S21-1. A new analysis method using surface electromyography to assess finger function in patients with severe stroke

Abstract

We have conducted our research into kinesthetic illusions induced by visual stimuli (KiNvis), which are sensations of being in motion that result from watching artificial images of the body part moving. Our previous studies revealed characteristic neural networks related to KiNvis; since then, we have initiated clinical studies adapting KiNvis in patients with stroke. In patients with severe stroke, it is often difficult to measure joint angles, because voluntary movement does not occur or simultaneous contraction of the agonist and antagonist muscles prevent controlled voluntary joint exercise. Therefore, we have developed an assessment method for finger function in these patients using surface electromyography (EMG). Our method aimed to assess “reciprocal muscle activity” during repetitive exercise. Hence, we calculated cross correlation coefficients between EMG signals recorded during reciprocal muscle activity and pseudo model signals during ideal reciprocal muscle contraction. During reciprocal muscle activity, the peak value of cross correlation coefficient calculated using this method was higher than at rest or during sustained muscle activity. Accordingly, we consider that even in patients with severe stroke in whom changes in motor function cannot be detected with the variable that analyzes EMG signals quantitatively, it may be possible to assess finger function alterations using the analysis method of the present study.