Health · Medical Research · 2 hrs ago
AI Uses Sleep ECGs to Predict Long-Term Cardiovascular Risk
Researchers led by scientists at Emory University found that heart signals during sleep may reveal future cardiovascular risk.
They trained an AI system using sleep-study data from nearly 16,000 patients, then tested it on data from about 23,000 more.
The system used one ECG lead recorded on the chest to estimate the next 10 years’ risk of an irregular heartbeat, stroke, heart failure or death.
It also predicted heart attack risk for one group, but not consistently across the study.
The model added information beyond established risk factors such as age, diabetes and smoking, and higher scores were linked to higher risks of several outcomes.
Sleep studies already record heart activity, but that information is not usually examined unless it signals an urgent problem.
Researchers say using these data could help identify people at risk earlier, but whether it improves standard risk prediction is the next question.
An artificial intelligence system used a single ECG lead recorded during sleep to predict cardiovascular risks over the next 10 years.
The study found that its risk score improved predictions of cardiac outcomes beyond established risk factors.
Researchers trained the system using data from nearly 16,000 patients and tested it on two other patient datasets.
The model predicted risks including atrial fibrillation, stroke, heart failure and death from any cause.
The findings were published in the journal SLEEP.
- Who
- Researchers led by scientists at Emory University developed the system.
- What
- An AI system used sleep ECG data to predict long-term cardiovascular risks.
- When
- The findings were published on October 21, 2026. The system predicted risks over the next 10 years.
- Where
- The research used patient data from Massachusetts General Hospital, Emory University Hospital and Beth Israel Deaconess Medical Center.
- Why
- The researchers aimed to identify people at risk of cardiovascular disease before clinical symptoms arise.
This story does not have two clearly opposing sides.
Researchers trained a deep learning system using data from nearly 16,000 patients enrolled in a sleep study at Massachusetts General Hospital.
Researchers tested the system on datasets including nearly 10,000 patients from Emory University Hospital and 13,000 from Beth Israel Deaconess Medical Center.
The study findings were published in the journal SLEEP.
- Training dataset
- Nearly 16,000 patients
- Emory test dataset
- Nearly 10,000 patients
- Beth Israel Deaconess test dataset
- 13,000 patients
- ECG recording
- One lead placed on the chest during sleep
- Prediction period
- 10 years
What people are saying
This new approach has the potential to identify individuals at risk for cardiovascular disease years before clinical symptoms arise
Importantly, we found that adding the score from our deep learning model improved prediction of cardiac outcomes beyond established risk factors
Adding these scores into sleep studies has great potential for improving early detection of cardiovascular disease, determining who is most at risk, and being more proactive in patient care.