In a recent research, an AI model was designed to detect irregular heartbeats 30 minutes before they occur

A new AI-based model has been created by researchers that has the ability to anticipate cardiac arrhythmia, or irregular heartbeat, around half an hour before it manifests.

The model, according to the researchers, was 80% accurate in foretelling the change from a normal cardiac rhythm to atrial fibrillation, the most prevalent kind of cardiac arrhythmia in which the ventricles and atria of the heart beat at different rhythms.

The group, which included University of Luxembourg experts, said that processing the data from smartwatches could be done simply with the installation of their AI-model in smartphones, which provides early warnings.

According to them, the alerts could enable patients to take preventative action to maintain a steady heart rhythm. The work appears in the Patterns journal.

The group recorded 350 patients for 24 hours at Tongji Hospital in Wuhan, China, in order to train the model.

The model is built on deep learning, a kind of machine learning AI algorithm that identifies patterns from historical data to provide predictions. The researchers have termed this sort of AI model WARN (Warning of Atrial FibRillatioN).

Because deep learning uses numerous layers to make decisions, it is highly specialized. WARN is the first technique to deliver a warning far from initiation, according to the researchers, who also discovered that WARN provided early warnings, on average 30 minutes before the onset of atrial fibrillation.

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

According to Goncalves, the chance grows as atrial fibrillation approaches and continues to do so until it reaches a certain threshold, acting as an early warning.

The AI-model is “ideal for integration into wearable technologies” because of its minimal processing cost, according to the researchers.

According to study author and LCSB researcher Arthur Montanari, “these devices can be used by patients on a daily basis, so our results open possibilities for the development of real-time monitoring and early warnings from comfortable wearable devices.”