Automatic diagnosis of insomnia using artificial intelligence: A narrative review

Background and aims: Numerous studies are focusing on the artificial intelligence (AI) and itsrelation to sleep medicine. Using AI in various fields of sleep researches such sleep disorders,automated scoring of polysomnography, and screening of obstructive sleep apnea syndrome arethe leading fields that has attracted the attention of researchers. One of the most common sleepdisorders is insomnia that characterized with difficulty falling asleep, staying asleep, or both.However, early diagnosis of insomnia is very challenging, but it has a great role in prevention offurther medical complications such as anger issues, heart disease, anxiety, depression and highblood pressure.Considering the tremendous capabilities of artificial intelligence, the aim of this review is to knowthe various capacities of AI that can help to diagnose insomnia automatically.Method: We performed a narrative review of the literature using the search for relevant articlesin PubMed and google scholar. Keywords include AI, sleep disorder, and insomnia and hence, weinitially found ۴۸ researches as the related studies in PubMed and Google scholar. Next, the titleand abstract of more relevant studies were screened and the full texts of ۲۲ studies were reviewed.Results: Finally ۱۲ studies satisfied the inclusion criteria. It turns out that insomnia researchesthat have sought to use AI have grown exponentially in recent years. Most of the studies used thefeatures extracted from electroencephalogram signal of polysomnography for make a predictivemodel of insomnia. Already the most researches in field of sleep medicine used AI algorithmssuch as machine learning (ML) and deep learning (DL) for prediction of differential brain activitypattern in patients with insomnia disorder and automatic prediction of Insomnia.Conclusion: Investigating the trend of using AI in the field of diagnosing sleep disorders illustratesthe unique potential of AI to play a strong role in sleep medicine to do better patient care,enhancing diagnostic abilities, reinforcing the management of insomnia, and to detect insomniaautomatically.It seems that DL has a great capability in diagnosing of insomnia. DL algorithms need more significantdata for training and testing and in the future, automated sleep disorders detection will beaddressed by DL and strong algorithms. However, it can be suggested that AI algorithms shouldbe standardized before being applied to patients in clinics.

کلیدواژه ها:

Artificial intelligence (AI) ، Electroencephalogram ، Insomnia ، Polysomnography ، Sleep disorder