Manifold Learning for Personalized and Label-Free Detection of Cardiac Arrhythmias

2025-06-19

AI Summary

Addressing key bottlenecks in arrhythmia detection—including scarce labeled data, high inter-subject variability, and overreliance on population-level homogeneity—this paper proposes an unsupervised, patient-specific detection framework based on nonlinear manifold learning. For the first time, t-SNE and UMAP are systematically applied to label-free feature learning from single- or multi-lead ECG signals (MLII/V1), enabling direct extraction of fine-grained, clinically interpretable, and patient-discriminative features without prior knowledge or class labels. The method circumvents the dependency of conventional supervised models on annotated datasets and cross-subject generalization, achieving end-to-end patient-level anomaly identification. Evaluated on the MIT-BIH Arrhythmia Database, it achieves ≥90% subject identification accuracy, a median patient-level arrhythmia classification accuracy of 98.96%, and a median F1-score of 91.02%.

Technology Category

• Machine Learning: Learning with Manifolds
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Computer Vision: Representation Learning for Vision

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Electrocardiograms (ECGs) provide direct, non-invasive measurements of heart activity and are well-established tools for detecting and monitoring cardiovascular disease. However, manual ECG analysis can be time-consuming and prone to errors. Machine learning has emerged as a promising approach for automated heartbeat recognition and classification, but substantial variations in ECG signals make it challenging to develop generalizable models. ECG signals can vary widely across individuals and leads, while datasets often follow different labeling standards and may be biased, all of which greatly hinder supervised methods. Conventional unsupervised methods, e.g. principal component analysis, prioritize large (and often obvious) variances in the data and typically overlook subtle yet clinically relevant patterns. If labels are missing and/or variations are significant but small, both approaches fail. Here, we show that nonlinear dimensionality reduction (NLDR) can accommodate these issues and identify medically relevant features in ECG signals, with no need for training or prior information. Using the MLII and V1 leads of the MIT-BIH dataset, we demonstrate that t-distributed stochastic neighbor embedding and uniform manifold approximation and projection can discriminate individual recordings in mixed populations with>= 90% accuracy and distinguish different arrhythmias in individual patients with a median accuracy of 98.96% and a median F1-score of 91.02%. The results show that NLDR holds much promise for cardiac monitoring, including the limiting cases of single-lead ECG and the current 12-lead standard of care, and for personalized health care beyond cardiology.

Problem

• Automated detection of cardiac arrhythmias without labeled data
• Handling ECG signal variations across individuals and leads
• Identifying subtle clinical patterns using nonlinear dimensionality reduction

Innovation

• Nonlinear dimensionality reduction for ECG analysis
• Label-free detection using manifold learning techniques
• High accuracy in arrhythmia discrimination without training