A new innovation in sleep studies is reshaping the traditional way of conducting sleep tests. This approach aims to simplify the process by allowing individuals to perform the study at home with just two wires, eliminating the need for the cumbersome array of wires typically used in sleep labs.
Professor Bhavin R. Sheth and former student Adam Jones from the University of Houston have introduced an innovative method for classifying sleep stages that could potentially replace the current standard practice of polysomnography. Polysomnography involves the use of numerous wires and sensors and is usually conducted in a clinical setting. The new approach, however, utilizes a single-lead electrocardiography-based deep learning neural network that can be easily used by individuals in their own homes.
Traditionally, undergoing a polysomnography test in a sleep lab is far from a relaxing experience, with multiple electrodes and wires attached to various parts of the body. The new method simplifies this process by reducing the number of electrodes to just two.
The research conducted by Sheth and Jones has shown that their approach achieves results comparable to polysomnography without the need for complex and expensive equipment or the presence of a clinician to interpret the results. This advancement challenges the reliance on electroencephalography (EEG) for accurate sleep staging, paving the way for more accessible and cost-effective sleep studies.
Furthermore, by enabling high-quality sleep analysis outside of clinical settings, this research has the potential to broaden the reach of sleep medicine significantly. The accurate classification of sleep stages is essential for diagnosing sleep disorders and understanding brain states in both medical and research contexts.
The electrocardiography-based model used in the study was trained on a dataset of 4000 recordings from subjects of various age groups. The findings demonstrate that the model’s performance is on par with that of a clinician scoring polysomnography.
Sheth highlights that their method surpasses current research and commercial devices that do not utilize EEG, achieving gold-standard levels of agreement using only a single lead of electrocardiography data. This breakthrough allows for more affordable and higher-quality sleep studies, promoting enhanced sleep research and personalized healthcare interventions.
To encourage further research and collaboration, Jones has made the complete source code available for free to interested parties at https://cardiosomnography.com. The project also involves the collaboration of Laurent Itti from the University of Southern California.
