非負値行列因子分解を用いた筋シナジー解析における筋電チャンネル数削減手法

抄録

Muscle synergy is used for analysis of body movements or as input to a physical assistive device, because they can effectively express features of body movements with a smaller amount of information than the raw signals of surface EMG. Besides, if similar muscle synergy analysis can be achieved with a smaller number of myoelectric channels, it can improve the efficiency of the measurement and to reduce the computational cost. Thus, the autors have proposed a method to reduce the number of myoelectric channels in the muscle synergy analysis by selecting channels according to their magnitude of influence on the temporal patterns of muscle synergy. In this paper, first, the detailed algorithm of the proposed method to evaluate the influence of each myoelectric channel on the temporal patterns of muscle synergy and to select proper channels is presented. Second, as an example to examine the validity of the algorithm, results of muscle synergy analysis of an object lifting motion is described. Then, the presented algorithms is applied to the example to study the relationship between the number of the adopted EMG channels and the results of the revised muscle synergy analysis. Thereby, it is shown that the channel can be selected properly to the purpose.