Heart failure is among medicine's quieter crises — millions carry it without knowing, and the tools to catch it early have long been out of reach for many. Researchers at Wake Forest University School of Medicine have trained an artificial intelligence model on more than a million electrocardiograms, teaching it to recognize three distinct forms of heart dysfunction — including the notoriously elusive preserved ejection fraction type — from a test already present in nearly every clinical setting. The model works nearly as well from a single electrical lead, the kind a smartwatch can capture, s
AI Model Detects Multiple Heart Failure Types from Routine ECGs
Early detection enabled by this AI tool could reduce hospitalizations and deaths among the 6+ million Americans with heart failure.