中文摘要：

目的为有效提取稳态视觉诱发脑机接口（SSVEP-based brain-computer interface）中的脑电特征，提出一种基于独立成分分析（independent component analysis，ICA）与希尔伯特黄变换（Hilbert-Huang transform，HHT）的特征提取方法。方法对采集得到的脑电信号进行带通滤波，得到预处理的脑电信号，将滤波后的脑电信号作为ICA的输入，经过ICA实现独立成分的快速获取。引入HHT对独立成分进行经验模态分解（EMD），分解获取固有模态函数（intrinsic mode function，IMF），通过对IMF的频域分析，即可提取出特征。将ICA和HHT法同WT法、ICA法以及HHT法等常用的特征提取方法在频域、功率谱估计、在时间消耗等多方面进行比对分析。结果频域分析和功率谱估计中，本文提出的方法明显优于WT法和ICA法，略优于HHT法。时间消耗方面，本文提出的方法略优于HHT法。结论基于ICA和HHT的特征提取方法在稳态视觉诱发脑机接口的特征提取中是可行的，并有效去除了脑电信号中的噪声。

英文摘要：

Objective In order to extract the feature of steady-state visual evoked potential （SSVEP）- based brain-computer interface （BCI） system more efficiently, a method based on independent component analysis （ICA） and Hilbert-Huang transform （HHT） is proposed in this paper. Methods In the method, band- pass filter is applied to preprocess the electroencephalograph （EEG） of SSVEP. Then the independent components are acquired from filtered signals with ICA. Furthermore, HHT is used, and its inputs are the independent components. Thus the intrinsic mode function （IMF） needed is obtained and frequency domain analysis is applied to analyze IMF. Finally, frequency domain, power spectrum estimation and time consumption of the methods are compared. Results In frequency domain and power spectrum estimation, the method based on ICA and HHT is obviously better than WT and ICA, and also more effective than HHT. In time consumption, the proposed method is more effective than HHT. Conclusions The proposed method is feasible in feature extraction and the noise also can be removed.