Multilayer network analysis of dynamic network reconfiguration in patients with moderate-to-severe obstructive sleep apnea and its association with neurocognitive function

BACKGROUND: Brain functional network disruption and neurocognitive dysfunction have been reported in obstructive sleep apnea (OSA) patients. Nevertheless, most research studies static networks, while brain evolution continues dynamically. PURPOSE: To investigate the characteristics of dynamical networks in moderate-to-severe OSA patients using multilayer network analysis of dynamic networks and compare their association with neurocognitive function. METHODS: Twenty-seven moderate-to-severe OSA patients and twenty-five matched healthy controls (HCs) who completed the examination of the Epworth sleepiness scale (ESS), neurocognitive function, polysomnography, and functional magnetic resonance imaging (fMRI) were prospectively included. The dynamic variations of resting-state functional networks in both groups were described via network switching rate. Switching rates and their correlation with clinical parameters were analyzed. RESULTS: At the global level, network switching rates were notably lower in the OSA group than in the HCs group (p = 0.002). More specifically, the differences include the default mode network (DMN), auditory network, and ventral attention network at the subnetwork level, and the right rolandic operculum, left middle temporal gyrus, and right precentral gyrus at the nodal level. Furthermore, these altered switching rates have a close correlation with ESS, sleep parameters, and neurocognitive function. CONCLUSION: Patients with moderate-to-severe OSA showed lower network switching rates, especially in the DMN, auditory network, and ventral attention network. The disruption of dynamic functional networks may be a potentially crucial mechanism of neurocognitive dysfunction in moderate-to-severe OSA patients.