By Deborah Borfitz
June 22, 2023 | As anyone who has ever needed a sleep study is aware, it’s neither convenient nor fun and can easily cost several thousand dollars. Polysomnography (PSG), the gold standard for diagnosing obstructive sleep apnea, typically requires an overnight stay at a sleep disorders unit within a hospital or at a sleep center with a long wait to even get scheduled, according to W. Hong Yeo, associate professor and Woodruff Faculty Fellow in the George W. Woodruff School of Mechanical Engineering at Georgia Tech.
The testing procedure requires up to 15 wired sensors and bulky electronics to capture brain waves, blood oxygen level, heart rate, and breathing during sleep, in addition to measures of eye and leg movements. Technicians may need to wake patients if cables need fixing, or some of the sensors fall off. Coupled with the new environment, patients invariably feel less comfortable than they would at home and consequently don’t sleep as well.
Yeo became familiar with issues in measuring sleep quality and detecting disorders from Lynn Marie Trotti, M.D., associate professor of neurology at Emory University, one of his collaborators on the development of an at-home sleep monitor patch system that could help put an end to the unpleasant diagnostic odyssey. In a study recently published in Science Advances (DOI: 10.1126/sciadv.adg9671), the device had an accuracy rate of 88.5% for sleep apnea detection.
That compares to an accuracy rate of about 80% when physicians are doing the diagnosing based on a manual read of recorded PSG data, says Yeo. In addition to the potential for human error, the conventional approach may underrepresent longer-term sleep patterns since recordings are being taken for one night.
Key Advantages
The new and soft wearable platform, made of silicone and thin-film metals, is the thickness of an adhesive bandage and includes two patches that attach comfortably to the forehead and chin, he says. Three embedded sensors wirelessly measure signals of brain, eye, and muscle activity—electroencephalogram, electrooculogram, and electromyogram, respectively—which are comparable to the data recorded by a PSG test at a sleep clinic.
The data gets sent to a smartphone app to automatically calculate sleep quality and detect sleep apnea, continues Yeo. Machine learning (convolutional neural networks) embedded in the portable sleep patches is better at sleep scoring with PSG, based on a standard performance yardstick (confusion matrix) used in the latest study.
Given its portability and real-time detection capabilities, the wearable system has the added advantage over PSG of detecting sleep apnea before it is severe—and even before symptoms are apparent, he says. And because data doesn’t have to be sent to a specialist for deciphering, significant time and cost savings can be realized.
According to the American Association of Sleep Medicine (AASM), obstructive sleep apnea in 2016 was afflicting an estimated 12% of the adult population in the U.S. (Journal of Clinical Sleep Medicine, DOI: 10.5664/jcsm.6034) with a high prevalence of undiagnosed cases. In addition to snoring and arrested breathing episodes, the afflicted often wake up suddenly, gasping or choking, and have high blood pressure.
The device was initially developed by Yeo with the idea of being integrated with other sleep sensors to record cardiac vibrations, heart rhythms, and blood volume changes from the sternum (Science Advances, DOI: 10.1126/sciadv.abl414), as a way for clinicians to investigate the cardiovascular response to sleep apnea during home sleep tests, he says. But that limited the ability to pick up signals from other motion activities of interest to clinicians, so efforts shifted to a wearable that could be mounted onto the face and used in conjunction with the soft sternal patch using Bluetooth to deliver data to a tablet or smartphone wirelessly.
It is now estimated that sleep apnea affects more than 18 million people, says Yeo, or roughly one in every 18 Americans. And those numbers have been increasing in time, in part due to the rise in recognition but also because of the obesity epidemic and long-term stress that can trigger overeating.
Sleep Tech Explosion
Yeo says he began developing a range of biomedical devices for home health monitoring after his father passed away suddenly in his sleep from an undetected heart issue. At the time, Yeo was serving in the military in South Korea as an ordnance officer.
Among other projects underway by Yeo and his group is a portable heart monitor where a simple screen-printing method is planned for large-scale production of the circuit boards and wireless sensors. A soft, flexible brain monitor is also under development, which is covered in tiny microneedles that can go around human hairs to maximize contact between electrodes and the scalp, lower electrical impedance, and capture more accurate measurements.
Although a vast number of smartphone apps are already available for sleep monitoring purposes, none of them capture all the data clinicians need to make a sleep apnea diagnosis, Yeo says, noting that doctors currently abide by the more rigorous diagnostics criteria set by the AASM. A headband device already on the market (product of Dreem) cannot measure eye movement and muscle activity and has a comparatively low accuracy rate due to motion artifact and the device’s rigidity.
It is but one of a slew of devices that have emerged over the past few years allowing patients to test for the disorder at home and they similarly fall short in providing all the AASM-required data, says Yeo. Bresotec Medical’s at-home sleep apnea test (BresoDX1) received 510(k) clearance from the Food and Drug Administration (FDA) earlier this year, and is based primarily on respiratory airflow. Belgian startup Sunrise also recently received FDA 510(k) clearance for its home sleep apnea test that utilizes a sensor analyzing bio-signals from mandibular jaw movements indicative of respiratory disturbances.
In 2021, UK-based Acurable announced that it received FDA clearance for a device that prescreens for sleep apnea by sensing physiological sounds generated by the body. And last year, the FDA gave the green light to RedApp Health’s SleepCheckRx prescription-only software device that screens for the condition by analyzing breathing and snoring sounds recorded via an iPhone.
Clinical Study Begins
Yeo and his team are now working on further improving the machine learning algorithm of their wearable platform to bring sleep apnea detection accuracy closer to 100%, he reports. They are also looking to improve the reusability of the sensor’s dry electrodes, which can already be used for multiple days of sleep recording—another advantage over gold-standard PSG requiring one-time-use gel electrodes.
A clinical study has been launched to validate the efficacy of at-home sleep monitoring and automated disease diagnosis, he says, and currently has over 30 enrollees in Atlanta and other areas. Participation is expected to expand internationally with greater financial support for recruitment efforts and device manufacturing.
Yeo is cofounder and chief technology officer of Wis Medical, which spun out of Georgia Tech last year and has a branch in South Korea as well as the U.S. The company will be seeking FDA 510(k) clearance for marketing the wireless sleep monitoring patch system, he says.