CTRL-labs: non-invasive neural interfaces for human augmentation

Patrick Kaifosh

Columbia University / CTRL-labs
Wednesday, February 13, 2019 at 12:00pm
560 Evans
As the nervous system’s evolved output, spinal motor neuron activity is from an evolutionary perspective a natural source of signals for a neural interface. Furthermore, the amplification of these signals by muscle fibers allows them to be measured non-invasively with surface electromyography (sEMG). CTRL-labs has developed a wireless wearable system that records state-of-the-art sEMG signals from the human forearm with dry electrodes and without the need for shaving or other skin preparation. Using this system, we demonstrate real-time detection and identification of individual motor unit action potentials, each of which corresponds to an action potential of an individual spinal motor neuron. This ability to monitor spiking activity of individual neurons sets sEMG apart from other non-invasive neural recording paradigms such as electroencephalography, functional magnetic resonance imaging, and near infrared spectroscopy. From the recorded sEMG signals, we also compute real-time predictions of joint angles, muscle tensions, and forces of the wrist and hand. The wireless and unobtrusive form factor of this recording system allows for long-duration monitoring of human neuromotor activity suitable not only for research and clinical applications, but also for real-time control in terms of motor unit action potentials, aggregate sEMG signals, and/or estimates of joint angles, muscle tensions, and forces. Relative to traditional human-computer interfaces, neuromotor interfaces have the potential to increase bandwidth by eliminating information loss as neural signals are converted to muscle tensions and then to input device signals. Meaningful reductions in latency are also achievable because sEMG signals precede forces and movement by tens of milliseconds. As a non-invasive neural interface with action potential resolution and the capability to augment human capacity for real-time control, this system expands the applicability of neural interface technology beyond research and clinical domains.