RESUMEN
Objective: To analyze dynamic functional connectivity (dFNC) states and influencing factors of brain network in male patients with obstructive sleep apnea (OSA). Methods: A total of 111 male patients diagnosed with obstructive sleep apnea or presenting with simple snoring, who visited the Sleep Clinic at the Second Affiliated Hospital of Soochow University between August 2020 and December 2021, were prospectively selected for this study. General information was collected, and polysomnography (PSG) was performed. Based on the oxygen desaturation index (ODI), the participants were divided into three groups: primary snoring group (ODI<5 events/hour, n=34), mild to moderate OSA group (5 events/hour≤ODI<30 events/hour, n=43), and sever OSA group (ODI≥30 events/hour, n=34). Cognitive function was assessed using the Montreal Cognitive Assessment (MoCA) scale, and daytime sleepiness was evaluated using the Epworth Sleepiness Scale (ESS). Resting-state functional magnetic resonance imaging (fMRI) data were collected and preprocessed. dFNC matrices were constructed using a sliding time window approach. The number of dFNC states was determined using k-means clustering analysis. Three parameters, namely, fractional time (FT), mean dwell time (MDT), and number of transitions (NT), were used to characterize the temporal properties of dFNC states. Differences in the temporal properties of dFNC states among the groups were compared. The correlations between temporal properties and PSG parameters, as well as MoCA and ESS scores, were further analyzed. Multiple stepwise linear regression analysis was performed to identify the influencing factors of the temporal properties of dFNC states. Results: The age of the patients was (40.2±8.6) years (range: 25-65 years). There were no significant differences in age, smoking history and alcohol history, and MoCA scores among the three groups (all P>0.05). Three dFNC states were extracted through k-means clustering analysis: state 1, characterized by strong connections within the visual and sensorimotor networks with a frequency of 31.7% (4 611/14 541); state 2, characterized by strong connections within the default mode network, attention network, and other cognitive networks, with the lowest frequency of 22.1% (3 213/14 541); state 3, characterized by weaker connections across the whole brain, with the highest frequency of 46.2% (6 717/14 541). The FT [0.28 (0.05, 0.35) vs 0.39 (0.26, 0.53)] and MDT [8.20 (4.35, 12.54) vs 11.68 (8.50, 16.69)] of state 2 in the sever OSA group were lower than those in the primary snoring group (both P<0.05), while there were no significant differences in the temporal properties of states 1 and 3 among the three groups (all P>0.05). The FT and MDT of state 2 were correlated with body mass index (BMI), apnea-hypopnea index (AHI), ODI, and minimum oxygen saturation (MinSaO2) (FT: r values were -0.218, -0.230, -0.249, 0.198, respectively; MDT: r values were 0.269, -0.253, -0.265, 0.209, respectively; all P<0.05). There were no significant correlations between the temporal properties and MoCA or ESS scores (all P>0.05). ODI was found to be an influencing factor for the temporal properties of state 2 (FT: ß=-0.225, 95%CI:-0.227 to -0.223; MDT: ß=-0.241, 95%CI:-0.289 to -0.195). Conclusions: Male patients with OSA exhibit alterations in specific temporal properties of brain network dynamic functional connectivity, which are associated with nocturnal oxygen parameters. This may be one of the mechanisms underlying brain functional damage in patients with OSA.
