We are delighted to share that our team, collaborated with team from the imaging center NICUM, has published a comprehensive study in NeuroImage systematically comparing 20 functional connectivity (FC) metrics across four large neuroimaging datasets, totaling 1187 participants. This work addresses a fundamental question in resting-state fMRI research: which connectivity metrics best capture biologically meaningful changes in brain networks?

Using multimodal MRI data, we examined how different FC metrics – including correlational, distance-based, spectral, and information-theoretic measures – perform in detecting known reductions in brain connectivity associated with healthy aging and, in a smaller dataset, malignant brain tumors. Our analyses reveal that correlational (e.g., Pearson, Spearman, Kendall’s tau) and distance-based metrics (e.g., Euclidean, cosine) are most sensitive to age-related decreases in global connectivity, whereas partial correlation – despite its popularity – performed poorly in capturing these effects.

Importantly, this study highlights that the choice of FC metric has a profound impact on scientific conclusions, affecting results in default-mode network connectivity, macroscale gradient composition, brain–behavior associations, and case-control comparisons. We also report that scanning parameters (e.g., repetition time, sequence length) and individual factors can influence which metric is most informative.

These findings offer a data-driven roadmap for selecting FC metrics in future studies of aging, cognition, and disease. They also support the growing call for explicit reporting of theoretical, methodological, and confounding properties in connectivity research, aligning with recent frameworks for mechanistic interpretation of FC.

Reference:

Roell, L., Wunderlich, S., Roell, D., Raabe, F., Wagner, E., Shi, Z., Schmitt, A., Falkai, P., Stoecklein, S., & Keeser, D. (2025). How to measure functional connectivity using resting-state fMRI? A comprehensive empirical exploration of different connectivity metrics. NeuroImage, 121195. https://www.sciencedirect.com/science/article/pii/S1053811925001983