Toward a Unified Framework for Cognitive Maps.

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Neuroimaging and Cognitive Neuroscience

Background:

• Spatial navigation relies on neural representations of self-location in the hippocampus (HPC) and entorhinal cortex (EC).
• HPC place cells exhibit unimodal spatial tuning, while EC grid cells show periodic tuning across modules with varying periods.

Purpose of the Study:

• To integrate neural encoding and decoding into a unified framework for spatial information processing.
• To investigate how modular grid cell representations in the EC are decoded by HPC neurons.

Main Methods:

• Developed an encoding model for modular grid cell representations in the EC.
• Implemented a belief propagation decoding model to simulate information flow to the HPC.
• Utilized numerical simulations to analyze the synergy and error-correcting capabilities of modular structures.

Main Results:

• The modular structure of EC grid cells demonstrates positive synergy, enhancing error-correction through inter-module coupling.
• Decoding accuracy significantly improved with an increased number of modules (e.g., four modules successfully decoded self-location, unlike two).
• Belief propagation dynamics revealed complex information updates due to interactions among modules with diverse scales.

Conclusions:

• The unified framework elucidates the encoding and decoding of self-location in the EC-HPC pathway.
• Modular organization in the EC is critical for robust spatial information processing and accurate self-location decoding.
• This research provides insights into the neural mechanisms underlying spatial navigation and information flow in the brain.