Current neural prosthetic decoding algorithms are finicky: they require daily calibration sessions and lengthy spike-sorting sessions to achieve optimal performance. To be clinically viable, decoding algorithms will need to run autonomously, with little-to-no user intervention for periods of weeks or months. One of the thrusts of our lab is on designing these algorithms. Working in collaboration with Valerie Ventura of Carnegie Mellon's Statistics Department and Wei Wang of the Department of Physical Medicine and Rehabilitation at the University of Pittsburgh, we are building the next generation of prosthetic decoding algorithms that will translate brain-computer interfaces from a research tool into clinical practice.

Funded by the DARPA RE-NET program