A noninvasive brain computer interface using visually-induced near-infrared spectroscopy responses

• Published in: Neuroscience letters Vol. 580; pp. 22 - 26
• Main Authors: Chen, Cheng-Hsuan, Ho, Ming-Shan, Shyu, Kuo-Kai, Hsu, Kou-Cheng, Wang, Kuo-Wei, Lee, Po-Lei
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier Ireland Ltd 19.09.2014
• Subjects: Adult; Brain - physiology; Brain computer interface; Brain-Computer Interfaces; Female; Flickering sequence; Hemodynamic response; Humans; Male; Near infrared spectroscopy; Photic Stimulation; Spectroscopy, Near-Infrared; Visual Perception; Visual response; Young Adult; Near infrared spectroscopy; Brain computer interface; Visual response; Flickering sequence; Hemodynamic response
• ISSN: 0304-3940; 1872-7972
• DOI: 10.1016/j.neulet.2014.07.042

•We utilized visually-induced NIRS signals to design BCI.•A method for mutual independencies among different visual stimuli was developed.•User's gazed target can be discerned though a simple averaging process.•The proposed NIRS-based BCI achieved detection accuracy higher than 90%. Visually-induced near-infrared spectroscopy (NIRS) response was utilized to design a brain computer interface (BCI) system. Four circular checkerboards driven by distinct flickering sequences were displayed on a LCD screen as visual stimuli to induce subjects’ NIRS responses. Each flickering sequence was a concatenated sequence of alternative flickering segments and resting segments. The flickering segment was designed with fixed duration of 3s whereas the resting segment was chosen randomly within 15–20s to create the mutual independencies among different flickering sequences. Six subjects were recruited in this study and subjects were requested to gaze at the four visual stimuli one-after-one in a random order. Since visual responses in human brain are time-locked to the onsets of visual stimuli and the flicker sequences of distinct visual stimuli were designed mutually independent, the NIRS responses induced by user's gazed targets can be discerned from non-gazed targets by applying a simple averaging process. The accuracies for the six subjects were higher than 90% after 10 or more epochs being averaged.