Estimation of Functional Independence Measure Motor Score Based on a Trunk Control Model

Abstract

The functional independence measure (FIM) is a metric of activities of daily living in stroke patients, but its accurate assessment requires experienced nurses. Therefore, we propose a system for estimating FIM scores easily at home based on angular velocities measured from specific bodily movements, performed only once, using gyro sensors. Here, we focus on the FIM motor score (mFIM) and determine whether “rising from a bed” or the “sit-to-stand movement” is the optimal specific bodily movement. We constructed a model considering the differences in convergence speeds between the movements of healthy people and those of stroke patients. We estimated the mFIM by performing regression analysis with the parameters utilized in the developed model as explanatory variables and mFIMs assessed by experienced nurses as objective variables. In our experiments, the mean absolute error of the angular velocity in the spine direction for rising from a bed was the smallest at 9.53.