UC Berkeley

Brain-machine interface (BMI) systems hold great promise for improving quality of life in patients with a number of motor and cognitive disabilities, but significant advancements must be made before these systems are clinically viable. Developing neuroprosthetic devices and adaptive deep brain stimulators that are neurobiologically-informed is essential for increasing the usability and scalability of these systems. To do so, we must first develop an understanding of how the brain interacts with these systems. Previous work has shown that learning to control a BMI is associated with neural adaptations that depend on large-scale networks across multiple brain regions. In this work, we aim to develop a better understanding of this large-scale learning process by analyzing how motor cortical population dynamics differ in groups of neurons used to directly control a BMI and the remaining recorded population. Then, we investigate the role of corticostriatal circuits in BMI control by simultaneously recording from dorsolateral prefrontal cortex, the caudate nucleus of the striatum, and motor cortex as a nonhuman primate controls a motor cortical BMI. Finally, we assess how high-frequency microstimulation administered in the striatum changes functional decision-making behavior and underlying neural representations of value in the caudate and anterior cingulate cortex with the goal of developing a better understanding of how stimulation-based therapies can be used for neuropsychiatric disorders. Altogether, this work aims to improve our understanding of the neurobiology of BMIs with the goal of developing neurobiologically informed neuroprosthetic devices.