Comparison of Resting-State Functional MRI Methods for Characterizing Brain Dynamics

Area of Science:

• Neuroscience
• Cognitive Science

Background:

• Resting-state functional MRI (fMRI) reveals dynamic functional connectivity patterns.
• Analysis methods like sliding window connectivity (SWC), phase synchrony (PS), and co-activation patterns (CAP) identify temporal brain states.
• Previous studies often use a single method, limiting comparative understanding.

Purpose of the Study:

• To directly compare k-means clustering results using SWC, PS, and CAP resting-state dynamics methods.
• To quantify brain state dynamics using high-resolution human connectome project data.
• To assess how these methods characterize brain states in relation to quasi-periodic patterns (QPPs).

Main Methods:

• Applied k-means clustering to frame-wise functional connectivity derived from SWC, PS, and CAP.
• Utilized high-resolution fMRI data from the Human Connectome Project.
• Compared clustering outcomes and brain state trajectories, particularly concerning QPP sequences.

Main Results:

• SWC, PS, and CAP methods produced distinct clusters and temporal trajectories.
• PS clustering grouped most QPP sequence mid-points into a single cluster.
• CAP clustering separated QPP sequence phases into different clusters, while SWC showed less sensitivity to QPPs.

Conclusions:

• The choice of resting-state dynamics method significantly influences brain state characterization.
• Understanding these differences is crucial for accurate interpretation of brain dynamics.
• This study enhances the conceptual and practical application of fMRI dynamics analysis tools.