基于稳态视觉诱发电位的脑电信号分类算法比较  

Comparison of Electroencepha Logram Signal Classification Algorithms Based on Steady State Visual Evoked Potential

作　　者：李新 吴迎年[1,2] 李睿 LI Xin;WU Ying-nian;LI Rui(School of Automation, Beijing Information Science and Technology University, Beijing 100192, China;Beijing International Science and Technology Cooperation Base for Intelligent Perception and Control of High End Equipment, Beijing 100192, China)

机构地区：[1]北京信息科技大学自动化学院,北京100192 [2]高端装备智能感知与控制北京市国际科技合作基地,北京100192

出　　处：《科学技术与工程》2021年第19期8106-8112,共7页Science Technology and Engineering

基　　金：促进高校内涵发展-应急攻关项目(5212010976);2019科技部高端专家引进项目(G20190201031);北京信息科技大学2019年教改重点资助项目(2019JGZD02);2019年北京高等教育本科教学改革创新项目(5112010813);北京信息科技大学促进高校内涵发展科研平台师资补充经费(5112011144)。

摘　　要：基于稳态视觉诱发电位(steady state visual evoked potential,SSVEP)的脑-机接口(brain computer interface,BCI)系统具有分类准确率高、用户不用长时间训练等优点而广受关注。如何高效地对SSVEP信号频率识别而实现更好的分类效果是SSVEP-BCI的核心问题。采用滤波器组典型相关分析(filter bank canonical correlation analysis,FBCCA)与任务相关成分分析(task-related component analysis,TRCA)对SSVEP信号分类比较研究,探讨了两种方法在数据长度、子带数以及通道数对SSVEP信号分类效果的影响。35位被试者的数据表明:在数据长度小、时间短的情况下,TRCA具有更高的分类准确率,且子带数设置为5时,分类准确率达到最大。通道数越多分类准确率越高,但是通道个数较少时,TRCA分类效果明显优于FBCCA。研究为SSVEP脑电数据有效性分析以及提高基于SSVEP的脑电信号分类准确率提供了新的思路。The steady-state visual evoked potentials(SSVEP)-based brain-computer interface(BCI)has attracted much attention due to its advantages such as high classification accuracy and little user training.How to effectively identify the SSVEP signal frequency and achieve better classification effect is the key issue of SSVEP-BCI.In this paper,filter bank canonical correlation analysis and task-related component analysis are used to compare SSVEP signal classification,and the effects of the two methods on the SSVEP signal classification effect are discussed in terms of data length,sub-band numbers and channel numbers.The data of 35 subjects show that the TRCA has higher classification accuracy in the case of small data length and short time,and the classification accuracy reaches the maximum when the number of sub-band is set to 5.The more channels,the higher the classification accuracy,however,when the number of channel is fewer,the classification performance of TRCA is better than FBCCA.The research provides a new idea to analyze the validity of the EEG data and improve the classification accuracy of EEG signal based on SSVEP.

关 键 词：稳态视觉诱发电位 滤波器组典型相关分析 任务相关成分分析 分类准确率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]