Bridging mathematical models and artificial intelligence for ECG analysis

SciML4Medicine aims to improve the analysis of electrocardiograms (ECGs) by combining different approaches: bespoke mathematical modeling and data-hungry AI models. AI models have the power to revolutionize health care by discovering new patterns in ECGs relevant for medical practice but need much data. Leveraging existing mathematical models of ECGs will make AI models less data-hungry and more robust. Conversely, augmenting traditional ECG analysis approaches with AI will lead to new fundamental insights. For this a new consortium of commercial partners (Cordys Analytics, ECG-Excellence) and the UMC Utrecht was established.

In the Netherlands roughly one in four people die of cardiovascular disease. Because ECGs are a cornerstone of cardiovascular, SciML4Medicine aims to improve ECG’s diagnostic utility and generate new insights in cardiovascular disease, leading to better cardiovascular health care and less unneeded loss of (healthy) life-years.

Mathematical modeling and data-driven AI methods are disparate approaches to ECG analysis that both have merit. By bringing together experts in both fields, SciML4Medicine will bring out the best of both worlds, making AI methods less data-hungry and more robust, and leading to new fundamental insights for ECG analysis.

The primary aim of SciML4Medicine is deepening and expanding scientific knowledge through research and publications., ultimately forming the building blocks of new technologies and products useful for cardiovascular health care.