English

3KG: Contrastive Learning of 12-Lead Electrocardiograms using Physiologically-Inspired Augmentations

Medical Physics 2021-09-22 v2 Machine Learning Signal Processing

Abstract

We propose 3KG, a physiologically-inspired contrastive learning approach that generates views using 3D augmentations of the 12-lead electrocardiogram. We evaluate representation quality by fine-tuning a linear layer for the downstream task of 23-class diagnosis on the PhysioNet 2020 challenge training data and find that 3KG achieves a 9.1%9.1\% increase in mean AUC over the best self-supervised baseline when trained on 1%1\% of labeled data. Our empirical analysis shows that combining spatial and temporal augmentations produces the strongest representations. In addition, we investigate the effect of this physiologically-inspired pretraining on downstream performance on different disease subgroups and find that 3KG makes the greatest gains for conduction and rhythm abnormalities. Our method allows for flexibility in incorporating other self-supervised strategies and highlights the potential for similar modality-specific augmentations for other biomedical signals.

Keywords

Cite

@article{arxiv.2106.04452,
  title  = {3KG: Contrastive Learning of 12-Lead Electrocardiograms using Physiologically-Inspired Augmentations},
  author = {Bryan Gopal and Ryan W. Han and Gautham Raghupathi and Andrew Y. Ng and Geoffrey H. Tison and Pranav Rajpurkar},
  journal= {arXiv preprint arXiv:2106.04452},
  year   = {2021}
}

Comments

11 pages, 3 figures, paper revision with new set of experiments and comparison to previous methods