An Assistive Communication Brain-Computer Interface for Chinese Text Input

Abstract

The performance of assistive communication brain-computer interfaces has been studied mostly for languages with alphabetic script. The viability of using such systems for languages with other types of script, such as Chinese, which has a logographic script, is currently poorly understood. Here, a performance analysis of the P300 Speller is presented for Chinese text input. The performance of six distinct paradigms, based on the established Row/Column (RC) and Single Character (SC) spellers, are tested and compared for 30 able-bodied, native Chinese readers. In terms of accuracy per trial, the optimal paradigm is based on the SC speller: 63.3% of participants were able to achieve 80% or better classification accuracy for 15 trials. However, because the RC speller has shorter trial duration than the SC speller, the optimal paradigm in terms of communication rate is a variant of the RC speller in which stimuli are intensified by changing background color. A communication rate of 14.5 bits per minute was attained using this paradigm. For a lexicon of ~11,000 Chinese characters, this corresponds to a projected mean input rate of ~1.1 characters per minute.