Assistive devices may aid motor recovery after a stroke, but access to such devices is limited. Exosuits can aid human movement and may be more accessible alternatives to current devices, but we know little about how they might assist post-stroke upper extremity movement. Here, we designed an exosuit actuator to support shoulder abduction, which we call an "exomuscle" and is based on a form of growing robot called a pneumatic-reel actuator. We also assembled a ceilingmounted support for a positive control. We verified that both supports reduce the activity of shoulder abductor muscles and do not impede range of motion in healthy participants (n=4). Then, we measured reachable workspace area in stroke survivors (n=6) with both supports and without support. Our exomuscle increased workspace area in four participants (180±90 cm 2 ) while the ceiling support increased workspace area in five participants (792±540 cm 2 ). Design decisions that reduce exomuscle complexity, such as leaving the forearm free, likely contribute to performance differences between the two supports. Heterogeneity amongst stroke survivors' abilities likely contribute to high variability in our results. Though both supports performed similarly for healthy participants, performance differences in stroke survivors highlight the need to validate assistive devices in the target population.

No abstract

Assistive devices may aid motor recovery after a stroke, but access to such devices is limited. Exosuits may be a more accessible alternative to current devices and have been used to successfully assist movement after a variety of motor impairments, but we know little about how they might assist post-stroke upper extremity movement. Here, we designed an exosuit actuator to support shoulder abduction, which we call an "exomuscle" and is based on a form of growing robot known as a pneumatic-reel actuator. We also assembled a ceiling-mounted support to serve as a positive control. We verified that both supports reduce the activity of shoulder abductor muscles and do not impede range of motion in healthy participants (n=4). Then, we measured reachable workspace area in stroke survivors (n=6) both with and without support. Our exomuscle increased workspace area in four participants (180±90 cm 2 ) while the ceiling support increased workspace area in five participants (792±540 cm 2 ). Design decisions that make the exomuscle wearable, such as leaving the forearm free, likely contribute to the performance differences between the two supports. Heterogeneity amongst stroke survivors' abilities likely contribute to a high variability in our results. Though both supports resulted in similar performance for healthy participants, the performance differences in stroke survivors highlight the need to validate assistive devices in the target population.

No abstract