BenchECG and xECG: a benchmark and baseline for ECG foundation models
- 1Digital Cardiology Lab, University Clinic of Internal Medicine III, Medical University of Innsbruck, Innsbruck, Austria
- 2University Clinic of Internal Medicine III, Cardiology and Angiology, Medical University of Innsbruck, Innsbruck, Austria
Abstract
Electrocardiograms (ECGs) are inexpensive, widely used, and well-suited to deep learning. Interest has grown in ECG foundation models that generalise across diverse downstream tasks. However, current evaluation rarely reflects the full capabilities expected of ECG foundation models, with prior work often using narrow task selections and inconsistent datasets, hindering fair comparison. We introduce BenchECG, the first standardised, open benchmark for rigorous ECG foundation model evaluation, comprising eight publicly available datasets (421,171 patients; 1,674,704 recordings). BenchECG assesses model performance across diverse signals, patient populations, and ten conceptually distinct tasks, spanning classification, detection, regression, survival analysis, and segmentation, including explicit out-of-distribution testing. It enables systematic comparison by measuring how well model representations transfer across tasks under controlled benchmark conditions. Alongside BenchECG, we propose xECG, an ECG foundation model combining the efficiency of xLSTM recurrent architecture (5x faster training than transformer baselines) with SimDINOv2 self-supervised learning. xECG achieves the highest BenchECG score (0.838) and best average rank (1.5 out of 5 models), outperforming current publicly available state-of-the-art methods and providing a strong baseline for future ECG foundation models.
Related
- Benchmark Readiness of Open-access Musculoskeletal Radiograph Datasets for Foundation Model Segmentation: A Systematic Audit
- Age-Normalized Baseline Renal Function: A Simple Risk Marker in Coronary Artery Bypass Grafting
- Baseline Age Disparities Across GLP‐1 Receptor Agonist–Based Trials
- Factors influencing baseline electrocardiographic parameters in rats: A systematic review of experimental studies published from 1960 to the present
- Temporal and therapeutic confounders in predicting efficacy of immunotherapy rechallenge via baseline intratumoral tertiary lymphoid structures.
BibTeX
@article{lunelli2026benchecg,
title = {{BenchECG and xECG: a benchmark and baseline for ECG foundation models}},
author = {Riccardo Lunelli and Angus Nicolson and Samuel Pröll and Sebastian J. Reinstadler and Axel Bauer and Clemens Dlaska},
journal = {npj Digital Medicine},
year = {2026},
month = sep,
doi = {10.1038/s41746-026-03196-y},
url = {https://doi.org/10.1038/s41746-026-03196-y},
note = {CC BY},
}
RIS
TY - JOUR
TI - BenchECG and xECG: a benchmark and baseline for ECG foundation models
AU - Riccardo Lunelli
AU - Angus Nicolson
AU - Samuel Pröll
AU - Sebastian J. Reinstadler
AU - Axel Bauer
AU - Clemens Dlaska
PY - 2026
DA - 2026/09/14
JO - npj Digital Medicine
DO - 10.1038/s41746-026-03196-y
UR - https://doi.org/10.1038/s41746-026-03196-y
AB - Electrocardiograms (ECGs) are inexpensive, widely used, and well-suited to deep learning. Interest has grown in ECG foundation models that generalise across diverse downstream tasks. However, current evaluation rarely reflects the full capabilities expected of ECG foundation models, with prior work often using narrow task selections and inconsistent datasets, hindering fair comparison. We introduce BenchECG, the first standardised, open benchmark for rigorous ECG foundation model evaluation, comprising eight publicly available datasets (421,171 patients; 1,674,704 recordings). BenchECG assesses model performance across diverse signals, patient populations, and ten conceptually distinct tasks, spanning classification, detection, regression, survival analysis, and segmentation, including explicit out-of-distribution testing. It enables systematic comparison by measuring how well model representations transfer across tasks under controlled benchmark conditions. Alongside BenchECG, we propose xECG, an ECG foundation model combining the efficiency of xLSTM recurrent architecture (5x faster training than transformer baselines) with SimDINOv2 self-supervised learning. xECG achieves the highest BenchECG score (0.838) and best average rank (1.5 out of 5 models), outperforming current publicly available state-of-the-art methods and providing a strong baseline for future ECG foundation models.
SN - 2398-6352
ER -