Subspace communication in the hippocampal-retrosplenial axis

Area of Science:

• Neuroscience
• Systems Neuroscience
• Computational Neuroscience

Background:

• Hippocampal circuits are crucial for navigation and memory.
• Mechanisms for adapting hippocampal computations across experiences are not well understood.

Purpose of the Study:

• To investigate circuit-level mechanisms of hippocampal adaptation.
• To identify how hippocampal-neocortical communication transforms inputs to outputs across experiences.

Main Methods:

• Performed large-scale (1024-channel) recordings in behaving mice across the hippocampal-retrosplenial cortex (RSC) circuit.
• Utilized partial canonical correlation analysis (a linear dimensionality reduction technique) to identify communication subspaces between brain regions.
• Analyzed spiking activity in dentate gyrus (DG), CA3, CA2, CA1, and RSC.

Main Results:

• Identified low-dimensional communication subspaces linking hippocampal regions (DG, CA3, CA2) to the RSC.
• Found that intrinsic neuronal properties and anatomical location constrained subspace membership within the CA3-CA1-RSC axis.
• Observed that these subspaces recombine neuronal pools to support distinct interareal interactions across different brain states and experiences.
• Discovered that CA1-CA3 subspace reactivation during sleep correlated with replay, indicating a plasticity-stability balance.

Conclusions:

• Hippocampal-neocortical communication involves reconfiguring circuit motifs for flexible experience encoding.
• Communication subspaces provide a framework for understanding adaptive input-output transformations in the hippocampus.
• The findings suggest a dynamic model for how the brain integrates information for memory and navigation.