High-capacity embedding of synfire chains in a cortical network model

Area of Science:

• Computational Neuroscience
• Neural Network Modeling
• Systems Neuroscience

Background:

• Synfire chains, crucial for timed spike sequence propagation, pose challenges for efficient cortical embedding.
• Understanding how these chains integrate into large-scale brain networks is a key question in neuroscience.

Purpose of the Study:

• To develop and analyze a model for embedding synfire chains within a cortical-scale recurrent network.
• To investigate the factors influencing embedding capacity and synfire wave propagation.

Main Methods:

• Utilized a computational model with conductance-based synapses, balanced chains, and variable transmission delays.
• Employed mean-field analysis to describe network equilibrium and explored embedding capacity limits.
• Conducted simulations to validate theoretical predictions across various network parameters.

Main Results:

• Achieved substantially higher embedding capacities compared to previous spiking neuron models.
• Demonstrated that recurrent background noise regulates the number of synfire waves.
• Identified an optimal inhibition level crucial for balancing stability and noise response.
• Showcased superior contrast with conductance-based synapses over current-based ones.

Conclusions:

• The model provides a framework for efficient synfire chain embedding in realistic cortical networks.
• Conductance-based synapses and variable delays are key to enhanced embedding capacity and wave regulation.
• Network parameters, including inhibition, must be finely tuned for optimal performance.