This project aims to develop a brain-computer interface based speller that would benefit disabled users or elderly people.
The primary purpose of this work is to develop the Thai language-based P300-BCI visual speller as an assistive technology option for disabled native Thai speakers. An experiment of a checkerboard paradigm with a 8 x 9 matrix layout containing 44 Thai alphabets, 16 vowels and 7 numbers was conducted. Results show that RegLDA with xDAWN outperformed other models on P300 speller performance, achieving cross validation accuracy of 93% and online accuracy of 76%. AIT Brain Lab will continue this project by employing hybrid P300-SSVEP for further improvements in speed and accuracy.
