A Continuous Attractor Model with Realistic Neural and Synaptic Properties Quantitatively Reproduces Grid Cell...

Area of Science:

• Computational neuroscience
• Systems neuroscience

Background:

• Understanding neural mechanisms of cognition requires detailed computational models.
• Realistic cellular properties are crucial for accurate simulations of brain activity.

Purpose of the Study:

• To investigate if data-driven physiological details can reproduce neural firing patterns observed in vivo.
• To model spatial coding in the medial entorhinal cortex using a spiking continuous attractor network.

Main Methods:

• Utilized cellular properties from Hippocampome.org for a spiking continuous attractor network model.
• Modeled the rodent hippocampal formation for computational efficiency and experimental grounding.
• Incorporated biological characteristics like excitability, connectivity, and synaptic signaling.

Main Results:

• Simulations generated grid cell activity matching experimental data in spacing, size, and firing rates.
• The model successfully recreated grid field properties across different scales along the dorsoventral axis.
• A wide range of neural and synaptic parameters produced accurate grid fields.

Conclusions:

• The continuous attractor network model is compatible with spiking neural network implementations using biophysical data.
• Data-driven parameters from Hippocampome.org enable realistic simulations of neural circuits.
• Open-source software release facilitates community reuse and novel applications.