Learning Cardiac Latent Representations in Vectorcardiogram Space

2026-05-29

AI Summary

This work addresses the limitations of conventional electrocardiogram (ECG) representation learning, which is prone to overfitting and poor generalization due to redundancy across leads and spurious correlations. To overcome these issues, the authors propose LVCG, the first universal self-supervised representation learning framework tailored to the Frank vectorcardiogram (VCG) space. By shifting the learning objective from raw observed signals to a physiologically meaningful latent space, LVCG captures the intrinsic, viewpoint-invariant characteristics of cardiac electrical activity, thereby circumventing lead-specific artifacts. Experimental results demonstrate that LVCG consistently outperforms ECG-space baseline methods across multiple downstream tasks and exhibits markedly enhanced robustness and generalization under domain shift scenarios.

Abstract

Electrocardiography (ECG) is a cornerstone of cardiac assessment, making the learning of informative ECG representations fundamental to tasks ranging from disease diagnosis to clinical report generation. However, existing methods operate almost exclusively in the observable ECG signal space. In practice, the standard twelve-lead ECG represents multiple projections of the same underlying cardiac electrical activity from different spatial orientations. Therefore, representation learning in the ECG space inevitably introduces substantial redundancy, which may lead to spurious correlations and increased risk of overfitting. To address this and motivated by the Frank vectorcardiogram (VCG) model, we propose learning a unified latent representation of cardiac electrical activity directly in the VCG space. We introduce LVCG, the first general self-supervised representation learning framework designed to operate in this physically grounded latent space. By learning view-invariant latent VCG representations rather than lead-specific artifacts, VCG minimizes redundancy and improves generalization. LVCG generally outperforms ECG-space baselines across tasks, demonstrating enhanced robustness and generalization, especially in domain shift settings.

Problem

• ECG representation learning
• redundancy
• vectorcardiogram
• overfitting
• spurious correlations

Innovation

• vectorcardiogram
• representation learning
• self-supervised learning
• domain generalization
• ECG