Performance-based robotic assistance during rhythmic arm exercises.

Only two stroke patients were tested, so this is a technical pilot rather than evidence that the method works in a wider population. The main outcome was real-time smoothness during the exercise, not lasting arm function, independence, or daily activities. The target movement was only 10% above the patient's comfortable performance, which the authors suggest may not have been challenging enough. The patients were too weak to hold the handle by themselves and had to be strapped to it, with the forearm supported. The robot had residual friction, and the simulation used prescribed movement rather than actually delivering assistive force. Amplitude and velocity assistance were barely used because the patients maintained the desired amplitude and velocity, so those modes were not tested well.

Funded by FRIA and Van Goethem Brichand. The supplied text gives no further author declarations.

Do not conclude that this robot-assisted therapy improves long-term arm function after stroke. It was a technical pilot, and the measured outcome was real-time smoothness during assisted trials, not lasting recovery.