Ooi et al.’s (2025) demonstration that longer fMRI scans improve BWAS reliability offers a welcome, empirically grounded response to the methodological concerns raised by Marek et al. (2022). This commentary celebrates this solution while arguing that it improves measurement precision without resolving conceptual limitations. BWAS assumes functional connectivity is a stable trait, but accumulating evidence suggests that brain function is inherently dynamic, characterized by metastable states and context-dependent transitions. Longer scans improve measurement within states but do not capture the essential transitions between them. The field needs both rigorous measurement and designs that treat the brain as a dynamical system, characterizing state spaces and trajectories rather than static averages. This synthesis offers a path where statistical rigour and mechanistic understanding reinforce rather than oppose each other.
Anthony R. McIntosh (Wed,) studied this question.