Abstract
Recently, it is focused on that research for Brain-Computer interface (BCI), which detects brain activity and controls electric devices. The technique of almost BCI systems aims at how the changes of EEG signal accompanying a certain cognitive phenomenon are detected with a sufficient precision. However, EEG signal is varied in individuals and drastically changes by the condition of a mental status or the body, even if an individual. Therefore, a fixation of the target feature changes of EEG signals makes high-precision BCI be fundamentally difficult. In this study, we attempted to search for the obvious feature of EEG changes accompanying a cognitive task, and suitable cognitive tasks for BCI control. Such BCI system has a potential to perform suitable control adapted to an individual. Template matching using a simplified fuzzy reasoning has been used for recognition of the EEG features. The antecedent parts of the fuzzy rules were utilized as the pattern template of the EEG feature and the pattern matching was performed using the fuzzy grade of each rule. The efficient feature searching was aimed by limiting the spectrum of the feature of EEG with the scheme of a fuzzy rule. Using this system, we succeeded in detecting the mu wave ERS and the occipital alpha-rhythm preponderance by a closing eyes.
