DOI: 10.5176/2251-2179_ATAI16
Authors: Koun-Tem Sun, Tzu-wei Huang, Min-chi Chen and Yao-Chuan Li
Abstract:
This study presented a brain-computer Chinese Spelling System. This system used ERPs to analyze the brain waves, and used specific components of ERPs to find out targets. Three experiments were carried out in research. In experiment 1, the amplitude of N2P3 signals at O1 achieved significant difference between the target and non-targets. Experiment 1 achieved an 89.9{6e6090cdd558c53a8bc18225ef4499fead9160abd3419ad4f137e902b483c465} detecting accuracy. And then, in experiment 2, we changed the procedure and interface presentation to improve the execution speed. The duration of selecting an item was reduced from 24s to 5.7s. The accuracy of experiment 2 was 86.7{6e6090cdd558c53a8bc18225ef4499fead9160abd3419ad4f137e902b483c465}. Finally, we carried out experiment 3 to further improve the performance. It spent 3.6s to select an item, and achieved a 100{6e6090cdd558c53a8bc18225ef4499fead9160abd3419ad4f137e902b483c465} accuracy rate. The bit-rate in experiment 3 achieved 59.75 bit/min, which is the fastest BCI for spelling Chinese words by brain waves.
Keywords: EEG, ERPs, BCIs, Speller
Updating...