AI-Powered Models Enhance Cardiovascular Disease Detection with Digital Stethoscope Biosignals


Summary:

Foundation models using heart sounds and ECG data improve cardiovascular disease detection.

Title: Foundation models for cardiovascular disease detection via biosignals from digital stethoscopes

Authors: George Mathew, Daniel Barbosa, John Prince, Subramaniam Venkatraman

Journal: npj Cardiovascular Health

 

Eko has applied a novel AI approach, similar to that used for Large Language Models (LLMs), by pre-training models on millions of unlabeled heart sounds and ECG signals, then refining them on smaller labeled datasets. This is the first time such techniques have been used for synchronously-captured PCG and ECG data.

View Research -->