Advertisement
Telegram ads for your business · TCA Media
RUUZEN
Advertisement
Telegram ads for your business · TCA Media
CBU ratesUSD 12 023,61EUR 13 236,79RUB 145,54Tuesday, 17 November 2026
Technology

AI detects hidden heart signal that may predict sudden death

Scientists at UC Berkeley identified a signal in routine ECGs linked to the risk of sudden cardiac death. AI recognized it more accurately than standard screening.

The algorithm was trained on over 440,000 electrocardiograms from Sweden, matching electrical indicators with death certificate data. It was then tested on medical records of patients from the U.S. and Taiwan to assess whether the model works beyond the original sample.

Standard risk assessment typically relies on how much blood the heart pumps with each contraction. This approach identifies a group of patients with an annual risk of sudden cardiac death of 4.6%. The ECG-based model identified a group with a 7% rate and simultaneously detected a broader category of people at elevated risk.

Most individuals in this category would be considered relatively low-risk during routine checks. Meanwhile, the algorithm requires no new examinations: it analyzes images of standard ECGs, which are regularly performed in hospitals and medical centers.

The results were published in *Nature*. In the future, the method could help doctors decide which patients should receive an implantable defibrillator or additional monitoring. Currently, such devices are given to people who ultimately do not need them, while some at-risk patients remain undetected. Outside of hospitals, about 90% of people die within minutes after sudden cardiac arrest.

Advertisement
Telegram ads for your business · TCA Media