In the recent past, it’s become easier to detect heart conditions with technology. The Apple Watch has become pretty good at detecting arrhythmia for instance.
But some researchers have been developing AI to detect heart problems, and one team may have the best version yet.
According to a recent study published in the Biomedical Signal Processing and Control Journal, a team of researchers from the Universities of Surrey, Warwick and Florence have a new neural network that can detect cardiac anomalies from a single heartbeat with 100% accuracy.
Their AI can quickly and accurately detect congestive heart failure (CHF) by analyzing one heartbeat on an electrocardiogram (ECG). CHF is a chronic progressive condition that affects how blood is pumped around the body. At least 5 million people live with the condition, and that’s just in the US. And we need better ways to diagnose the condition, because there’s a high mortality rate if left unchecked, not to mention treatment is very expensive.
The AI uses convolutional neural networks (CNN) advanced signal processing and machine learning to analyze raw ECG signals. “First, by assessing the ECG directly, we confirm that with AI it is possible to accurately detect CHF looking beyond heart rate variability analysis. Thus, we have in general results that are more adherent to the real behavior of the affected heart,” said Dr Sebastiano Massaro at the University of Surrey.
Because of this they take a much shorter period of time to study a person’s heart, meaning diagnosis is faster. This is especially important if a person is in the hospital in a severe state and doctors don’t have 24 hours to observe their heart.
The team is currently working on the technology, hoping to eventually apply it to other more everyday healthcare systems, and make the technology more accessible in hospitals around the world.