How fundyn/dfcwalk works
Project type and purpose
• This is an open-source MATLAB toolbox (“dFCwalk”) for analyzing dynamic Functional Connectivity (dFC) in neural time-series. It implements two core workflows:
a. dFC Speed Analysis – computes time-resolved reconfiguration rates (“speeds”) of functional networks.
b. Meta-Connectivity (MC) Analysis – groups links whose fluctuations co-vary into modules and analyzes module-specific dFC speed.
• There is no traditional frontend/backend: it’s a computational library invoked in MATLAB, providing functions, utilities, and plotting routines.
File structure
• All .m files reside at top level – no subfolders.
• Key clusters of functions by name patterns:
• Speed analyses: BuildSpeedHistogram.m, PlotSpeedHisto.m, dFC_Speeds.m
• Stream conversion & metrics: TS2FC.m, TS2dFCstream.m, TS2eFC.m, dFCstream2MC.m, dFCstream2dFC.m
• MC analyses: Pick_MCModuleStream.m, Which_MCIndex_ThisLink.m, Which_StreamIndices_ThisMCmodule.m
• Utility / indexing: Which_Link_ThisMCIndex.m, Which_StreamIndex_ThisLink.m, Which_StreamIndex_ThisLink.m
• Statistical surrogates and info theory: PhaseRand_surrogates.m, Mutual_Information_Labels.m
• Community detection: community_louvain.m (from Brain Connectivity Toolbox)
• Vectorization: Matrix2Vec.m
• No build scripts or deployment configs beyond MATLAB path instructions.
README for additional insights
• Confirms toolbox scope: speed analyses and MC analyses.
• Suggests companion toolboxes: Brain Connectivity Toolbox, Fluctuation Analysis Toolbox.
• Encourages using function‐level help for syntaxes.