Making our way through the world: Towards a functional understanding of the brain's spatial circuits

Many animals make return trips from a home base to gather food and supplies, mate, or survive the seasons. In a world of unreliable and ambiguous cues, localizing within familiar environments and mapping new environments - functions critical for making successful return trips - is a complex problem requiring memory, integration, and inference. We review some key features of the mammalian brain's navigation system and its computational challenges, as well as the task neuroscientists face in understanding how its components interact and function. We argue that synthesizing the wide body of neural phenomenology requires formalization of the navigation problem as one of sequential probabilistic inference, as done in the robotics field of simultaneous localization and mapping (SLAM).