Cluster Synchronization of Kuramoto Oscillators and the Method of Averaging

TL;DR

This paper establishes rigorous conditions for cluster synchronization in Kuramoto oscillators using an advanced averaging method and nonmonotonic Lyapunov functions, with applications to brain network analysis.

Contribution

It extends averaging stability theory with nonmonotonic Lyapunov functions to derive new cluster synchronization conditions for Kuramoto oscillators.

Findings

Synchronization occurs under weak inter-cluster coupling or large frequency differences.

Invariant manifolds ensure cluster phase cohesiveness.

Application to brain networks reveals parameter-connectivity relations.

Abstract

Rigorous conditions for cluster synchronization of Kuramoto oscillators are presented. The method of averaging plays an important role in stability analysis, but the standard Lyapunov's second method is not applicable due to the lack of uniform continuity. This paper contributes to overcoming this difficulty with the help of nonmonotonic Lyapunov functions. Our extensions of averaging in stability theory are key to derive the two interrelated cluster synchronization conditions: (i) the coupling strengths between clusters are sufficiently weak and/or (ii) the natural frequencies are largely different between clusters. Cluster phase cohesiveness in the absence of network partitions ensuring the existence of invariant manifolds is also investigated. Moreover, we apply our theoretical findings to brain networks and exhibit certain relations among network parameters and functional connectivity.