点击次数：

论文类型： 期刊论文

发表时间： 2017-12-01

发表刊物： MEDICAL & BIOLOGICAL ENGINEERING & COMPUTING

收录刊物： Scopus、SCIE、EI、PubMed、SSCI

卷号： 55

期号： 12

页面范围： 2245-2256

ISSN号： 0140-0118

关键字： Brain-computer interface; Event-related potential; P300; Paradigm

摘要： Recently, many studies have been focusing on optimizing the stimulus of an event-related potential (ERP)-based brain-computer interface (BCI). However, little is known about the effectiveness when increasing the stimulus unpredictability. We investigated a new stimulus type of varied geometric pattern where both complexity and unpredictability of the stimulus are increased. The proposed and classical paradigms were compared in within-subject experiments with 16 healthy participants. Results showed that the BCI performance was significantly improved for the proposed paradigm, with an average online written symbol rate increasing by 138% comparing with that of the classical paradigm. Amplitudes of primary ERP components, such as N1, P2a, P2b, N2, were also found to be significantly enhanced with the proposed paradigm. In this paper, a novel ERP BCI paradigm with a new stimulus type of varied geometric pattern is proposed. By jointly increasing the complexity and unpredictability of the stimulus, the performance of an ERP BCI could be considerably improved.