Sarah has covered news topics for digital and print publications. She has a degree in broadcast journalism from the University of Nevada.
Sleep staging, the process of dividing sleep quality by different types of sleep, is hugely informative for health professionals, but the process itself is time-consuming and leaves room for error. It can also interfere with natural sleep or require hours in a lab. Now, a significant breakthrough in the field has emerged with the help of artificial intelligence.
Danish computer scientists Christian Igel and Mathias Perslev, in partnership with the Danish Center for Sleep Medicine, have developed an artificial intelligence algorithm that can improve diagnoses, treatments, and our overall understanding of sleep disorders. The software is called U-Sleep, a publicly available, automated sleep staging system that eliminates the hours it would typically take a doctor to properly diagnose a patient.
U-Sleep uses a neural network that provides information, leading to new ways of analyzing sleep stages. Using polysomnography recordings, or PSG — a comprehensive test used to diagnose sleep disorders — U-Sleep was created through the use of more than 15,000 participants and 16 clinical studies.
In conversation with Sleep Foundation, Igel recalls the birth of U-Sleep beginning with a simple question: can we automate the sleep staging process?
“By thinking about sleep staging, we thought, ‘Okay, once we automate the standard sleep staging procedure, can we perhaps even refine it and do it, for example, on faster timescales?” Igel said.
The potential benefits of this AI technology could mean major breakthroughs in sleep medicine. According to Perslev, the benefit of this automated process isn’t just for the patient — it also has real-time benefits for the doctors who spend countless hours pouring over sleep staging data.
“We hope that with this tool, we’ll reduce some of the time that the expert has to stay in front of the computer, and that time could be used on doing the more advanced stuff that we really need human experts for,” Perslev said.
In the age of technology, there is still a level of hesitancy when it comes to artificial intelligence. If a job is meant to be done by a human being, it can be challenging to turn over that process to a machine. Igel says that while the U-Sleep team takes this concern seriously, the software’s primary function is to assist in the process; not to overtake the human element.
“It’s mostly doing a monitoring job, which a human expert then has to inspect,” Igel said.
“Based off of that, a diagnosis would be made.”
“We know that having sleep issues correlates with long-term negative health conditions,” Perslev said. “There are many people who don’t get diagnosed for sleep disorders, in particular sleep apnea.”
One of the reasons U-Sleep stands out among other artificial intelligence softwares is that it is readily available and free to the public on their website. Igel says U-Sleep is meant to help those who are unable to access sleep staging on a clinical level.
“There are situations where there is no qualified expert who can actually do this analysis, simply because the person has too much to do,” Igel said. “To have a system available which is on par with the top trained experts in sleep staging may be helpful for people who can not have an expert look at their polysomnography tests.”
For Igel and Perslev, sleep staging is just the beginning of how artificial intelligence can impact the sleep field.
“There are some neurodegenerative diseases, which may be actually visible in the sleep signals, and our method may allow for an early diagnosis,” Igel said.
Getting enough sleep and improving sleep quality isn’t just a smart choice for your well-being—it’s essential to your health. The potential that U-Sleep has to improve a sleep professional’s ability to track our sleep stages may lead to larger breakthroughs in sleep disorders across the board.