不同编码分组调制的VEP脑机接口研究

Study on VEP brain-computer interface with different coding grouping modulation

在现有的各种脑机接口实现范例中,基于c-VEP的脑机接口取得了最高的信息传输率,但这种脑机接口系统必须在刺激目标数与检测精度和检测速度之间进行折衷。要增加刺激目标数并保证识别率高,必须增加码长,然而,增加码长会导致识别一个目标所需的时间增加,降低信息传输率。针对这个问题,提出了使用不同编码分组调制刺激目标的方法。基于该方法,采用Golay码和近完美码各调制16个刺激目标,目标识别的方法是先获取两组目标的参考模板,分别对两个模板做移位得到所有目标的模板,再运用模板匹配法对目标进行识别。7位受试者在一个刺激周期的平均分类准确率高达92.34%,研究结果表明该方法增加了刺激目标数,同时提高了识别率,缩短了检测时间。

In current various implementation paradigms of brain-computer interface（BCI）,the c-VEP based BCI achieved the highest information transfer rate,but this BCI system must have a compromise among the stimulation target quantity,detection accuracy and detection speed. To increase the stimulation target quantity and ensure the high recognition rate,the code length must be increased,but it will lead to time increasing to recognize a target,and information transfer rate reducing. To solve this problem,a stimulation target method applied with different coding grouping modulations is proposed. Based on this method,the Golay code and nearly perfect code are used to modulate 16 stimulation targets respectively. The target identification method is used to obtain the reference templates of two target groups. The two templates are shifted to obtain all target templates,and then the template matching method is used to recognize the target. The average classification accuracy of a stimulation period for 7 subjects can reach up to 92.34%. The research results show that the method can increase the stimulation target quantity,improve the recognition rate,and shorten the detection time.