Machine learning algorithms—also known as artificial intelligence (AI)—can better detect sleep apnea compared with traditional linear approaches, according to a study being presented at the CHEST Annual Meeting 2019.
The researchers included 620 patients who were referred to a sleep lab in a suburban community sleep center. Researchers collected information on 12 select parameters: height, weight, waist, hip, body mass index, age, neck side, Modified Friedman stage, snoring, Epworth sleepiness scale, sex, and daytime sleepiness.
During phase I, researchers used a binary particle swarm optimization technique to select the best sub-features that characterize sleep apnea. In phase II, they built an artificial neural network model based on a feedforward algorithm to detect sleep apnea.
Four models were developed, including two models for each sex with and without feature selection. A set of 93 different experiments were performed.
The researchers found that neck and daytime sleepiness were the most valuable features—they were selected 86% and 87%, respectively, of the time by binary particle swarm optimization in all experiments.
Modified Friedman stage was weakly associated with sleep apnea—it was chosen 44% of the time in the 93 experiments.
The detection ratio obtained by the AI model was 80% for men and 75% for women. The researchers noted that benefits of machine learning include the ability to discover the hidden information by feature selection methods and the ability to reduce the computational time to achieve the final results.
“Algorithms based on machine learning can help reduce the risk of sleep apnea by predicting [it] in an early stage,” the researchers concluded.
Surani S, Sheta A, Turabieh H, et al. Diagnosis of sleep apnea using artificial neural network and binary particle swarm optimization for feature selection. Presented at the CHEST Annual Meeting 2019. October 19-23, 2019; New Orleans, Louisiana.