New AI-based model predicts irregular heartbeat 30 minutes before onset

Researchers have developed a new AI-based model that can predict an irregular heartbeat or arrhythmia about 30 minutes before its onset.

The researchers found that the model proved to be 80 percent accurate in predicting the transition from a normal heart rhythm to atrial fibrillation, the most common type of heart rhythm disorder in which the upper chambers (atria) of the heart beat irregularly and out of sync with the heartbeat . the lower (ventricles).

The team, including researchers from the University of Luxembourg, said their AI model, which provides early warnings, can be easily installed on smartphones to process the data recorded on smartwatches. The alerts could allow patients to take preventative measures to keep their heart rhythms stable, they said. The research has been published in the journal Patterns.

To develop the model, the team trained it based on 24-hour recordings of 350 patients at Tongji Hospital in Wuhan, China. The model, which the researchers have called WARN (Warning of Atrial fibRillatioN), is based on deep learning, a type of machine learning AI algorithms that learn patterns from previous data to make predictions.

Deep learning is more specialized because it has multiple layers in the decision-making process.

READ ALSO| Heat wave and heart attack: what’s the connection? How to beat extreme heat and manage your heart health

The researchers found that WARN provided early warnings, on average 30 minutes before the onset of atrial fibrillation, and that this is the first method to provide an alert far from the onset, they said.

“We used heart rate data to train a deep learning model that can recognize different phases – (normal) sinus rhythm, pre-atrial fibrillation and atrial fibrillation – and calculate a ‘danger probability’ that the patient will have an impending episode ,” said Jorge Goncalves, from the Luxembourg Center for Systems Biomedicine (LCSB), University of Luxembourg, and corresponding author of the study.

As atrial fibrillation approaches, its likelihood increases until it exceeds a specific threshold, providing an early warning, Goncalves said.

Because computational costs are low, the AI ​​model is “ideal for integration into wearable technologies,” the researchers said.

“These devices can be used daily by patients, so our results open opportunities for the development of real-time monitoring and early warnings from comfortable wearable devices,” said study author Arthur Montanari, an LCSB researcher.