Feasibility of force myography for the direct control of an assistive robotic hand orthosis in non-impaired individuals.

The study included only 10 participants, all male and neurologically-intact, so it does not show performance in people with stroke or spinal cord injury, the intended users of the device. The system was not tested online; the orthosis state was manually set and known, which simplified the classification problem. Only binary open-versus-try-to-close and closed-versus-try-to-open decisions were assessed, not a full multi-gesture control system. Performance varied between participants, and only 6 of 10 achieved above 90% accuracy for both states. The closed-state task was significantly less accurate than the open-state task, and the most critical error was unintended opening while holding an object, although it occurred less than 7% of the time. The band was custom-built, sensor placement and tightness were adjusted by subjective report, and the number of active sensors varied with forearm size. The training protocol required many repetitions and arm configurations, which the authors note may be burdensome for real users. The authors state that online performance in people with neurological hand impairments is expected to be lower than the offline healthy-participant results.

The work was funded by the National Centre of Competence in Research Robotics, the National Research Foundation Singapore under its Campus for Research Excellence and Technological Enterprise (CREATE) programme, and the Swiss Federal Institute of Technology Zurich.

Do not conclude that force myography can reliably control a robotic hand orthosis for patients with hand weakness. The study tested only 10 healthy men, classified intended gestures offline rather than driving the device online, and only 6 of 10 reached 90% accuracy for both states.