Artificial Intelligence for Diagnosis of Obstructive Sleep Apnea

This review focuses on the different methodologies and artificial intelligence (AI) technologies being utilized to improve rapid diagnosis of sleep disordered breathing (SDB).

Many recent studies have examined the application of AI neural networks on imaging techniques, single lead diagnostics, and multi-lead technologies to improve screening and automated diagnosis of SDB. While several techniques are better at identifying moderate or severe sleep apnea, the ability to detect mild cases of sleep apnea still lags. A major limitation of these studies is the small and homogenous patient populations in training datasets, limiting generalization and thus application.

Application of AI in the diagnosis of SDB is promising and great advances have been made. Further research is needed to develop more generalizable techniques before widespread application.