New home-based method may revolutionize traditional polysomnography

A University of Houston engineer is redesigning sleep studies, eliminating the need for numerous wires traditionally used. This new procedure requires only two wires and can be performed at home, eliminating the necessity to spend the night in a sleep study lab.

University of Houston associate professor of electrical and computer engineering Bhavin R. Sheth and former student Adam Jones have introduced an innovative approach to sleep stage classification that could replace the current gold standard in sleep testing, polysomnography. The traditional method involves multiple wires and is conducted in a clinic. Their new procedure uses a single-lead electrocardiography-based deep learning neural network and can be performed by users at home.

“If you’ve ever had a problem sleeping and ended up in a sleep lab, you know the polysomnography test is anything but restful,” Sheth said. “With a multitude of leads — sensors and wires — dangling from every part of your body, you are asked to sleep, which is nearly impossible with such encumbrance.”

Sheth reports in Computers in Biology and Medicine: “We have successfully demonstrated that our method achieves expert-level agreement with the gold-standard polysomnography without the need for expensive and cumbersome equipment and a clinician to score the test.” He added that this advancement challenges the traditional reliance on electroencephalography (EEG) for reliable sleep staging, paving the way for more accessible, cost-effective sleep studies.

The research by Sheth and Jones holds potential to significantly expand access to high-quality sleep analysis outside clinical settings. Reliable classification of sleep stages is crucial in providing valuable insights, diagnoses, and understanding brain states within sleep medicine and neuroscience research. While commercial devices like Apple Watch, Fitbit, and Oura Ring track sleep, their performance falls short compared to polysomnography.

The electrocardiography-based model was trained on 4,000 recordings from subjects aged 5–90 years old. It demonstrated robustness comparable to clinician-scored polysomnography.

“Our method significantly outperforms current research and commercial devices that do not use EEG,” said Sheth. “It makes less-expensive, higher-quality studies accessible to a broader community, enabling improved sleep research and more personalized, accessible sleep-related healthcare interventions.”

Jones has made the complete source code freely available for researchers, clinicians, or anyone interested at cardiosomnography.com. The collaboration also includes Laurent Itti at the University of Southern California.