Engineers develop wearable heart attack detection technology

https://news.ycombinator.com/rss Hits: 12
Summary

Researchers are developing a lightweight, efficient chip that can be implanted on wearable devices to detect heart attacks in real time. Detecting a heart attack as it happens can save patients critical time in getting lifesaving treatment. Credit: John McCustion / University Marketing and Communications Every second counts when it comes to detecting and treating heart attacks. That's where a new technology from the University of Mississippi comes in to identify heart attacks faster and more accurately than traditional methods. In a study published in papers from the International Conference on Intelligent Systems, Blockchain and Communication Technologies, electrical and computer engineering assistant professor Kasem Khalil shows that a new technology developed at his lab could improve heart attack detection methods without sacrificing accuracy. "For this issue, a few minutes or even a few extra seconds is going to give this person the care they need before it becomes worse," Khalil said. "Compared to traditional methods, our technology is up to two times faster, while still highly accurate. Our target was not only to increase performance for classifying heart attacks. We are also focusing on the design. If we want to make this device a usable machine for any person, that means it has to be something lightweight and economical." In the United States, someone dies from a heart attack every 40 seconds. Heart disease—a collection of underlying conditions that can lead to a heart attack—is the leading cause of death in the United States. Khalil and his team used artificial intelligence and advanced mathematics to design a chip that can analyze electrocardiograms, known as ECGs—graphs of the heart's electrical signals—and detect a heart attack in real-time. The resulting technology is lightweight and energy-efficient enough to be embedded in wearable devices while still being 92.4% accurate—higher than many current methods. "We wanted to be able to implement this in a w...

First seen: 2025-05-11 05:21

Last seen: 2025-05-11 16:23