摘要
为了实现脑一机接口通信信号的单次提取,基于视觉诱发电位在脑部不同位置的相关性应该比自发脑电信号强这一特点,采用多通道的同步平均信号作为特征,用支持向量机分类算法,对一名受试者200次选择作业的记录进行了分类.结果表明,在仅利用Cz与Pz两个通道的信号叠加平均后,取250~550ms时段的信号作为分类器的特征值,能达到97%以上的分类精度.这可为简化脑一机接口系统的设计、提高通信速率提供参考.
The communication signals should be estimated by single trial in brain-computer interface. Based on the fact that the relativity of visual evoked potentials from different sites should be stronger than those of spontaneous EEGs, using the time-lock averaged signals from multi-channels as features, 200 trials of EEG recordings evoked by target or non-target stimuli were classified by support vector machine (SVM). The results show that a classification accuracy of 97 % can be obtained only using the 250-550 ms time section of the averaged signals from channel Cz and Pz as features. It suggests that it be possible to boost communication speed and simplify the designation of BCI system in this way.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2007年第1期11-13,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(30370393)
关键词
脑-机接口
相干平均
诱发电位
单次提取
brain-computer interface
in-phase average
evoked potentials
single-trial estimation