Wearable sEMG interfaces for computer interaction after paralysis

People with cervical spinal cord injury often cannot sense or move their hands and fingers — but many retain zones of partial preservation in the forearm where motor units still fire under volitional control. This project asks whether those residual signals can be measured with a wearable surface-EMG interface and converted into reliable, everyday computer interaction.

I lead Carnegie Mellon’s collaboration with Meta Reality Labs on the EMG-driven gesture models used by the Meta Neural Band, and I am project lead on the HD-EMG offline validation and online decoding experiments behind it. A first study assessed longitudinal volitional control of individual motor units in persons with cervical SCI, validated against research-grade HD-EMG textile arrays. A follow-up study (now concluding) pairs the Neural Band with trackball pointer navigation in a hybrid interface: point with the trackball, click with a decoded motor-unit gesture — no hand ever leaves the trackball. The benchmark tasks are browser-based and hosted at chr.nml.wtf.

Selected publications

• Despradel Rumaldo DL, Murphy MD, et al. Enabling skilled human-computer interaction after paralysis via a wearable sEMG interface. Science Robotics (provisionally accepted, 2026).
• Despradel D, Murphy MD, et al. Enabling advanced interactions through closed-loop control of motor unit activity after tetraplegia. ACM UIST (adjunct), 2024.