Hippocampal predictive representations of moving auditory objects in echolocating bats

Central to many natural behaviors is the use of past or current information to predict future states. For example, echolocating bats rely on internal models to predict target motion trajectories under occlusion, which is essential for intercepting erratically moving prey in cluttered environments. However, the brain regions and neural mechanisms underlying the generation and representation of predictive models guiding action selection are poorly understood. To bridge this gap, bats were trained to track a sonar target moving in 1-D along the range axis, while we recorded neural activity from CA1 of the hippocampus. Hippocampal CA1 neurons encoded the allocentric object location or egocentric object distance, but only when the bats actively tracked a moving target using echolocation. These findings demonstrate that active sonar tracking drives hippocampal object representations. In a second experiment, bats tracked targets moving in 2-D in the presence of occluders that temporarily blocked sensory access to moving stimuli. We analyzed the directional aim of the bat’s sonar and hippocampal neural responses over repeated presentations of 2-D target trajectories in combination with experimentally controlled occlusion placements. These results lay the foundation for understanding predictive coding of moving auditory objects in the bat hippocampus.