Oxford, Mississippi – A team of researchers at the University of Mississippi has developed a compact chip capable of detecting heart attacks in real time. The chip, small enough to be used in wearable devices, combines advanced data analysis with artificial intelligence to deliver high detection accuracy.
This breakthrough technology merges a proven mathematical method known as Fast Fourier Transform (FFT) — commonly used to analyze electrocardiogram (ECG) signals — with artificial neural networks that classify heart rhythms as normal or indicative of a heart attack. The result is a detection system with an accuracy rate of 92.41%, outperforming many existing solutions.
Unlike current wearables that rely on cloud computing or smartphone apps to analyze ECG data, this chip processes everything locally. That means real-time results — a critical factor when every second counts in preventing severe damage or death from a myocardial infarction, commonly known as a heart attack.
Although the FFT method is widely used in the medical device industry, including by companies like Fourth Frontier, this on-chip AI-based detection system marks a new step forward. However, the chip is not yet available for clinical use and cannot currently be prescribed by doctors.
Developing a reliable, wearable heart attack detector involves several complex challenges. First, the hardware must be small and comfortable enough for everyday wear. At the same time, it must be powerful enough to accurately record ECG data within the limits of its form factor.
Then comes the task of fine-tuning the software to correctly detect key ECG changes, particularly ST elevation — a classic sign of a heart attack — while avoiding false alarms or missed warnings. Finally, the technology must pass the rigorous standards of regulatory bodies like the U.S. Food and Drug Administration (FDA), which only approve devices that meet strict safety and accuracy thresholds.
This is one reason why even popular consumer devices like the Apple Watch, which can record ECGs, are not cleared to diagnose heart attacks directly.
The researchers published their findings in March in the journal Intelligent Systems, Blockchain and Communication Technologies. They now aim to explore further applications of their AI-integrated analysis method.
While commercial use may still be years away, the University of Mississippi team’s innovation shows promise in bringing real-time heart attack detection to wearable technology — and potentially saving lives in the future.
Related topics:
