Greater hypoxic burden predicts weaker coordination between brain pulsation and CSF flow on 7T MRI independent of...

Obstructive sleep apnea (OSA) is a risk factor for neurodegeneration, and glymphatic impairment may be one mechanistic pathway. Anti-phase coordination between brain pulsations and CSF flow, reflecting compensatory CSF displacement during each vascular pulsation cycle, supports glymphatic activity. This study examined whether hypoxic burden and sleep fragmentation, two distinct OSA pathologies, are differentially associated with brain pulsation-CSF flow dynamics.

This cross-sectional study included 28 individuals with newly identified OSA and 8 without OSA. Participants completed in-lab polysomnography or WatchPAT testing, providing measures of hypoxic burden, quantified as time below 90% oxygen saturation (T90), and non-hypoxic sleep fragmentation, quantified as respiratory effort-related arousals (RERAs). Participants also completed 7T resting-state functional MRI to estimate global BOLD-CSF, defined as anti-phase cross-correlation between blood-oxygen-level-dependent (BOLD) signal and CSF inflow. Associations were examined using correlation and hierarchical regression. Exploratory analyses examined region-specific BOLD-CSF and brain pulsation strength, quantified as BOLD amplitude.

Greater T90 was associated with weaker global BOLD-CSF, independent of RERAs and covariates (β=0.08,p=0.03). Greater T90 was also associated with higher BOLD amplitude across temporal, frontal, and parietal regions, but this elevation in amplitude was not accompanied by stronger region-specific BOLD-CSF coupling (β=0.001,p>0.05). In contrast, among regions where BOLD amplitude was not associated with T90, greater BOLD amplitude predicted stronger region-specific BOLD-CSF (β=-0.004,p<0.001). RERAs were not associated with global BOLD-CSF or BOLD amplitude.

In OSA, hypoxic burden may be the primary feature associated with impaired brain pulsation and CSF dynamics that support glymphatic activity. These alterations may be pronounced in the temporal lobe, where elevated pulsations were uncoupled from compensatory CSF displacement.