Home // COGNITIVE 2012, The Fourth International Conference on Advanced Cognitive Technologies and Applications // View article
Authors:
Jeng-Ren Duann
Tien-Fu Chang
Yung-Jiun Lin
Jin-Chern Chiou
Keywords: sensorimotor rhythm (SMR); motor imagery; EEG; brain-computer interface (BCI); neurorehabilitation; event-related desynchronization (ERD); independent component analysis (ICA)
Abstract:
This study tested the robustness of the sensorimotor rhythm (SMR) features induced by motor imagery task and compared the result to those of other motor tasks, such as motor execution and observation. Thirteen subjects participated in the study and performed 5 runs of motor tasks with three different motor conditions (imagery, execution, and observation). Each run consisted of 15 motor task trials (5 for each motor condition), cued in a random sequence. 64-ch EEG was recorded while subjects performed the motor tasks. The separate runs of EEG data were concatenated and processed using independent component analysis (ICA) to separate the motor components from other brain or non-brain sources (including artifacts). Equivalent independent components from different subjects were selected using a K-means clustering method based on the features summarized from dipole location, time-frequency response, as well as scalp map. Finally, the average alpha (8-13 Hz) power changes were computed according to the motor conditions. The significance level of the event-related desynchronization (ERD) as compared to the baseline prior to the motor cue onset and the duration with significant ERD were compared across different motor conditions. Our result showed that no significant difference in terms of onset of motor induced ERD among the three motor conditions (around 370 ms after motor cue onset). However, the level of ERD was most pronounced for the motor execution condition (around -8 dB from the baseline). Both motor imagery and motor observation had similar level of the motor induced ERD. For the onset duration of the motor induced ERD, motor execution also showed the longest duration of alpha power decrease as compared to two other motor conditions. Both motor imagery and motor observation had much shorter duration of alpha power decrease; however, motor imagery elicited slightly longer ERD. As a result, the SMR-based BCI-controlled neurorehabilitation using a motor imagery should be quite challenging, if not impossible, for extracting SMR feature in real time given such a limited duration of the alpha ERD induced by the motor imagery.
Pages: 98 to 100
Copyright: Copyright (c) IARIA, 2012
Publication date: July 22, 2012
Published in: conference
ISSN: 2308-4197
ISBN: 978-1-61208-218-9
Location: Nice, France
Dates: from July 22, 2012 to July 27, 2012