Classification of severe obstructive sleep apnea with cognitive impairment using degree centrality: A machine...

Area of Science:

• Neuroimaging
• Machine Learning
• Sleep Medicine

Background:

• Obstructive sleep apnea (OSA) is linked to cognitive impairment.
• Mild cognitive impairment (MCI) presents challenges in diagnosing OSA's neurological impact.

Purpose of the Study:

• To utilize voxel-level degree centrality (DC) features and machine learning to differentiate OSA patients with and without MCI.
• To explore neuroimaging evidence for cognitive impairment in OSA.

Main Methods:

• rs-MRI scans from 99 OSA patients (51 with MCI, 48 without).
• Extraction and selection of degree centrality (DC) features using AAL brain atlas and LASSO regression.
• Classification models built using Support Vector Machine (SVM), Random Forest, and Logistic Regression.

Main Results:

• Support Vector Machine achieved the highest classification efficiency with an AUC of 0.78.
• Random Forest (AUC = 0.71) and Logistic Regression (AUC = 0.77) also showed classification capabilities.
• Selected DC features effectively distinguished between OSA patients with and without MCI.

Conclusions:

• Machine learning models, particularly SVM, can effectively differentiate OSA patients with and without MCI.
• Degree centrality features from rs-MRI offer potential neuroimaging biomarkers for OSA-related cognitive impairment.