Quality sleep is essential to survival, yet it remains a scientific mystery despite spending a third of our lives in slumber. Sleep analysis, known as polysomnography, is typically used to diagnose sleep disorders by recording various data types including brain and heart activity. Recently, researchers at the University of Southern California have developed an approach that matches the performance of polysomnography using just a single-lead echocardiogram. This open-source software allows for the creation of a low-cost, DIY sleep-tracking device. The model, which assesses sleep stages, outperforms other non-EEG models and commercial sleep-tracking devices, promoting accessibility and affordability in sleep analysis.
Published in the journal Computers in Biology and Medicine, the study was co-authored by computer science professor Laurent Itti, along with lead author Adam Jones and collaborator Bhavin R. Sheth. The researchers found that sleep, a predictor of cognitive decline, becomes shorter and more fragmented with age. Chronic poor sleep can also contribute to the accumulation of beta-amyloid plaques, a hallmark of Alzheimer’s disease. The researchers trained their model on a diverse dataset ranging in age from 5 to 90 years old, using only heart data and a deep-learning neural network. The automated ECG-only network successfully categorized all five stages of sleep, showcasing the importance of sleep for memory consolidation and overall health.
This insight into the connection between the heart and the brain could lead to new interventions and treatments for sleep-related issues. The autonomic nervous system, which links the brain and the heart, may play a more significant role in sleep than previously understood. By simplifying the typically costly and cumbersome process of monitoring sleep, the research aims to improve understanding and accessibility to sleep analysis. The use of ECG data in sleep studies could also shed light on the origins and functions of sleep, particularly in remote populations. Further research is planned to explore the hidden information within the heart data and how it impacts sleep.
The impact of interrupted sleep in one’s 30s and 40s on memory problems a decade later was highlighted in a recent study, emphasizing the importance of addressing sleep issues early on. The researchers’ neural network found a decline in sleep earlier than expected with age, indicating the urgency for interventions and accessible solutions. The researchers’ software could help individuals better understand their nightly sleep patterns and make informed decisions about their sleep health. By developing a model that matches the performance of expert-scored polysomnography using just a single-lead echocardiogram, the researchers are paving the way for more accessible and affordable sleep analysis methods for individuals of all ages.
Overall, the research conducted by the USC computer science researchers has shown promising results in the development of a simpler and cost-effective method for monitoring sleep using just a single-lead echocardiogram. By training a deep-learning neural network to assess sleep stages based on heart data, the researchers have highlighted the significant role the heart plays in sleep analysis. This breakthrough could lead to improved interventions for sleep-related issues and a better understanding of the connection between the heart and the brain. The software developed by the researchers has the potential to revolutionize the field of sleep analysis, providing valuable insights into individuals’ sleep patterns and overall health.