TUM soft glove restores grasp via forearm EMG in single ALS case with 90% intent detection

Researchers at Technical University of Munich built a low-cost fabric exoskeleton with air cushions that individually flex fingers and rotate the wrist. Muscle signals from the flexor pollicis longus were classified by machine learning at 97% reliability in healthy volunteers and 90% in the patient, with motion sensors preventing accidental release during transport. The design costs little because the lead researcher sewed the prototype, yet it supported secure holding of forks, plates, and small cubes.

The work rests on a single-patient observational test rather than an RCT. The ALS participant retained only thumb-joint motion; five minutes of a video-game training task improved performance, but no control condition, larger cohort, or long-term retention data exist. Stroke adaptation is planned yet untested.

Prior myoelectric prostheses and rigid exoskeletons have shown higher grip forces in small trials but at greater cost and weight. This soft approach trades force for affordability and ease of donning; whether those advantages translate to daily use remains unknown.

Next steps require a multi-center study enrolling at least 20 stroke and ALS participants with pre-specified functional endpoints such as the Action Research Arm Test at three months.