基于HHT运动想象脑电模式识别研究

脑机接口是一种变革性的人机交互, 其中基于运动想象(Motor imagery, MI)脑电的脑机接口是一类非常重要的脑机交互. 本文旨在探索有效的运动想象脑电特征模式提取方法. 采用在时域、频域同时具有很高分辨率的希尔伯特--黄变换(Hilbert-Huang transform, HHT),进而提取自回归(Auto regressive, AR)模型参数并计算运动想象脑电平均瞬时能量,从而构造特征向量, 最后利用能较好地适应运动想象脑电单次试验分类的支持向量机(Support vector machine, SVM)进行分类. 结果表明在Trial的5.5～7.5s期间, HHT特征提取方法平均分类正确率为81.08%, 具有良好的适应性;最高分类正确率为87.86%, 优于传统的小波变换特征提取方法和未经HHT的特征提取方法;在Trial的8～9s期间, HHT特征提取方法显著优于后两种特征提取方法. 本研究证实了HHT对运动想象脑电这一非平稳非线性信号具有很好的特征提取能力, 也再次验证了运动想象事件相关去同步(Event-related desynchronization, ERD)现象, 同时也表明运动想象脑电的脑--机交互系统性能与被试想象心理活动的质量密切相关. 本文可望为基于运动想象脑电的在线实时脑机交互控制系统的研究打下坚实的基础.

Identification of EEG Induced by Motor Imagery Based on Hilbert-Huang Transform

Brain-computer interface is a revolutionary human-computer interaction. The brain-computer interface based on electroencephalogram (EEG） induced by motor imagery (MI) is a very important kind of brain-computer interface. The purpose of this paper is to explore the effective features extraction method for EEG induced by motor imagery. Hilbert-Huang transform (HHT) is used, which has a high resolution both in time domain and frequency domain. Auto regressive (AR) parameters are then extracted and the average instantaneous energy of motor imagery is calculated. Thus structural feature vector is constructed. Finally, support vector machine (SVM) is used for classification of EEG induced by motor imagery. The results show that for the 5.5 to 7.5 seconds of the trial, the average classification accuracy of HHT feature extraction method is 81.08%, and thus this method has a good adaptability.　Moreover, the highest classification accuracy of 87.86% is achieved by HHT which is superior to the feature extraction methods using the traditional wavelet transform and without HTT. For the 8 to 9 seconds of the trial, HHT feature extraction method is also significantly better than other two feature extraction methods. This study confirms that HHT has good feature extraction ability for EEG induced by motor imagery which is nonstationary and nonlinear signal. It also confirms the event-related desynchronization (ERD) phenomenon of motor imagery. It is shown that the performance of brain-computer interaction system based on EEG induced by motor imagery is closely related to the performance of the subject's imagination mental activity. This paper can lay a solid foundation for research of online real-time brain-computer interaction control system based on motor imagery.