文献类型：会议论文

英文题名：A Comparison Between Two Motion-Onset Visual BCI Patterns: Diffusion vs Contraction

作者：Huang, Minqiang[1];Zhang, Hanhan[1];Jin, Jing[1];Zhang, Yu[1];Wang, Xingyu[1]

机构：[1]East China Univ Sci & Technol, Minist Educ, Key Lab Adv Control & Optimizat Chem Proc, Shanghai, Peoples R China

会议论文集：14th International Symposium on Neural Networks (ISNN)

会议日期：JUN 21-26, 2017

会议地点：JAPAN

语种：英文

外文关键词：ERPs; Visual; Motion; P300; BCI; Classification accuracy; Diffusion; Contraction

摘要：Two motion-onset visual patterns are proposed in this paper. The difference between two patterns is the motion pattern of stimuli. One motion pattern is dot diffusion and another motion pattern is dot contraction. The performances of two patterns in terms of classification accuracy, comfort level and interference level (feedback from the participants) are compared with each other. Although no significant difference can be found at the online accuracy (p = 0.0861) and comfort level (p = 0.7804), the single-trial accuracy of the contraction pattern is significantly higher than the diffusion pattern (p = 0.0084) and the interference level is significantly lower than the diffusion pattern (p = 0.0243). The results demonstrate that the contraction pattern can decrease the interference from other stimuli and improve the classification accuracy.