EMG-Based Simultaneous Online Control for Rehabilitation Exercises of the Upper Limb in Stroke Survivors

Myoelectric control is gaining interest as an attractive approach to increase patients’ motivation during neurorehabilitation exercises by providing means to perform engaging online control tasks with immediate feedback of performance. Applying this approach clinically is, however, challenging as traditional interfaces based on wet electrodes require time to set up. Here, we present a novel dry textile interface with regression-based control that can be applied in a clinical context. We demonstrate the system by implementing a target reaching task using an EMG-based simultaneous control of 2 Degree-of-Freedom (DoF) for wrist exercises performed by 5 healthy participants and 5 stroke survivors. Preliminary results have shown that healthy participants could successfully reach all the targets generating stable trajectories using combined wrist motions (flexion/extension and ulnar/radial deviations). The two less affected patients could reach half of the targets generating longer and more unstable trajectories. The system has therefore shown promising results for clinical applications, although future work needs to be done to reduce the entry-level complexity of the tasks as well as evaluate the patients’ engagement and motivation.