Hand Motor Function Evaluation by Integrating Multi-Tasks Using Home Rehabilitation Device

Abstract

Many hemiplegic patients, especially those having paralysis at hand, need continuous rehabilitation after leaving hospitals. To support rehabilitation at home, we have developed a robotic device which provides automated rehabilitation and evaluation of hand motor function recovery. The device monitors the patient's finger movements during voluntary finger tasks. Our idea is to calculate the extent of recovery by asking a patient to perform a specific voluntary movement (task) and calculating the dissimilarity between a typical healthy person's movement and that measured by this device. In our previous studies, we only clarified the evaluation methods for a single task. This paper proposes the integration of these values into a single measurement value by using binary logistic regression. This integration provides the patient with a probability of being classified as a patient and a healthy subject. A home user should easily understand the value. The experimental results show that the integrated evaluation has higher performance over the single task evaluation.