During this years (2020) European Society of Cardiology conference Marinka Oudkerk Pool received the best abstract award e-cardiology for her presentation entitled: “Distinguishing sinus rhythm from atrial fibrillation on single-lead ECGs using a deep neural network”. In this study a residual neural network has been used to distinguish sinus rhythm from atrial fibrillation on 600 single-lead ECGs. The network has been trained on the publicly available PhysioNet dataset, after which it was tested on 600 Dutch kardia users. The algorithm achieved an accuracy of 92% and an area under the curve of 0.98.
