Mount Sinai-led researchers have augmented an artificial intelligence (AI)-based algorithm to analyze video recordings of clinical sleep tests, ultimately improving accurate diagnosis of a common sleep disorder that affects more than 80 million people worldwide. . The findings were published in the journal Annals of Neurology On January 9th.
REM sleep behavior disorder (RBD) is a sleep condition during the rapid eye movement (REM) stage of sleep that causes unusual movements or physical behaviors that distract from dreaming. RBD that occurs in healthy adults is called “isolated” RBD. It affects more than a million people in the United States and in almost all cases, it is an early sign of Parkinson’s disease or dementia.
RBD is very difficult to diagnose because symptoms may not be noticeable or may be confused with other diseases. A final diagnosis requires a sleep study, known as video polysomnography, performed by a medical professional in a facility equipped with sleep monitoring technology. Additionally, the data is subjective and can be difficult to universally interpret based on a variety of complex variables such as sleep stages and muscle activity. Video data is systematically recorded during sleep tests, but is rarely reviewed and is often discarded after test interpretation.
Previous limited research in this area suggested that research-grade 3D cameras may be needed to detect movement during sleep, as sheets or blankets cover the activity. This study is the first to outline the development of an automated machine learning method to analyze video recordings routinely collected with a 2D camera during a nighttime sleep test. This method also defines additional “classifiers” or features of the movement, resulting in an RBD detection accuracy of nearly 92%.
“These automated approaches can be integrated into the clinical workflow during sleep study interpretation to enhance and facilitate diagnosis and prevent missed diagnoses,” said corresponding author, Neurology (Movement Disorders) and Medicine (Pulmonology, Critical Care and Sleep). Medicine), Icahn School of Medicine at Mount Sinai. “This method may inform treatment decisions based on the severity of movements seen during a sleep test and ultimately help doctors personalize treatment plans for individual patients.”
The Mount Sinai team replicated and extended a proposal for automated machine learning analysis of movement during sleep studies created by researchers at the Medical School of Innsbruck, Austria. This approach uses computer vision, a field of artificial intelligence that allows computers to analyze and understand visual data, including images and videos. Based on this framework, Mount Sinai experts monitored patients’ sleep throughout the night using 2D cameras routinely used in clinical sleep laboratories. The data set included analysis of records from a sleep center consisting of approximately 80 patients with RBD and a control group of approximately 90 patients without RBD with or without other sleep disorders. An automated algorithm that calculates pixel motion between successive frames of video was able to detect movement during REM sleep. Experts reviewed the data to extract the speed, rate, size, velocity, and float rate of movement. By analyzing these five characteristics of short movements, they achieved 92%, the highest accuracy researchers have achieved to date.
Researchers from the École Polytechnique Fédérale de Lausanne in Lausanne, Switzerland, contributed to the study by sharing their expertise in computer vision.
