Introduction:: Over the last decade exoskeletal rehabilitation robots have become ever more popular. Robotic therapy has established itself as new and powerful tool for stroke rehabilitation, as it can provide a much deeper understanding of the biomechanical properties of the patient’s limb, as well it can make the rehabilitation more effective and efficient. Our group has developed an arm exoskeleton (Fig. 1) capable of actuating the shoulder, elbow wrist and metacarpophalangeal (MCP) joint and simultaneously measuring the torques and forces around and along the three orthogonal axes at each joint.
Materials and Methods:: In this work we present sample results from four healthy and four stroke subjects. We begin with measuring the passive properties of each joint. To do so, after securely fastening the subject’s arm to the robot and having accurately aligned each anatomical axis with the actuation axis, each joint, one at time, is moved at 10 deg/s through the passive range of motion (PROM), which had been previously determined by a therapist, for five cycles. Simultaneously measuring the torques at every joint, enables us to not only have an accurate representation the local stiffnesses, but also of the coupled torques at the joints that are not being moved by the robot. To measure the active properties of the arm, we first lock each joint at the neutral position (determined during the PROM assessment), and then release one joint at a time, using an impedance control algorithm, which allows the subject to move the intended joint freely. The subject is then instructed to move the unlocked joint through their active range of motion (AROM) five times. This enables us not only to accurately measure the subjects AROM at each joint, but also to observe the coupled torque generated at all other joints when the intention is to only move one at a time.
Results, Conclusions, and Discussions:: As it has been shown in literature, stoke has profound effects on both passive and active properties of the arm. Our unique approach, enables us to observe the coupling among active and passive properties, adding a new dimension to the assessment of the effects of stroke. Figures 2 and 3 show respectively the typical active and passive couplings in a healthy control and stroke subject. We can clearly observe how these coupling relationships that normally exist in healthy subjects, together with ordinary biomechanical measures such as active/passive range of motion and local stiffness, are exacerbated through all measurements after stroke.
