Determining Binary Intentions Using a Steady State Visual Evoked Potential Based BCI: A Comparison of Centered and off Centered Stimulus Presentation

A brain-computer interface (BCI) is a system that can measure and convert brain activity into artificial outputs which can be used to control external devices (Wolpaw & Wolpaw, 2012). A Steady-State Visual Evoked Potential or SSVEP-based BCI that relies on electroencephalographic (EEG) signals can be helpful in understanding the binary (yes/no) intentions of people who have severe oculomotor dysfunction due to trauma or disease (Jeong-Hwan Lim et al., 2013). The present study explores the effect of foveally-centered or foveally-off-centered stimulus presentation on the binary classification of the SSVEP with the overall goal of developing a robust and easy-to-use bedside communication system. Sixteen channel EEG was measured in six subjects (5 females and 1 male, average age 46.8 ±23.4, median age 59). Participants were instructed to attend to either the left or right eye with their eyes open, while LEDs frequencies were presented at 23Hz and 31Hz simultaneously to left and right eyes. The results showed qualitative differences in average classification accuracy (i.e., which eye the person was attending to) for centered (92.9 % (± SD10%)) and off centered (95.8 %( ±SD 3.2%)) conditions; and ideal bit rates of 10.7 (± SD5) and 13.7 (±SD 5) bits/min respectively. These results suggest that taking advantage of the off-centered LED geometry may provide optimized information for binary communication system that is fast, accurate and can operate at the bedside. Further studies are planned to optimize frequency pairs, and quantify pupillary distance and LED brightness, to help standardize a bedside system