An efficient coding theory for a dynamic trajectory predicts non-uniform allocation of entorhinal grid cells to...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cognitive Science

Background:

• Grid cells in the entorhinal cortex are crucial for spatial navigation, exhibiting periodic tuning curves.
• Recent findings show grid cells are organized into modules with uniform spacing.
• Understanding the coding principles of grid cells is key to deciphering neural representations of space.

Purpose of the Study:

• To develop a theory for efficient position coding by grid cells, incorporating temporal motion statistics.
• To predict the relationship between module population size and grid spacing.
• To identify a neural readout scheme for the grid cell code.

Main Methods:

• Theoretical modeling of grid cell network function.
• Analysis of temporal statistics of animal movement.
• Comparison of theoretical predictions with experimental data on grid cell modules.
• Development of a neural circuit model for code readout.

Main Results:

• The theory predicts a decrease in module population size with increasing grid spacing, matching experimental trends.
• A simple readout scheme for the grid cell code was identified.
• This readout scheme achieves accuracy comparable to optimal Bayesian decoders.
• The readout mechanism requires timescale persistence dependent on grid cell module.

Conclusions:

• The brain may use an efficient position representation by leveraging spatiotemporal statistics.
• This efficient coding principle is similar to those in early sensory processing.
• The proposed theory and readout scheme offer insights into neural spatial cognition.