P300-based brain-neuronal computer interaction for spelling applications

Abstract

A brain-neuronal computer interaction system can provide a communication channel for severely disabled people or a supplementary control channel for able-bodied subjects. In this paper, a physiological hybrid P300-based speller that uses a modified stimulus presentation paradigm-the half checkerboard paradigm (HCBP)-is evaluated. The speller uses electrooculography and electroencephalography signals for selecting alphanumeric characters or commands arranged in an 8 × 9 matrix. In this study a group of subjects, who can voluntarily gaze at a target, used the checkerboard paradigm- and HCBP-based spellers in a counterbalanced fashion for comparing their performances under a series of online tests. A 16-character-long text was spelled by each subject, while a 13-character-long text was used for calibrating the system. By using the HCBP, the time required for spelling one character is reduced, resulting in higher information transfer rates. The results suggest that the HCBP has the potential to provide a more effective P300 paradigm with a major importance for people with neuromuscular diseases and also for healthy people as a supplementary communication channel.