Direct interfaces with the brain provide exciting new ways to restore and repair neurological function. For instance, motor Brain-Machine Interfaces (BMIs) can bypass a paralyzed person’s injury by repurpose intact portions of their brain to control movements. Recent work shows that BMIs do not simply “decode” subjects’ intentions—they create new systems subjects learn to control. To improve BMI performance and usability, we must therefore better understand learning and control in these systems. I will present a survey of recent work and new directions exploring how the design of BMI systems influence BMI performance. I’ll touch on the importance of brain-decoder interactions and multi-learner approaches, control loop design, and neural signal selection. These examples highlight the important role of learning and closed-loop control in BMIs, and demonstrate the promise of engineering approaches based on optimizing learning and control rather than purely “decoding”.
Dr. Amy Orsborn is the Clare Boothe Luce Assistant Professor in Electrical & Computer Engineering and Bioengineering at the University of Washington. She works at the intersection of engineering and neuroscience to develop therapeutic neural interfaces. Her lab applies engineering techniques to optimize control and learning in neural interfaces, and uses neural interfaces as a tool to study the neural mechanisms of learning. Dr. Orsborn completed her Ph.D. at the UC Berkeley/UCSF Joint Graduate Program in Bioengineering. She trained as a postdoctoral researcher at NYU’s Center for Neural Science. Her work has been supported by NSF Graduate Research Fellowship, a pre-doctoral award from the American Heart Association, and a L’Oreal USA for Women in Science postdoctoral award.