Abstract
Functional brain imaging has revealed two types of dynamic patterns in the resting-state: spontaneous brain activity (BOLD dynamics) and the functional interconnections between these activities (functional connectivity dynamics, or FCD). While often studied separately, recent evidence suggests these dynamics share similar spatial patterns. This study characterizes local BOLD dynamics and FCD in the resting-state to investigate their between-region associations. Using a sliding-window method on resting-state fMRI data, we found that BOLD dynamics and FCD are significantly associated across brain regions. Crucially, these associations are heterogeneous: within-network FCD is negatively correlated with BOLD dynamics, while between-network FCD is positively correlated. These findings suggest distinct roles for local activity fluctuations in coordinating internal network stability versus global network flexibility.
Introduction
In neuroimaging, researchers typically distinguish between local signal fluctuations (local dynamics) and the coordination between distant brain regions (functional connectivity). It is now understood that neither is static; both evolve over time even during rest. While the spatial maps of local BOLD fluctuations and connectivity variability appear similar, the underlying relationship between them has remained unclear. Does high local variability destabilize connections, or does it facilitate large-scale communication? This study aims to quantify how local BOLD dynamics relate to the variability of functional connections across different intrinsic connectivity networks (ICNs).
Methods
The study utilized resting-state fMRI data from 102 healthy subjects (ages 20–23) from the 1000 Functional Connectomes Project. Brain regions were defined using the Dosenbach atlas, comprising 160 regions of interest (ROIs) across six ICNs: Cerebellum (CER), Cingulo-opercular (CON), Default Mode (DMN), Fronto-parietal (FPN), Occipital (OCC), and Sensorimotor (SMN).
Two primary metrics were calculated:
- Local BOLD Dynamics (LBD): The temporal variance of the BOLD signal within each ROI.
- Functional Connectivity Dynamics (FCD): Calculated using a sliding-window method (30-second windows) to determine the variance of correlation coefficients over time. FCD was further categorized into within-network (connections within the same ICN) and between-network (connections between different ICNs).
Key Results
The analysis revealed a "spatial mirror effect," where regions with high local signal variability, such as the DMN and FPN, also exhibited the most dynamic connectivity changes.
Quantitative analysis showed that BOLD dynamics have opposite effects on different types of connectivity:
- Within-network stability: LBD and within-network FCD were negatively correlated ($r = -0.0921, p < 0.0001$), suggesting that higher local variability stabilizes a network's internal connections.
- Between-network flexibility: LBD and between-network FCD were positively correlated ($r = 0.2272, p < 10^{-36}$), indicating that local variability promotes more dynamic interactions between different networks.
The DMN emerged as a unique hub, showing the strongest negative correlation between local dynamics and internal connectivity variability. This suggests the DMN is a highly coherent system that utilizes high activity levels to maintain internal stability while remaining flexible in its global interactions.
Discussion
The authors propose that these heterogeneous associations may be explained by "stochastic resonance." High-amplitude local signals might act as a buffer against noise within a network, maintaining internal stability. Simultaneously, this activity might push a region toward a "critical state," enabling it to switch rapidly between different global network configurations.
While the study provides significant insights, limitations include the use of a fixed sliding-window length and potential atlas-related biases. Future research should explore these dynamic associations during cognitive tasks to determine how they shift under different mental loads.
Conclusion
This study demonstrates that local BOLD dynamics and functional connectivity dynamics are deeply intertwined but functionally distinct. The heterogeneous nature of their relationship—stabilizing internal networks while facilitating external flexibility—underscores a potential fundamental mechanism for brain coordination and co-evolution.
Summary updated