Left-right-alternating theta sweeps in the entorhinal-hippocampal spatial map

Place cells in the hippocampus and grid cells in the entorhinal cortex are elements of a neural map of self-position ^1–5 . To benefit navigation, this representation must be dynamically related to surrounding locations ² . A candidate mechanism for linking places along an animal’s path has been described in place cells, where the sequence of spikes within each cycle of the hippocampal theta oscillation encodes a trajectory from the animal’s current location towards upcoming locations ^6–8 . In mazes that bifurcate, such trajectories alternately traverse the two upcoming arms as the animal approaches the choice point ^9,10 , raising the possibility that the trajectories express available forward paths encoded on previous trials ¹⁰ . However, to bridge the animal’s path with the wider environment, beyond places previously or subsequently visited, an experience-independent spatial sampling mechanism might be required. Here we show in freely moving rats, that within individual theta cycles, ensembles of grid cells and place cells encode a position signal that sweeps linearly outwards from the animal’s location into the ambient environment, with sweep direction alternating stereotypically between left and right across successive theta cycles. These sweeps were accompanied by, and aligned with, a similarly alternating directional signal in a discrete population of parasubiculum cells with putative connections to grid cells via conjunctive grid×direction cells. Sweeps extended into never-visited locations that were inaccessible to the animal and persisted during REM sleep. Sweep directions could be explained by an algorithm that maximizes cumulative coverage of surrounding space. The sustained and unconditional expression of theta-patterned left-right-alternating sweeps in the entorhinal-hippocampal positioning system provides an efficient ‘look-around’ mechanism for sampling locations beyond the travelled path.