觉障碍的脑卒中患者的康复训练。
Objective A auditory stimulus method was proposed in this paper. The auditory EEG of imaging left-right hand movement was used to infer the intention of the stroke patients. The aim of this work was to control the hand rehabilitation robot to satisfy the rehabilitation requirement of the stroke patients. Methods The experimental objects were two individuals and each one experimented 80 times. According to the non- stationary and non-linear feature of the EEG signals, Hilbert-Huang transform was applied to extract the frequency feature. The energy of the signal was used as the feature to recognize the movement of the left-right hand by using the SVM algorithm. Results The signals from C3 and C4 on the movement feeling regions of the brain showed that the event-related desynchronization and the event-related synchronization of the auditory EEG were significant and the recognition rates were above 80%. Conclusions This auditory stimulus method has preferable recognition that can be used for the rehabilitation training of the stoke patients with visual impairment.
出处
《北京生物医学工程》
2015年第6期607-611,共5页
Beijing Biomedical Engineering
基金
国家自然科学基金(51275101)资助
