François Marelli

Atrial Fibrillation (AF) is the second leading cause of death from heart disease, affecting between 1 and 2% of the worldwide population aged over 35. AF episodes are characterised by a highly irregular cardiac pulse, leading to dangerous complications such as stroke and heart failure. Although effective prevention methods exist, such as pacing techniques and antiarrhythmic oral drugs, there lacks a practical system to predict the onset of AF episodes sufficiently ahead to trigger these preventive measures. The aim of this research is to develop an intelligent forecasting Holter monitor that will be continuously recording electrocardiogram signals and forecasting AF episodes in real time, using machine learning. This wearable device would help suppress arrhythmia episodes for high-risk patients, improving their quality of life and reducing healthcare costs. We aim to demonstrate the practical applications of our results by performing proof-of-concept clinical tests of our prototype.