CoFE: A Framework Generating Counterfactual ECG for Explainable Cardiac AI-Diagnostics

2025-08-21

AI Summary

To address the limited clinical interpretability of AI-based electrocardiogram (ECG) models, this paper proposes CoFE—the first counterfactual explanation framework specifically designed for ECG signals. CoFE integrates temporal modeling with causal counterfactual generation to precisely localize perturbations in clinically meaningful features—such as waveform amplitudes and interval durations—and quantify their causal effects on model predictions (e.g., atrial fibrillation classification or serum potassium level regression). The generated counterfactual ECG signals adhere to domain-specific physiological priors, enabling clinicians to scrutinize and validate model reasoning. Experimental results demonstrate that CoFE substantially enhances model transparency and clinical trustworthiness. It provides a novel, interpretable paradigm for deploying AI-ECG systems in real-world clinical settings.

Technology Category

• Humans and AI: Explainable AI (XAI) for Human Understanding
• Machine Learning: Causal Learning
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Recognizing the need for explainable AI (XAI) approaches to enable the successful integration of AI-based ECG prediction models (AI-ECG) into clinical practice, we introduce a framework generating extbf{Co}unter extbf{F}actual extbf{E}CGs (i,e., named CoFE) to illustrate how specific features, such as amplitudes and intervals, influence the model's predictive decisions. To demonstrate the applicability of the CoFE, we present two case studies: atrial fibrillation classification and potassium level regression models. The CoFE reveals feature changes in ECG signals that align with the established clinical knowledge. By clarifying both extbf{where valid features appear} in the ECG and extbf{how they influence the model's predictions}, we anticipate that our framework will enhance the interpretability of AI-ECG models and support more effective clinical decision-making. Our demonstration video is available at: https://www.youtube.com/watch?v=YoW0bNBPglQ.

Problem

• Generating counterfactual ECG signals for explainable AI diagnostics
• Illustrating how specific ECG features influence model predictions
• Enhancing interpretability of AI-based ECG prediction models

Innovation

• Generates counterfactual ECG signals for explainability
• Reveals feature changes aligning with clinical knowledge
• Enhances interpretability of AI-ECG diagnostic models