UCSF

The infinite range of human behaviors is made possible by the anatomic and functional complexity of our brains. Our brains are arranged as networks of interacting neural populations which perform computations both within and across areas. Past research has focused on the specific roles of different brain regions, parceling out computational steps in sensory, motor, cognitive, and affective processes. However, our understanding of how brain regions interact is extremely preliminary, and is bottlenecked by limitations in experimental approaches, recording technologies, interventional methods, and computational analyses. These limitations impact not only our comprehension of the nervous system, but also our ability to design, optimize, and implement new therapies for patients with neurological diseases and disorders.

This thesis first investigates cross-area communication in the motor system in the context of natural movement learning and closed-loop brain-machine interface (BMI) learning. It then proposes a framework for understanding and manipulating cross-area communication in the context of chronic pain, a disorder driven by pathological activation and coupling of sensory, cognitive, and affective regions. We find that, during natural motor learning, cross-area activity dynamics can (1) be distinguished from local dynamics, (2) develop representations of learned movements which predict single-trial behavior, (3) become coordinated with local dynamics over the course of learning, and (4) causally influence downstream local activity to drive learned behaviors. Preliminary results from BMI learning show that tasks requiring only local neural modulation also engage neural populations in partner brain regions, suggesting circuit-wide participation in new task learning. This knowledge of the brain’s ability to learn new cross-area activity patterns in the context of natural behaviors and external, device-based feedback informs our framework for designing closed-loop neuromodulatory therapies for refractory chronic pain given the devices currently available for treating nervous system disorders.