Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Finding "Good Views" of Electrocardiogram Signals for Inferring Abnormalities in Cardiac Condition

Hyewon Jeong, Suyeol Yun, Hammaad Adam

TL;DR

This project investigates several ways to define positive samples, and assess which approach yields the best performance in a downstream task of classifying arrhythmia, and finds that learned representations invariant to patient identity are powerful in arrhythmia detection.

Abstract

Electrocardiograms (ECGs) are an established technique to screen for abnormal cardiac signals. Recent work has established that it is possible to detect arrhythmia directly from the ECG signal using deep learning algorithms. While a few prior approaches with contrastive learning have been successful, the best way to define a positive sample remains an open question. In this project, we investigate several ways to define positive samples, and assess which approach yields the best performance in a downstream task of classifying arrhythmia. We explore spatiotemporal invariances, generic augmentations, demographic similarities, cardiac rhythms, and wave attributes of ECG as potential ways to match positive samples. We then evaluate each strategy with downstream task performance, and find that learned representations invariant to patient identity are powerful in arrhythmia detection. We made our code available in: https://github.com/mandiehyewon/goodviews_ecg.git

Finding "Good Views" of Electrocardiogram Signals for Inferring Abnormalities in Cardiac Condition

TL;DR

This project investigates several ways to define positive samples, and assess which approach yields the best performance in a downstream task of classifying arrhythmia, and finds that learned representations invariant to patient identity are powerful in arrhythmia detection.

Abstract

Electrocardiograms (ECGs) are an established technique to screen for abnormal cardiac signals. Recent work has established that it is possible to detect arrhythmia directly from the ECG signal using deep learning algorithms. While a few prior approaches with contrastive learning have been successful, the best way to define a positive sample remains an open question. In this project, we investigate several ways to define positive samples, and assess which approach yields the best performance in a downstream task of classifying arrhythmia. We explore spatiotemporal invariances, generic augmentations, demographic similarities, cardiac rhythms, and wave attributes of ECG as potential ways to match positive samples. We then evaluate each strategy with downstream task performance, and find that learned representations invariant to patient identity are powerful in arrhythmia detection. We made our code available in: https://github.com/mandiehyewon/goodviews_ecg.git

Paper Structure

This paper contains 16 sections, 1 equation, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Approach
  3. Self-Supervised Contrastive Learning
  4. Defining Transformation-Invariant Positive Sample
  5. Evaluating learned representation on arrhythmia classification
  6. Experiments
  7. Dataset
  8. Experimental Results
  9. Conclusion
  10. Acknowledgements and Author Contributions
  11. Experimental Details
  12. Major Arrhythmia Classes
  13. Pre-training Implementation
  14. Evaluation on Downstream task
  15. Baseline
  16. ...and 1 more sections

Figures (1)

  • Figure 1: Concept. ECG 1, ECG 2, and ECG 3 on the left are sampled according to the strategies in Section \ref{['sec2.1']} (e.g., given ECG 1 drawn from the male subgroup (strategy 3), we can sample ECG 2 from the same gender group and ECG 3 from distinct gender group). The contrastive learning objective attracts the learned representation from positive samples and repels those from negative samples. We then evaluate the learned representation on the downstream task.