Spike-timing computation properties of a feed-forward neural network model

Area of Science:

• Neuroscience
• Computational Neuroscience

Background:

• Brain function relies on dynamic neural network interactions.
• Network topology significantly influences signal processing and information flow.
• Understanding synaptic plasticity is key to deciphering neural computation.

Purpose of the Study:

• To investigate how network topology and background activity influence signal propagation through feed-forward neural networks.
• To determine the effectiveness of spike-triggered stimulation (STS) in reorganizing neural network dynamics under varying topological and activity conditions.

Main Methods:

• Modeled feed-forward networks (7-22 neurons) with spike-timing dependent plasticity (STDP).
• Simulated activity under different network topologies (single vs. parallel pathways) and background input correlations.
• Applied spike-triggered stimulation (STS) to perturb network dynamics and assess changes.

Main Results:

• Polysynaptic spike-timing relationships were absent in networks with single pathways and uncorrelated activity.
• Correlated background activity or parallel pathways induced robust polysynaptic spike-timing relationships.
• STS application led to predictable changes in synaptic strengths and spike-timing relationships when temporal relationships were present.

Conclusions:

• Precise temporal relationships in network activity, induced by input or topology, are essential for polysynaptic signal propagation.
• Network topology plays a critical role in enabling effective spike-triggered stimulation and reorganizing functional brain networks.
• Higher-order topological structures may be vital for maintaining polysynaptic correlations, even with weak synapses.