Brain microstructural alterations and cognitive impairment in obstructive sleep apnea: a diffusion kurtosis imaging study.

BACKGROUND

This study aimed to investigate brain microstructural alterations and their association with neurocognitive impairment in adults with moderate to severe obstructive sleep apnea (OSA) using diffusion kurtosis imaging (DKI), to better understand the neuropathological mechanisms contributing to cognitive decline in OSA.

MATERIALS AND

METHODS

40 OSA patients and 40 matched healthy controls (HCs) underwent cognitive assessments using the Montreal Cognitive Assessment (MoCA) and MRI scans including DKI and 3D T1-weighted imaging. DKI were processed to generate kurtosis parameter maps, including axial kurtosis (AK) and radial kurtosis (RK). Brain region values were extracted using SPM12. Group comparisons were conducted for cognitive scores and kurtosis values. Receiver operating characteristic (ROC) curves were used to assess the diagnostic performance of imaging biomarkers. Partial correlation analysis examined relationships between imaging metrics, cognitive scores, and sleep-related variables. Multiple comparisons were corrected using the false discovery rate (FDR) method.

RESULTS

The OSA group showed increased AK in 9 brain regions and decreased RK in 28 regions. MoCA scores, particularly in visual space and executive function, abstraction, and delayed recall, were significantly lower in the OSA group. ROC analysis showed that RK in specific brain regions had strong diagnostic accuracy for OSA (AUC = 0.817). Lower oxygen saturation (LSpO2) was associated with altered kurtosis values in key regions related to cognition. Cognitive scores were positively correlated with RK values in regions such as the frontal cortex, cingulate cortex, and hippocampus.

CONCLUSION

DKI effectively detects microstructural brain changes in OSA patients. These alterations are associated with cognitive decline, providing valuable insights into the potential mechanisms underlying neurocognitive impairments in OSA.