Automated Prediction of Spontaneous Initiation and Termination of Atrial Fibrillation from the ECG

D. Hayn and G. Schreier (Austria)


ECG analysis, atrial fibrillation, and risk assessment


A software system capable to automatically predict the initiation of atrial fibrillation (AF) by analyzing the electrocardiogram (ECG) prior to AF's appearance, and to predict, whether AF will terminate spontaneously has been developed. Prediction of the initiation of AF was done by detecting premature atrial beats and analyzing the variations of their P waves. AF classification according to "terminating" and "none-terminating" was done by eliminating ventricular components of the ECG and by comparing the major frequency of the remaining signal with a certain threshold value. Both algorithms were validated by participating at the Computers in Cardiology Challenges 2001 and 2004.

Important Links:

Go Back