Madhav, Manu

Dr. Madhav studies cognitive representations in the brain, specifically in the hippocampal formation. The neural correlates of these representations can be recorded as the activity of place cells, grid cells, and head-direction cells, among several other cell types. His lab will investigate how these representations are formed from sensory inputs, how they are flexibly modulated by task demands, sensory availability and behavioral states, and how they contribute to navigation and control of behavior. There will be two initial research thrusts;

(1) How do the topology of cognitive representations change to adapt the task demand? In a virtual reality 'dome' apparatus, rats will be trained to perform complex tasks that require the representation of non-spatial cues in addition to spatial variables. The study will examine how the topology of the representation, as decoded from high-density electrophysiological recordings, conform to the topology of the task-related cues.

(2) How do flexible representations result in robust navigation? Both rats and human participants in head-mounted virtual reality will navigate a 'maze' apparatus. The study will examine how varying levels of information related to the task and goal will affect the behavioral strategies of these subjects. These strategies will be compared to robotic path planning, reinforcement learning and optimal control algorithms. In addition, the study will look for neural correlates of these algorithms in rodents using high-density electrophysiology.

Dr. Madhav's research aims to, in the long term, develop virtual-reality-based non-invasive strategies for the early detection and management of neurodegenerative dementias such as Alzheimer's disease.