My interests are in machine learning for closed-loop human-machine interfaces. To this end, I'm pursuing a number of paths:

--Increases in the amount and variety of BCI data being recorded in our lab and others places a new emphasis on transfer learning as a method for increasing BCI effectiveness across subjects. I am interested in applying new techniques from the machine-learning literature to BCI as well as studying what sort of techniques can be developed that take advantage of the unique nature of brain-based data for classification or regression.

--Despite decades of work on decoding prosthesis movements from EMG signals, a robust prosthesis for home use is still beyond our grasp. I am interested in whether machine learning can be used either to help the user learn more reliable control patterns, or whether it can be used to refine a prosthesis's performance during natural use.