Continuous attractor network model for conjunctive position-by-velocity tuning of grid cells

Area of Science:

• Neuroscience
• Computational Neuroscience
• Cognitive Science

Background:

• Grid cells in the medial entorhinal cortex (MEC) form triangular spatial patterns crucial for navigation.
• These grid cells exhibit selectivity for head direction and are modulated by animal speed.
• Existing models include attractor networks, theta oscillations, and single-unit mechanisms.

Purpose of the Study:

• To present a novel attractor network model for MEC grid cells.
• To account for the combined position and velocity selectivity of grid cells.
• To demonstrate robust path integration capabilities of the proposed model.

Main Methods:

• Development of a new attractor network model.
• Simulations to test the model's ability to replicate grid cell properties.
• Analysis of path integration performance under perturbed network conditions.

Main Results:

• The proposed network model successfully accounts for conjunctive position-by-velocity selectivity in grid cells.
• The model demonstrates robust path integration.
• The network maintains functionality despite random perturbations in recurrent connections.

Conclusions:

• Attractor network dynamics provide a viable mechanism for grid cell function.
• The model offers a unified explanation for position and velocity coding in MEC.
• This work advances our understanding of neural mechanisms underlying spatial navigation.