Disorder of the Rem sleep behavior, i.e. RBD, is a state that causes incorrect movements or low repetitive vibrations during sleep and sporadic episodes of sleep adopting.
Challenge
RBD affects over a million Americans and is almost always the early sign of Parkinson or dementia, often preceding other symptoms by 10-15 years. This means that it is an unprecedented opportunity to develop therapy against Parkinson’s disease or dementia and ultimately identify those who would benefit from early prevention therapy.
However, RBD was very tough to diagnose.
“The simple question whether people play their dreams is poorly sensitive, because many people with RBD rarely have full episodes of dreams, but they only have small vibrations that they or their partners are not aware of,” explained the clinicist from Mount Sinai, Dr. Emmanuel.
During the professor of neurology at Icahn School of Medicine, where he sees patients in the Integrative Sleep Center Mount Sinai and specializes in movement and sleep disorders, as well as in pulmonary medicine and critical care.
Mount Sinai Health System is one of the largest academic medical systems in the Metro region in Up-to-date York, from 48,000 employees working in eight hospitals, over 400 outpatient practices, over 600 research and clinical laboratories, nursing schools and a leading school of medical and education of graduates.
“A simple screening question is also poorly specific because more common conditions – Like sleep apnea or a form of restless leg movements during sleep, it can cause symptoms of registration of imitating RBD dreams “during continuous.” RBD questionnaires have no accuracy. “
The Gold Standard RBD diagnosis test is a sleep test at LAB in the sleep center, also known as a polysomnogram that measures muscle activity (increased in RBD) during REM sleep using muscle sensors or electromyography.
But RBD was very tough to diagnose with this sleep test, because it is very tough to interpret and undergo artifacts, to such an extent that even sleep experts may not agree to the diagnosis.
“In particular, although the video camera records all movements during sleep, the current standard of interpretation of the test does not take this into account and there is currently no automated tool for interpretation of video data”, again.
“In fact, in most sleep centers, video data is rejected after the sleep test and only the rest of the collected data – EEG, breathing signals, ECG, etc. – will be stored.”
In addition, he said, unless the person undergoing a sleep test is particularly suspected of RBD by a sleep specialist – based on the history of sleep adopting – the diagnosis can be easily omitted, because in 99% of cases the test is carried out to assess sleep apnea, which does not require an assessment of muscle activity, “he added.
According to research, this leads to the missing “random cases” of RBDs, which occur in at least 1% of the adult population. To satisfy this need, Mount Sinai research team has developed a method of automating RBD diagnosis by analyzing sleep video recordings during sleep tests.
APPLICATION
The team has developed an algorithm for automatic interpretation of the frequency and pattern of body movements detected during ReM (brisk eye movement) of sleep and determine whether the person has RBD on the basis of these movements or not.
This was done only by one group in Austria, but their study used a specific 3-D research class camera, which required adding to the standard sleep test equipment.
“No study before ours tested video data routinely collected with a 2-D-infrared camera used in all clinical sleep laboratories around the world,” he said.
Fulfillment of the challenge
“We collected a large set of data – larger than in the previous study – covering 81 sleep records of patients with RBD (” cases “”) and 91 without RBD (“controls”), including 63 with various sleep disorders and 28 fit sleeping “during explanation.
Some The computer vision algorithm of the optical flow automatically detected movements during ReM sleep, from which the features of speed (frequency), ratio (interest in sleeping time showing movements), the size and speed of movement and the ratio of real estate (measure of the pattern of the distribution of movements in the San REM). Of these five functions, the machine learning classifier was trained to distinguish RBD from other sleep conditions and normal sleep.
“We were also interested in testing the accuracy of the resulting classifier in order to detect RBD in patients who actually have RBD, but during the review of their sleep test it was not reported that it was carried out, based on a manual review of video figs by a sleep expert,” during withdrawal. “11 such patients with RBD were identified, but not visible movements in the eyes were identified, and 71 patients with RBD with visible movements.”
RESULTS
The Mount Sinai team stated, as they expected, that people from the RBD showed an increased number of movements in Rem sleep, especially low movements shorter than 2 seconds, including jerking or vibration known as a myoclonus. The accuracy of RBD detection was from 84.9% (with only two functions) to 87.2% (with five features).
Combining all five functions, but analyzing only low (less than 2 second times) movements have reached the highest accuracy of 91.9%.
Of the 11 patients with RBD without noticeable movements during the sleep test, seven were correctly identified or detected as having RBD, based on the Mount Sinai algorithm.
Tips for others
“This is the first study showing that a simple algorithm analyzing video recordings purchased during sleep tests, carried out as part of routine clinical care, can diagnose RBD and with a very high accuracy of 91.9%” during noteworthy. “This work improves earlier frameworks using a 2D camera, which is routinely used in sleep laboratories and improving performance by adding only three functions.
“This approach can be implemented in clinical sleep laboratories to facilitate and improve the diagnosis of RBD,” he concluded. “In combination with the automatic detection of REM sleep, it should also be tested in the home environment, using conventional infrared cameras for RBD detection and monitoring.”
Watch now: Chief AI officer for children in Seattle says better technology results