Brain-Computer Interfaceのためのカーソル制御: 利激呈示方法の検討

摘要

われわれは脳波を用いた意思伝達システムの開発を目指して研究を行っている．過去の研究においてディスプレイ中央に4種類の矢印画像をランダムに呈示する刺激を用いて意思伝達実験を行った．結果，一部の被験者では刺激に対しての脳波の応答にほとんど変化がなく，刺激方法の改善が必要であることが分かった．そこで，本研究では，刺激時の画像呈示を従来の中央画像呈示の他に，画像定位置呈示，画像背景色変化の2種類を試し，最適な画像呈示方法を検討した．実験は8名の健常者（男性21-26歳）を対象に行った．その結果から，画像呈示の違いによる有意差を検定すると，画像背景色変化による画像呈示は他の画像呈示と比べてP300振幅差，周波数帯域1-2Hz，2-3Hz，3-4Hz，4-5Hzのパワー差が有意に大きくなった．また，P300振幅値によるtarget弁別成功率は画像背景色変化が最大となり，68．7％となった．周波数解析によるtarget弁別成功率は画像背景色変化が最大となり88．3％となった．以上の結果から，画像背景色変化による刺激が最適であることが示唆された．%We have reported the EEG based brain-computer interface (BCI) which employed four kinds of arrow images as visual stimuli on the center of the display. However, in some cases, it was hard to discriminate between EEG responses because of their similarity. Therefore, in this report we investigated on which of stimulation methods, such as central area presentation (CAP, conventional), peripheral area presentation (PAP) or background color change (BCC), was effective to discriminate between EEG responses. Eight healthy men, aged 21-26 years old, were participated and examined. In the time domain comparison, the amplitude difference of P300 by the BCC method was significantly larger than those by the other methods. Furthermore, the maximum accuracy (68.7%) of distinction between the target response and the non-target ones was obtained by the BCC method. In the frequency domain comparison, the band power differences of 1-2 Hz, 2-3 Hz, 3-4 Hz and 4-5 Hz by the BCC method were also significantly larger than those by the other methods. Moreover, the maximum accuracy (88.3%) of distinction between the target response and the non-target ones was obtained by the BCC method. It was suggested that the BCC is effective stimulation method for our developing BCI.