|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ИНХС РАН |
||
The P300 BCI is more than an oddball and more than a P300 BCIтезисы доклада
- Авторы: Shishkin S.L., Ganin I.P., Kaplan A.Y.
- Сборник: Society for Psychophysiological Research. 49th Annual Meeting
- Том: 46
- Тезисы
- Год издания: 2009
- Место издания: Berlin
- Первая страница: 47
- Последняя страница: 47
- Аннотация: In Farwell-Donchin (P300) BCI, attention acts as a factor mediating operator's control over a computer. However, attention can be captured by various distractors. In some context, a command may be unwanted or even dangerous (e.g., deleting an important file). Is it easy for a user to prevent triggering such commands occasionally? Moreover, can the expected loss associated with a command make it act as an emotional distractor, so that the attention will preferably trigger this command? 11 normal subjects were trained to use BCI (3x3 matrix with simple pictorial symbols) with ~0.7-0.8 accuracy, and then played a game: correct "entering" the target symbol gave one point; entering the symbol which was the previous trial's target (PTT, or "dangerous symbol") led to 5 point loss in "dangerous" condition (DC, 21-42 trials) but did not change the score in control condition (CC, same n of trials); other errors did not change the score. In each trial the target position was adjacent to the target position of the previous trial. All but one subject at least once entered a PTT in the DC. Mean number of entering PTT was 2.0 in DC and 1.1 in CC; the difference was not significant (p=0.16). Among the cases of high number of PTT in DC were a student who had an exam 15 min after the experiment, and a person who felt herself especially nervous during a block where she made all such errors. Though we cannot be certain yet that harmful commands may indeed be activated preferably, it is clear that blocking their random activation may be not a simple task for a BCI user.
- Добавил в систему: Ганин Илья Петрович