上原 信太郎

OBJECTIVES: To investigate the feasibility of robot-assisted hand movement training using a novel end-effector robot in individuals after stroke. METHODS: Eleven individuals with subacute stroke with hand motor impairment underwent robot-assisted repetitive finger flexion/extension for 20 min daily and repeated this training on 7 non-consecutive days. The robot was designed to allow the flexion and extension of the metacarpophalangeal and proximal interphalangeal joints of the index to the little fingers, and to provide assistive torque if the movement did not reach the target angle within a limited time. We assessed the co-contraction index (CCI) of the flexor digitorum superficialis and extensor digitorum muscles and assessed the active range of motion (AROM) of the index finger before and after training each day (intra-day effect). We performed clinical assessments of motor function and spasticity and evaluated the CCI and AROM before and immediately after the 7-day training (inter-day effect). RESULTS: Ten participants completed the 7-day training. For the intra-day effect, the CCI was significantly decreased immediately after training, particularly during active finger flexion, and the AROM tended to improve from the middle of the training days. For the inter-day effect, there were no significant changes in the Stroke Impairment Assessment Set for Finger Function, modified Ashworth scale, CCI, or AROM after the 7-day training. CONCLUSIONS: Repetitive finger movement training with the assistance of the novel robot improves muscle activation patterns, reducing co-activation between the agonist and antagonist muscles immediately after training.

Abstract We aimed to evaluate the applicability of a newly developed joint angle measurement system comprising six-axis inertial measurement unit sensors and tablet-based application for estimating joint angles from angular velocity and acceleration data. The application calculated orientation angles from single sensor data, with relative angles calculated using multiple sensors. In experiment 1, validity and reliability were examined using a test device. In experiment 2, static angles of five joints were calculated in four healthy participants using attached sensors and compared with universal goniometer values. In experiment 3, usability and satisfaction were evaluated using the System Usability Scale (SUS) and Quebec User Evaluation of Satisfaction with Assistive Technology (QUEST)-like scale. In experiment 1, mean difference and root mean squared error (RMSE) between the developed system and test device were < 0.2° and < 1.0°, respectively, across all axes. In experiment 2, when data from all joints were pooled, mean difference and RMSE were 0.2° and 3.8°, respectively. Mean difference and RMSE across each joint were < 5°, indicating the system is comparable to universal goniometer. In experiment 3, median SUS and QUEST-like scores were 73.8 and 4.0, respectively, indicating good usability and satisfaction. The developed system has high accuracy and sufficient validity for human joint angles, with good usability and satisfaction.