A dynamic functional connectivity toolbox for multiverse analysis

Abstract

ABSTRACTVarious methods for estimating dynamic functional connectivity from fMRI data and subsequent analyses through graph theoretic approaches have been introduced in recent years. But with many of the ground truths unknown, researchers are often faced with arbitrary yet defensible decisions for their analyses, which raises concerns about the reproducibility of results. We here aim to address this issue through the implementation of a wide array of dynamic functional connectivity methods in a unified Python software package, facilitating a diverse exploration of brain dynamics. Anchored in the framework of multiverse analysis, the present work further introduces a workflow for the systematic examination of various methodological decisions. The developed toolbox is supplemented by a graphical user interface for easy usability and accessibility if usage outside of a script-based Python analysis pipeline is desired. Further, extensive demo scripts are provided for researchers to easily adapt this approach for their own analyses.