Achievement

Matt Golub, a doctoral candidate in Electrical and Computer Engineering, works on brain-machine interface (BMI) technology that allows a monkey to control a video cursor with its mind. BMI technology shows great promise for paralyzed patients or amputees to control prosthetic limbs by thinking about desired motions. One of the key technical problems is how to best extract the required control information from the firing patterns of cells in motor cortex. During his IGERT training, Golub performed BMI experiments with a rhesus monkey, testing a novel algorithm called the speed-dampening Kalman filter that improves controllability of movement. His results were reported at the IEEE Engineering in Medicine and Biology Conference in San Diego in August 2012. Subsequently, Golub applied machine learning and a novel probabilistic framework to extract the monkey's internal model of the cursor's dynamics. He found that the model could account for 80% of the error on a cursor movement task.