| 104B5 |
|---|
| Pattern Recognition of EEG in Right or Left Hand Motion Imagination |
| AUTHORS |
| Katsuhiro Inoue, Kazumasa Ishii, Kouta Sugioka, Kousuke Kumamaru (Kyushu Institute of Technology) Gert Pfurtscheller (Graz University of Technology) |
| ABSTRACT |
| In this paper, the pattern recognition of EEG in right or left hand motion imagination based on autoregressive (AR) model was tried. The EEG wave pattern changes along with the time passage from stimulation. Therefore, it is necessary to decide the optimal interval in which AR parameter is estimated. We adopted the 112 kinds of intervals (7 kinds of processing length and 13-19 kinds of start points). The subject experimented 4-5 sessions. Sessions were divided into the training session and the test session. AR parameters were estimated on each interval and were used were used in pattern recognition in the training session. The optimal interval was decided the interval of high recognition accuracy in the training session. The optimal interval and AR parameters were used in pattern recognition of the test session data. The recognition accuracies were compared among the test sessions. Finally, the optimal decision methods of AR parameters were discussed. |
Back to Program
Time Table
Top of This Site
Annual Conference 2002