3KG: Contrastive Learning of 12-Lead Electrocardiograms using Physiologically-Inspired Augmentations
This work addresses the challenge of limited labeled data for ECG diagnosis, offering an incremental improvement in representation learning for biomedical signals.
The paper tackles the problem of improving self-supervised learning for 12-lead electrocardiograms by proposing 3KG, a contrastive learning method using physiologically-inspired 3D augmentations, resulting in a 9.1% increase in mean AUC over the best baseline when fine-tuned on 1% of labeled data for 23-class diagnosis.
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 $9.1\%$ increase in mean AUC over the best self-supervised baseline when trained on $1\%$ 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.