TolerantECG: A Foundation Model for Imperfect Electrocardiogram

2025-07-13

AI Summary

ECG diagnosis is highly susceptible to lead dropout and noise corruption, severely compromising reliability. To address this, we propose the first robust foundation model for incomplete and noisy ECGs, integrating contrastive learning, self-supervised pretraining, and knowledge-enhanced text-signal alignment—enabling flexible input of arbitrary lead subsets while jointly modeling multi-lead waveforms and clinical report semantics. Evaluated on PTB-XL, our model achieves top-1 or top-2 performance across most settings; on MIT-BIH Arrhythmia Database, it attains state-of-the-art accuracy. This work is the first to systematically tackle the joint modeling of lead incompleteness and noise robustness, establishing a general-purpose framework for reliable diagnosis of low-quality ECGs.

Technology Category

• Natural Language Processing: Safety and Robustness
• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

The electrocardiogram (ECG) is an essential and effective tool for diagnosing heart diseases. However, its effectiveness can be compromised by noise or unavailability of one or more leads of the standard 12-lead recordings, resulting in diagnostic errors or uncertainty. To address these challenges, we propose TolerantECG, a foundation model for ECG signals that is robust to noise and capable of functioning with arbitrary subsets of the standard 12-lead ECG. TolerantECG training combines contrastive and self-supervised learning frameworks to jointly learn ECG signal representations alongside their corresponding knowledge-retrieval-based text report descriptions and corrupted or lead-missing signals. Comprehensive benchmarking results demonstrate that TolerantECG consistently ranks as the best or second-best performer across various ECG signal conditions and class levels in the PTB-XL dataset, and achieves the highest performance on the MIT-BIH Arrhythmia Database.

Problem

• Addresses ECG noise and missing leads issues
• Develops robust model for imperfect ECG signals
• Improves diagnostic accuracy across varied ECG conditions

Innovation

• Robust foundation model for noisy ECG signals
• Functions with arbitrary subsets of 12-lead ECG
• Combines contrastive and self-supervised learning