Sleep disorders, sleep apnea and narcolepsy are among a range of sleep disorders that affect thousands of Danes. In addition, it is estimated that up to 200,000 Danes go undiagnosed for sleep apnea.
In a new study, researchers from the Department of Computer Science at the University of Copenhagen worked with the Danish Center for Sleep Medicine at the Danish hospital Rigshospitalet to develop an artificial intelligence algorithm that can improve diagnoses, treatments and our general understanding of sleep disorders.
“The algorithm is exceptionally precise. We carried out various tests in which his performance could keep up with that of the best doctors in this field worldwide, ”says Mathias Perslev, doctoral student at the Institute for Computer Science and first author of the study, which was recently published in. the journal npj Digital Medicine (Link) was published.
Can support doctors with their treatments
Today’s examinations for sleep disorders usually begin with admission to a sleep clinic. The night sleep of a person is monitored with various measuring devices. A sleep disorder specialist then reviews the 7-8 hour readings from the patient’s nighttime sleep.
The doctor manually divides these 7-8 hours of sleep into 30-second intervals, all of which have to be divided into different sleep phases, such as REM sleep (Rapid Eye Movement), light sleep, deep sleep, etc. It is a time-consuming work, which the algorithm can execute in seconds.
“With this project we were able to prove that these measurements can be carried out very safely with machine learning – which is of great importance. By saving many hours of work, many more patients can be assessed and diagnosed effectively, ”explains Poul Jennum, Professor of Neurophysiology and Director of the Danish Center for Sleep Medicine.
In the capital region of Denmark alone, more than 4,000 polysomnographic tests – known as PSG or sleep studies – are carried out annually on patients with sleep apnea and more complicated sleep disorders. It takes 1.5-3 hours for a doctor to analyze a PSG study. In the capital region of Denmark alone, the use of the new algorithm could save between 6,000 and 12,000 medical hours.
The algorithm works across sleep clinics and patient groups
By collecting data from various sources, the researchers behind the algorithm were able to ensure optimal functionality. A total of 20,000 sleep nights from the USA and a large number of European countries were collected and used to train the algorithm.
“We collected sleep data from continents, sleep clinics and patient groups. The fact that the algorithm works well under such different conditions is a breakthrough, ”explain Mathias Perslev and Christian Igel, who led the project on the IT side, and add. :
“Achieving this kind of generalization is one of the greatest challenges in medical data analysis.”
They hope the algorithm will help doctors and researchers around the world learn more about sleep disorders in the future. The sleep analysis software is available free of charge at sleep.ai.ku.dk and can be used by anyone and anywhere – even in places where there is no sleep clinic around the corner.
“Only a few measurements with common clinical instruments are required for this algorithm. The use of this software could therefore be particularly relevant in developing countries, where you may not have access to the latest equipment or an expert, ”says Mathias Perslev.
The researchers are now working with Danish doctors to approve the software and algorithm for clinical use.