Control of Dynamic Activity Across Neural Representations
Speaker: Seth Egger
Abstract:
As we perceive, act, and think, our brains generate dynamic patterns of activity. It has been proposed that these dynamics reflect computations performed by neural circuits to drive changes in state across neural representations of thought and action. To act as the mechanism of computation, however, neural dynamics must integrate sensory information and reflect different behavioral contexts. I will discuss two experiments designed to better understand how dynamics are controlled by sensory input. The first probes the internal representation of time and addresses how neural dynamics adjust to flexibly anticipate the timing of upcoming events. The second addresses how sensory neurons drive dynamics across the internal representation of sensorimotor gain. Together, these experiments lay a foundation for understanding how the dynamic activity generated by circuits of neurons interact with sensory drive to compute the changes in internal states necessary for intelligent behavior.