P300 Chinese input system based on Bayesian LDA

被引：48

作者：

Jin, Jing ^{[1
,2
]}

Allison, Brendan Z. ^{[1
]}

Brunner, Clemens ^{[1
]}

Wang, Bei ^{[2
]}

Wang, Xingyu ^{[2
]}

Zhang, Jianhua ^{[2
]}

Neuper, Christa ^{[1
,3
]}

Pfurtscheller, Gert ^{[1
]}

机构：

[1] Graz Univ Technol, Inst Knowledge Discovery, Lab Brain Comp Interfaces, A-8010 Graz, Austria

[2] E China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200237, Peoples R China

[3] Graz Univ, Dept Psychol, Graz, Austria

来源：

BIOMEDIZINISCHE TECHNIK | 2010年 / 55卷 / 01期

基金：

中国国家自然科学基金;

关键词：

Bayesian linear discriminant analysis; brain-computer interface; Chinese character input system; P300; particle swarm optimization; BRAIN-COMPUTER-INTERFACE; MENTAL PROSTHESIS; BCI; AMPLITUDE; ERPS;

D O I：

10.1515/BMT.2010.003

中图分类号：

R318 [生物医学工程];

学科分类号：

0831 ;

摘要：

A brain-computer interface (BCI) is a new communication channel between humans and Computers that translates brain activity into recognizable command and control signals. Attended events can evoke P300 potentials in the electroencephalogram. Hence, the P300 has been used in BC1 systems to spell, control cursors or robotic devices, and other tacks. This paper introduces a novel P300 BCI to communicate Chinese characters. To improve classification accuracy, an optimization algorithm (particle swarm optimization, PSO) is used for channel selection (i.e., identifying the best electrode configuration). The effects of different electrode configurations on classification accuracy were tested by Bayesian linear discriminant analysis offline. The offline results from I I subjects show that this new P300 BCI can effectively communicate Chinese characters and that the features extracted from the electrodes obtained by PSO yield good performance.

引用

页码：5 / 18

页数：14

共 35 条

[1] Workload assessment of computer gaming using a single-stimulus event-related potential paradigm [J].