Static and dynamic functional connectome reveals reconfiguration profiles of whole-brain network across cognitive states

• Published in: Harvard data science review 2023-10, Vol.7 (3), p.1034-1050
• Main Authors: Zhang, Heming, Meng, Chun, Di, Xin, Wu, Xiao, Biswal, Bharat
• Format: Article
• Language: English
• Subjects: Brain; Brain architecture; Cognition & reasoning; Cognitive ability; Cognitive process; Drift diffusion model; Dynamic functional connectivity; Functional integration; Network reconfiguration; Network topologies; Neural networks; Reconfiguration; Sensorimotor system; Static functional connectivity; Time
• ISSN: 2472-1751; 2472-1751; 2644-2353
• DOI: 10.1162/netn_a_00314

Assessment of functional connectivity (FC) has revealed a great deal of knowledge about the macroscale spatiotemporal organization of the brain network. Recent studies found task-versus-rest network reconfigurations were crucial for cognitive functioning. However, brain network reconfiguration remains unclear among different cognitive states, considering both aggregate and time-resolved FC profiles. The current study utilized static FC (sFC, i.e., long timescale aggregate FC) and sliding window–based dynamic FC (dFC, i.e., short timescale time-varying FC) approaches to investigate the similarity and alterations of edge weights and network topology at different cognitive loads, particularly their relationships with specific cognitive process. Both dFC/sFC networks showed subtle but significant reconfigurations that correlated with task performance. At higher cognitive load, brain network reconfiguration displayed increased functional integration in the sFC-based aggregate network, but faster and larger variability of modular reorganization in the dFC-based time-varying network, suggesting difficult tasks require more integrated and flexible network reconfigurations. Moreover, sFC-based network reconfigurations mainly linked with the sensorimotor and low-order cognitive processes, but dFC-based network reconfigurations mainly linked with the high-order cognitive process. Our findings suggest that reconfiguration profiles of sFC/dFC networks provide specific information about cognitive functioning, which could potentially be used to study brain function and disorders. A key question about functional brain connectome is whether a specific cognitive process depends on stronger or weaker network reconfiguration. A wide range of reconfiguration profiles based on both sFC and dFC should be taken into account to provide more detailed information. The current study first investigated static and dynamic network reconfigurations between three cognitive states and then linked them with behavioral parameters reflecting sensorimotor and cognitive processes. We further explored the relationship between sFC and dFC reconfiguration patterns. This work contributes to better understanding of relationships between cognitive functioning and network reconfiguration.