Cyclic transitions between higher order motifs underlie sustained asynchronous spiking in sparse recurrent networks

Area of Science:

• Computational Neuroscience
• Systems Neuroscience
• Network Science

Background:

• Neocortical circuits require stable spiking activity for efficient computation and information propagation.
• This stable activity is characterized by asynchronous, low-rate, and critical dynamics.
• The mechanisms underlying the maintenance of this asynchronous spiking regime remain unclear.

Purpose of the Study:

• To algorithmically construct spiking neural network models that mimic neocortical topological statistics.
• To identify the conditions and mechanisms that sustain asynchronous neural activity.
• To investigate the role of higher-order network patterns in maintaining stable spiking dynamics.

Main Methods:

• Algorithmic construction of spiking neural networks (5000 neurons) using neocortical topological statistics and objective functions for naturalistic activity.
• Analysis of network activity using functional graphs based on pairwise spike dependencies and recruitment networks.
• Quantification of motif transitions as a Markov process to identify critical dynamics for sustained activity.

Main Results:

• Sustained asynchronous activity was observed under specific initial conditions, while others led to truncated activity, indicating factors beyond synchrony, rate, or criticality.
• A periodic, low-variance transition between isomorphic triangle motifs in recruitment networks was identified in sustained simulations.
• Failure to engage in this stereotyped motif dominance cycling resulted in early truncation of spiking activity.

Conclusions:

• Excitatory higher-order patterns, specifically the cycling of triangle motifs, play a crucial role in sustaining asynchronous activity in sparse recurrent networks.
• This motif cycling phenomenon is robust across manipulations of synaptic weights and topologies.
• The findings offer a potential explanation for observed connectivity and activity patterns in the neocortex.