Comparison of Skin Deformation between Passive and Active Movements by Reconstruction Error of Principal Component Analysis

Abstract

Since skin deformation can be obtained during passive movement by therapists, even stroke patients in acute who have difficulty with active joint movement, comparing skin deformations between passive and active movement and extracting the amount of active movement may be useful in assessing motor function. In this study, skin deformation was compared between passive and active movements based on the reconstruction error of a principal component analysis (PCA) model trained on passive movement data. Experimental results of measuring skin deformation of the upper arm during flexion and extension of the elbow joint showed that the PCA reconstruction error increased/decreased proportionally to the joint movement during active movement, and the average reconstruction error of active movement was larger than that of passive movement.