Investigation of the effect of standard polysomnographic parameters on the diagnosis of sleep apnea using machine learning methods

Obstructive Sleep Apnea (OSAS), which is one of the sleep diseases that is most confused with snoring, can result in death when not treated medically without early diagnosis. Obstructive Sleep Apnea is a health problem that causes problems such as interruption of sleep, decrease in sleep quality, paralysis as a result of complete or partial closure of the respiratory tract during sleep, and even serious problems that may result in death. The prevalence of OSA in Turkey from 0.9% to 1.9%. The prevalence of OSAS disease is related to variables such as age, neck diameter, cigarette-alcohol consumption, and gender. Polysomnography signal recordings (PSG) are used in sleep laboratories for the diagnosis of sleep apnea. Polysomnography is a test used in the diagnosis and treatment of sleep disorders such as snoring, sleep apnea, parasomnia (abnormal behavior during sleep), narcolepsy (sleep attacks that develop during the day), restless leg syndrome. It is done by recording various parameters such as brain waves, eye movements, heart and chest activity, respiratory activities, and the amount of oxygen in the blood with the help of electrodes placed in different parts of the body during night sleep. In this study, information obtained from sleep laboratories, sleep apnea, sleep stages affected by sleep apnea, PSG data, and engineering studies and our research are presented. In our study, signal data were obtained from healthy individuals and individuals with sleep apnea. These data are classified and given to the artificial neural network to reveal their effects on the diagnosis of sleep apnea. The effects of sleep apnea on sleep stages and REM sleep are tried to be proven by studies in the field of engineering. By conducting the necessary literature reviews, a conclusion was drawn about what different studies can be done with the studies we have completed. More innovative and not only clinically concluded studies, but also other studies are envisaged. Since clinical studies are mostly based on physician observations and predictions, more engineering studies are required for these processes to be more scientific.