We study general principles of how the neocortex represents and updates hypotheses, using the spatial computations that occur as part of complex natural behaviors in rodents.

Most behaviors in which rodents maintain a working memory and plan ahead are expressed by localizing their body in, or reasoning about, space:   navigation ,  foraging, and  predator/social interactions. We seek to use the computations that are inherent to these behaviors to probe the underlying neural mechanisms.

We aim to approach this question from two sides that support each other: 1) identify the computations performed by animals during complex ethological behaviors and make them tractable, and 2) exploit this to design targeted experiments to gain insights into their neural implementation, in turn informing better models.