Place-cell capacity and volatility with grid-like inputs.

Area of Science:

• Computational Neuroscience
• Systems Neuroscience

Background:

• Place cells in the hippocampus are fundamental for spatial memory and navigation.
• Their activity patterns, specifically the arrangement of multiple firing fields, are not fully understood.
• Grid cells provide periodic input, potentially influencing place cell organization.

Purpose of the Study:

• To investigate the factors constraining the spatial arrangement of place cell fields.
• To quantify the repertoire and capacity of place cell arrangements based on grid cell inputs.
• To explore the implications of these constraints on the stability of the spatial code.

Main Methods:

• Modeling place cells as perceptrons receiving multiscale, periodic grid-cell inputs.
• Analytically enumerating the repertoire (possible unique field arrangements).
• Deriving the capacity (spatial range for achieving any arrangement).

Main Results:

• Place cells possess a very large and noise-robust repertoire of field arrangements.
• However, this repertoire represents a small fraction of all possible arrangements.
• Capacity is constrained, scaling with grid periods, limiting arrangements over larger distances.
• Altering grid-place weights significantly impacts existing arrangements.

Conclusions:

• Grid-cell inputs impose strong constraints on place cell field arrangements, forming a structured scaffold.
• The limited capacity and sensitivity to weight changes may explain the observed volatility in the spatial code.
• This provides a framework for understanding how structured inputs shape neural representations.