Key result
Adding synthetic signals to real-world data boosts ECG classification accuracy by a modest ~0.7%.
Why the study?
Obtaining well-balanced datasets is challenging, and whether synthetic data is accurate enough to improve deep learning model training for electrocardiography classification remains unclear.
Does supplementing real-world electrocardiography data with synthetic signals improve the classification performance of deep neural networks?
Does supplementing real-world electrocardiography data with synthetic signals improve the classification performance of deep neural networks?
Effect estimate: Accuracy boost up to 0.7%
Supplementing real-world ECG datasets with synthetic data can provide a modest improvement in deep learning model accuracy, particularly for balancing class distributions in rare diseases.
Obtaining well-balanced datasets with quality-assured and clinically-validated labels remains a challenge in healthcare. Many datasets suffer from a class imbalance with a majority of healthy subjects. Synthetic data, which is artificially generated, might be a suitable solution to counter class imbalance; however, the degree to which synthetic data is "accurate enough" to actually improve model training is unclear in many areas. In this work, we evaluate in how far synthetic electrocardiography signals as a supplement to real-world data can improve performance of a deep neural network for electrocardiography classification. Real data stems from the public PTB-XL dataset and synthetic data is based on electrophysiological simulations from the recently-published MedalCare-XL dataset. We measure the classification performance of four different classes including three cardiac conduction abnormalities and evaluate different oversampling and training data shuffling strategies, borrowing from the concept of Curriculum Learning. Results indicate that adding synthetic data yields a modest yet consistent accuracy boost up to 0.7%. Even though synthetic signals are not as complex as real-world measurements, they might be promising to balance class distributions in the context of rare disease.
No takes yet. Share an insight, caveat, or question.
Schmale et al. (2025) studied Cardiac conduction abnormalities. Synthetic electrocardiography signals combined with real-world data vs. Real-world data alone was evaluated on Classification performance (accuracy) (Accuracy boost up to 0.7%). Adding synthetic electrocardiography signals to real-world data using curriculum learning strategies yielded a modest yet consistent accuracy boost of up to 0.7% for ECG classification.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: