Clinically Relevant Automatic Assessment of Upper-limb Motor Function Impairment

Abstract

Assessment and evaluation of motor function in stroke patients is important in the neurorehabilitation interventions. In this paper, I proposed a system for automatically evaluating the motor function of stroke patients through the sensor fusion (depth camera sensor and force sensing resistor) and algorithm implementation. This research aims to overcome the limitations of conventional FMA and existing previous automated FMA related studies. In particular, two different algorithms (rule-based binary logic algorithm and fuzzy logic algorithm) were proposed to verify the feasibility and applicability of the proposed automated FMA system. Clinical trials with 51 stroke patients were performed for the system validation. The proposed system shows high FMA score classification accuracy (more than 90% agreement) through rule-based binary logic algorithm. The calculated continuous FMA also showed a high correlation (Pearson's correlation r = 0.923). The proposed automated FMA system can be applied to robot-aided therapy and remote / home rehabilitation in near future.