Researchers at the Technical University of Munich have unveiled a soft pneumatic exoskeleton glove capable of restoring grasping functionality for individuals with severe hand paralysis. The development, detailed in Nature Machine Intelligence, integrates surface electromyography sensors, machine learning classification, and targeted pneumatic actuators to translate residual neuromuscular signals into controlled hand movements. The device operates by capturing electrical activity from intact forearm muscles through adhesive electrodes. A machine learning algorithm processes these signals to predict the user’s grasping intention with 97 percent accuracy. Upon detection, a network of thirteen pneumatic lines inflates air cushions embedded in a textile sleeve. These cushions individually flex each finger and enable wrist rotation, allowing users to securely hold everyday items such as utensils and drinking glasses. To ensure stability during manipulation, auxiliary motion sensors maintain continuous grip pressure throughout transport movements, effectively preventing accidental drops. Clinical validation was conducted through close collaboration with a patient experiencing advanced amyotrophic lateral sclerosis. Despite profound neuromuscular degradation, the patient retained isolated motor control in one thumb joint. By isolating electromyographic signals from the flexor pollicis longus muscle, researchers successfully calibrated the system. Within five minutes of training, the participant achieved reliable object manipulation, including lifting a fork for the first time in four years and successfully interacting with a motion-controlled video game. The trial confirms the exoskeleton’s capacity to bypass severe neurological impairment and restore practical dexterity. Unlike conventional rigid robotic braces, this design prioritizes accessibility and patient comfort. The fabric-based construction can be assembled using low-cost materials and basic sewing techniques, making the technology financially viable for broad clinical distribution. Lead researcher Dr. John Nassour noted that the system successfully merges precise signal interpretation with ergonomically optimized hardware. Gordon Cheng, director of TUM’s Institute for Cognitive Systems, emphasized that the prototype demonstrates functional recovery is achievable even under extreme neurological conditions. The technology’s applicability extends beyond motor neuron diseases. Neurologist Tobias Wächter of Klinik Passauer Wolf highlighted the glove’s potential to assist patients with flaccid paralysis, including individuals recovering from peripheral nerve trauma following vehicular accidents or those with polyneuropathy. The TUM research team is currently adapting interface protocols and training regimens for stroke survivors and additional rehabilitation demographics. By combining advanced signal processing with affordable soft robotics, the prototype establishes a scalable clinical pathway for restoring fine motor function and improving long-term quality of life for paralyzed patients worldwide.