NEURAL CODING AND POPULATION GEOMETRY OF AGENT CLASS AND IDENTITY IN THE BAT HIPPOCAMPUS

The hippocampus is essential for navigation and contains place cells, neurons that encode an individual’s self-location. However, natural navigation occurs in complex environments populated by various agents, including conspecific animals, other species, and behaviorally relevant objects. To understand how the hippocampus encodes such agents during navigation, we performed wireless neural recordings from dorsal-CA1 hippocampal neurons of wild-caught bats, as they flew toward conspecifics (other bats), heterospecifics (humans), or objects. Hippocampal neurons encoded both the agent’s class (humans/bats/objects) and identity (e.g. human1 or human2) – via gain-modulation (‘rate-remapping’) – which allowed excellent decoding of agent class and identity, while preserving spatial tuning. The neuronal population activity formed a hierarchical representation, with agent identity being nested within the representation of agent class. Overall, hippocampal neurons encoded both agent class and identity during navigation, creating a structured representation of diverse agents that supports class abstraction.