Navigational systems in the human brain dynamically code for past, present, and future trajectories

Abstract Navigational trajectory planning requires the interaction of systems that include spatial orientation and memory. Here, we used a complex navigation task paired with fMRI pattern classification to examine head and travel direction tuning throughout the human brain. Rather than a single, static network, we report multiple simultaneous subnetworks that 1) have strong connections with both allocentric (world-centered) and egocentric (viewer-centered) movement trajectories, 2) change during the course of exploration, 3) code for past and future movements as well as the present direction, and 4) are strongest for individuals who convert their trajectories into egocentric movements once they have learned the environment. These findings shift our understanding of the neural processes underlying navigation from static structure-function relationships to a dynamic understanding of the multiple brain networks that support active navigation. The insights into the nature of individual navigation abilities uncovered here challenge the dominant framework of largely allocentric coding for successful navigation in complex environments, and replace this with a new framework that relies on multiple co-existing dynamic computations.