Design and Preliminary Evaluation of a Perturbation-based Robot-assisted Assessment of Hand Sensorimotor Impairments

Abstract

Neurological injuries often lead to hand sensorimotor deficits, such as impaired corrective responses to mechanical disturbances. That is frequently caused by difficulties in integrating sensory feedback when executing motor responses and affects the general ability to interact with the external environment. It is challenging to quantitatively assess such impairments using clinical measures, whereas the existing robotic approaches mainly target shoulder joints. This work presents design and preliminary evaluation of a robot-assisted assessment of hand sensorimotor function, in which subjects need to rely on proprioceptive feedback to actively correct for load perturbations. The task is implemented on a one degree-of-freedom robotic platform focused on the index finger metacarpophalangeal joint, which can apply a precise stimuli and measure kinematic responses. In a study with six healthy young subjects we investigated four different load perturbations parameters (1.5N step-, 3N step-, 2.5N ramp and 5N ramp, each on top of a 2.5N load) to find the most promising perturbation scenario for future studies with neurological subjects. We found that a ramp perturbation applied over 1 second with 2.5 N amplitude is least prone to confounds such as involuntary responses or high intra-subject variability, hence it is expected to be the most accurate in distinguishing between impairments. Overall, the presented perturbation-based assessment provides insights into sensorimotor system integrity in a quantitative and reproducible way and hence has the potential to add value to the battery of existing clinical methods.