色彩の効果によるBCIにおける意思判読能力の改善

Abstract

Pictures of four kinds of filled shapes (circle, equilateral triangle, square and regular pentagon) colored by red, green, blue and corresponding grays were prepared. As a typical brain-computer interface (BCI) paradigm using P300, pictures were presented in an oddball paradigm. Target stimulus was selected from four filled shapes of specific color and non-target stimuli were the other filled shapes of the same color. Twelve experiments (twice for each color) were implemented. Subjects counted whenever targets were detected. For each color, the P300 peak amplitude for target minus that for non-target which relate to signal-to noise ratio (S/N) were calculated. S/N for red and blue were larger than those for the corresponding grays. Correct rates of BCI calculated by the statistical method for red, green and blue were higher than those for the corresponding grays. Peak latency for red was shorter than that for the corresponding gray. These results showed the probability of improvement of noise proof property, correct rate and rapidity of BCI by using effects of colors.