基于P300的脑-机接口:视觉刺激强度对性能的影响  

P300-based brain-computer interface:Effect of stimulus intensity on performance

作　　者：马忠伟[1] 高上凯[1] 

机构地区：[1]清华大学医学院生物医学工程系,北京100084

出　　处：《清华大学学报（自然科学版）》2008年第3期415-418,共4页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60318001);清华-裕元医学科学研究基金项目

摘　　要：脑-机接口(BCI)是大脑与外部世界直接的交流通道。为了研究视觉刺激强度对基于P300的脑-机接口性能的影响,设计并实现了一种基于5个选择oddball的P300诱发电位范式的脑-机接口系统,并在此系统中研究2种不同强度下视觉刺激(高强度和低强度)下脑-机接口的信息传输率差异。9名受试者参加了实验,每位受试者在高低2种强度视觉刺激下各采集40组数据,数据在预处理后使用支持向量机进行分类,最终的目标识别率分别为84%和81%。平均波形表明在所设计的范式下2种强度视觉刺激均可以诱发出稳健的P300电位,离线分析表明高强度视觉刺激下平均信息传输率可以达到4.9 bit/min,而低强度视觉刺激下为4.5 bit/min。A brain-computer interface （BCI） is direct communication channel between the human brain and the external world. The effect of stimulus intensity on the performance of the brain-computer interface was investigated using a P300-based BCI system having a five-choice oddball paradigm. 9 subjects participated the tests with each subject completing 40 epochs with high intensity visual stimuli （HIVS） and 40 epochs with low intensity visual stimuli （LIVS）. The results show a difference of the BCI performance between the HIVS and the LIVS. The preprocessed data were classified by a support vector machines. The identification correct rates of 84% with HIVS and 81% with LIVS were acquired. The averaged waveforms from the HIVS and the LIVS show that this new paradigm could elicit evident P300 potentials. Further analyses indicate that the HIVS information transfer rate could reach 4.9 bit/min, which is higher than the 4.5 bit/min with the LIVS.

关 键 词：脑-机接口 P300 支持向量机 信息传输率 

分 类 号：R318[医药卫生—生物医学工程]