A practical hybrid BCI speller for Chinese Character Input: Integrating an Eye Tracker into a P300-Based Spelling approach

Abstract

In this study, we proposed a practical hybrid brain-computer interface (BCI) speller to enhance the performance for Chinese character input by incorporating eye tracker into the traditional P300 event-related potential spelling paradigm. The spelling process is activated asynchronously when subjects gazes at the Start-button. Then, the interface turns to the classic row-column P300 paradigm. A Chinese character can be inputted in three steps, i) input the initial consonant, ii) input the vowel component, and iii) selecting the target character. With the combination of an eye-tracker, the proposed paradigm achieves asynchronous operation, and achieves accurate and fast-reading Chinese character input without adding additional cognitive load compared to the traditional P300 method. The improved system showed an average speed of about 1.14 sinograms per minute. The experimental results revealed that the proposed paradigm can achieve character input asynchronously with a high accuracy and speed, improving the traditional P300-based spelling paradigm.