数据区间优化对SSVEP算法性能的影响

稳态视觉诱发电位(steady-state visual evoked potential， SSVEP)响应的个体差异性较大，不同环境下各被试者脑电信号的质量有差别.针对这个问题，研究了SSVEP中数据区间的优化对CCA(典型相关分析)和ECCA(扩展典型相关分析)方法分类结果的影响.首先通过网格搜索方法找到脑电信号的最优数据区间，然后使用CCA和ECCA方法对该区间数据进行特征识别，使得识别效果得到提升.实验结果表明，同时对数据区间起始点和终点进行优化能够有效提升信息传输率(ITR)，数据区间优化后的CCA和ECCA分类平均ITRs为(61.18±27.20)bit/min和(71.37±32.24)bit/min，比使用传统的仅优化数据区间终点的方法提高了29.89%和8.3%，证明了通过数据区间优化能够提升SSVEP算法的性能.

Influence of Data Interval Optimization on SSVEP Algorithm Performance

Steady-state visual evoked potential(SSVEP)responses vary greatly among individuals， which results in the quality difference of EEG signals from subjects in different environments. So， the effect of EEG data interval on CCA(canonical correlation analysis) and ECCA(extended canonical correlation analysis) classification results is investigated. The optimal data interval of EEG signals is determined through grid search， then， the EEG features in the optimal data interval are identified by CCA and ECCA， with the recognition results improved. The results show that the information transfer rate(ITR)can be effectively improved by optimizing the starting and ending points of the data interval. The average ITRs of CCA and ECCA classification after interval optimization are (61.18±27.20)bit/min and(71.37±32.24)bit/min， which are 29.89% and 8.3% higher than that of the traditional method which only optimizes the ending point of data interval. The results proved that the performance of SSVEP algorithm can be improved by optimizing data interval.