Mean field theory for a balanced hypercolumn model of orientation selectivity in primary visual cortex

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Theoretical Neuroscience

Background:

• Cortical neurons exhibit orientation selectivity, crucial for visual processing.
• Understanding the balanced state and firing dynamics in neural networks is a key challenge.

Purpose of the Study:

• To develop a quantitative mean field theory for balanced states in orientation hypercolumns.
• To self-consistently determine firing rates and correlations without neuron model restrictions.

Main Methods:

• Developed a complete mean field theory for a balanced state model.
• Implemented a numerical procedure to solve mean-field equations.
• Applied the model to integrate-and-fire neurons.

Main Results:

• Naturally reproduced key properties of orientation-selective neurons: irregular firing, linear gain function, and contrast-invariant tuning width.
• Showed firing irregularity is sensitive to synaptic strengths and consistent across orientations (Fano factor).
• Demonstrated input current noise tuning matches external input, while mean input current tuning depends on connectivity.

Conclusions:

• The mean field theory provides a robust framework for understanding neural network dynamics.
• Synaptic strengths play a critical role in shaping firing irregularity and network properties.
• The model offers insights into the computational principles underlying orientation selectivity in the cortex.