Multiscale phase oscillations induced by cluster synchronisation in human connectome core network

TL;DR

This study explores how multiscale phase oscillations and cluster synchronization in the human connectome core network relate to brain dynamics, emphasizing the influence of network topology and edge weights on synchronization stability.

Contribution

It reveals the connection between multiscale oscillations and cluster synchronization in the connectome, highlighting the role of network topology and edge weights in these processes.

Findings

Identified three significant clusters of brain regions.

Demonstrated the primary role of network topology in cluster formation.

Showed that edge weights stabilize synchronization and reduce multifractality.

Abstract

Brain imaging data mapping onto human connectome networks enables the investigation of global brain dynamics, where the brain hubs play an essential role in transferring activity between different brain parts. At this scale, the synchronisation processes are increasingly investigated as one of the key mechanisms revealing many aspects of brain functional coherence in healthy brains and revealing deviations due to various brain disorders. For the human connectome core network, consisting of the eight brain hubs and the higher-order structure attached to them, previous simulations of Kuramoto phase oscillators at network nodes indicate instability of the global order parameter for a range of positive coupling strengths. In this work, we investigate the multiscale oscillations of the global phase order parameter and show that they are connected with the cluster synchronisation processes occurring in this range of couplings below the master stability threshold. We use the spectral graph analysis and eigenvector localisation methodology, where the clusters of nodes playing a similar role in the synchronisation processes have a small mutual distance in the eigenvector space. We determine three significant clusters of brain regions and show the position of hubs in them. With the parallel analysis of the weighted core network and its binary version, we demonstrate the primary role of the network's topology in the formation of synchronised clusters. Meanwhile, the wights of edges contribute to the hub's synchronisation with the surrounding cluster, stabilise the order parameter variations and reduce the multifractal spectrum.