Multiscale fMRI analysis reveals hierarchical network disruptions underlying disorders of consciousness

Disorders of consciousness following anoxic brain injury present profound clinical and scientific challenges, with current diagnostic methods often failing to capture underlying neuronal network pathophysiology. Here, we develop a novel multiscale framework combining global, macro-, and local-level network analyses to characterize functional connectivity alterations in vegetative (VS) and minimally conscious states (MCS) in patients with anoxic brain damage/ after anoxic brain damage. While global network architecture remained preserved, macro-level analyses revealed selective disruptions in the cingulate operculum and ventral attention networks, with opposing centrality patterns suggesting network-specific reorganization. Modified participation coefficients demonstrated widespread imbalances in integration–segregation across large-scale networks, particularly affecting perception and salience systems. Network-based statistics identified a conserved triangular hypo-connectivity pattern (anterior cingulate-orbitofrontal-temporal) alongside state-specific hyperconnectivity profiles in VS and MCS patients compared to healthy controls. Our multigraph integration revealed asymmetric reorganization: VS patients exhibited extensive compensatory hyperconnectivity, while MCS showed predominant hypo-connectivity with partial VAN-mediated compensation. These findings establish a hierarchical model of consciousness impairment, where core network disruptions interact with state-dependent compensatory mechanisms. The multiscale approach provides clinically actionable biomarkers while advancing theoretical understanding of neural correlates of consciousness, offering new avenues for targeted interventions.