Special thanks to the Kavli JHU for funding my current work on how the bat hippocampus forms internal models of targets moving along naturalistic 2D trajectories.

A big shout to my co-authors, and to my mentor Cindy Moss, without whom none of this would have happened.

Collectively, we suggest a framework where hippocampus encodes the location of any object actively tracked by the animal.

We hypothesize that previous reports of neurons encoding conspecifics and human experimenters likely co-opted intrinsic object-tracking mechanisms

Finally, leveraging the active nature of echolocation, we show that hippocampal object representations emerge only when bats actively track the target.

We show that a subpopulation of neurons encoded the allocentric location of the target, and another encoded the egocentric distance of the target.

To distinguish between egocentric and allocentric coding, we conducted additional experiments in which the bat tracked the object from a different location in the room.

As the bats actively tracked the target, a population of ‘auditory object cells’ represented the moving target.

Using this metric, we trained bats to track a moving object in the dark via echolocation and recorded from hippocampal CA1 neurons, during which they increased their call rate as the target approached.

(Video credits - Mel Wohlgemuth)

Testing this idea requires a model system with an explicit metric of object tracking.

Echolocating bats adjust the rate and duration of their calls when tracking objects, providing a quantifiable measure of spatial attention to moving targets

Further, some hippocampal studies have reported neurons encoding the locations of moving objects, whereas others have not. We hypothesized that this discrepancy depends on whether the animal attended to the object, with the hippocampus encoding the location of any moving object under active tracking

Since echolocating bats use sound to explore complex 3D spaces, they present powerful model systems to answer how the brain builds spatial representations from non-visual inputs.