{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T15:01:24Z","timestamp":1785423684725,"version":"3.56.0"},"posted":{"date-parts":[[2026,7,17]]},"group-title":"Cardiovascular Medicine","reference-count":14,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,7,17]],"date-time":"2026-07-17T00:00:00Z","timestamp":1784246400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"id":[{"id":"https:\/\/ror.org\/01cwqze88","id-type":"ROR","asserted-by":"publisher"}]},{"name":"Wellcome Leap Visible"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,7,17]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>We conducted a scaling evaluation of unlabeled pretraining for electrocardiogram foundation model performance. One-dimensional vision transformer (1D-ViT) masked autoencoders were pretrained across increasing ECG volumes and fine-tuned for rhythm, morphology, diagnostic, and structural heart disease tasks. Models pretrained at \u2264400,000 ECGs failed to consistently exceed controls without self-supervised pre-training, whereas 600,000\u2013800,000 ECGs improved AUROC across tasks, suggesting a minimum threshold for effective ECG representation learning.<\/jats:p>","DOI":"10.64898\/2026.07.15.26358182","type":"posted-content","created":{"date-parts":[[2026,7,17]],"date-time":"2026-07-17T22:10:11Z","timestamp":1784326211000},"source":"Crossref","is-referenced-by-count":0,"title":["Scaling ECG Foundation Models and Identifying a Threshold for Effective Representation Learning"],"prefix":"10.64898","author":[{"given":"Raghav","family":"Sriram","sequence":"first","affiliation":[{"name":"Lampe Joint Department of Biomedical Engineering, North Carolina State University, Raleigh, NC"},{"name":"Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ivan","family":"Nenadic","sequence":"additional","affiliation":[{"name":"Division of Cardiology, Department of Medicine, Northwestern University, Feinberg School of Medicine Chicago, Illinois, United States"},{"name":"Duke Heart Center, Duke University Hospital, Durham, North Carolina, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Elan","family":"Shahrabani","sequence":"additional","affiliation":[{"name":"Department of Internal Medicine, University of California in Los Angeles, Los Angeles, California, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sascha N.","family":"Goonewardena","sequence":"additional","affiliation":[{"name":"Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shanshan","family":"Yao","sequence":"additional","affiliation":[{"name":"Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Brennen","family":"Farrell","sequence":"additional","affiliation":[{"name":"Department of Computer Science, North Carolina State University, Raleigh, North Carolina, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zak","family":"Loring","sequence":"additional","affiliation":[{"name":"Duke Heart Center, Duke University Hospital, Durham, North Carolina, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7901-1321","authenticated-orcid":false,"given":"Venkatesh L","family":"Murthy","sequence":"additional","affiliation":[{"name":"Department of Internal Medicine, Division of Cardiovascular Medicine, University of Michigan, Ann Arbor, United States"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"key":"2026072014400554000_2026.07.15.26358182v1.1","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0268-3"},{"key":"2026072014400554000_2026.07.15.26358182v1.2","doi-asserted-by":"publisher","DOI":"10.1038\/s41591-018-0240-2"},{"key":"2026072014400554000_2026.07.15.26358182v1.3","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-025-09227-0"},{"key":"2026072014400554000_2026.07.15.26358182v1.4","unstructured":"Bommasani R , Hudson DA , Adeli E , Altman R , Arora S , von Arx S , et al. On the opportunities and risks of foundation models. arXiv. 2021;2108.07258."},{"key":"2026072014400554000_2026.07.15.26358182v1.5","unstructured":"Kaplan J , McCandlish S , Henighan T , Brown TB , Chess B , Child R , et al. Scaling laws for neural language models. arXiv. 2020;2001.08361."},{"key":"2026072014400554000_2026.07.15.26358182v1.6","first-page":"30016","article-title":"Training compute-optimal large language models","volume":"35","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"2026072014400554000_2026.07.15.26358182v1.7","first-page":"1","article-title":"Self-supervised representation learning from 12-lead ECG data","volume":"48","year":"2021","journal-title":"Computing in Cardiology"},{"key":"2026072014400554000_2026.07.15.26358182v1.8","doi-asserted-by":"crossref","unstructured":"Sawano S , Kodera S , Katsushika S , et al. Applying masked autoencoder-based self-supervised learning for electrocardiogram analysis. PLOS ONE. 2024;19.","DOI":"10.1371\/journal.pone.0307978"},{"key":"2026072014400554000_2026.07.15.26358182v1.9","unstructured":"Na Y , Park M , Tae Y , Joo S. Guiding masked representation learning to capture spatiotemporal relationship of electrocardiogram. In: International Conference on Learning Representations. 2024."},{"key":"2026072014400554000_2026.07.15.26358182v1.10","doi-asserted-by":"crossref","unstructured":"Moody JB , Poitrasson-Rivi\u00e8re A , Renaud JM , Hagio T , Alahdab F , Al-Mallah MH , Vanderver MD , Goonewardena SN , Ficaro EP , Murthy VL . A foundation transformer model with self-supervised learning for ECG-based assessment of cardiac and coronary function. NEJM AI. 2025;2.","DOI":"10.1056\/AIoa2500164"},{"key":"2026072014400554000_2026.07.15.26358182v1.11","doi-asserted-by":"crossref","unstructured":"McKeen K , Oliva L , Masood S , Toma A , Rubin B , Wang B. ECG-FM: an open electrocardiogram foundation model. npj Digital Medicine. 2025.","DOI":"10.1093\/jamiaopen\/ooaf122"},{"key":"2026072014400554000_2026.07.15.26358182v1.12","doi-asserted-by":"crossref","unstructured":"He K , Chen X , Xie S , Li Y , Doll\u00e1r P , Girshick R. Masked autoencoders are scalable vision learners. In: Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition. 2022:16000\u201316009.","DOI":"10.1109\/CVPR52688.2022.01553"},{"key":"2026072014400554000_2026.07.15.26358182v1.13","unstructured":"Gow B , Pollard T , Nathanson LA , Johnson A , Moody B , Fernandes C , et al. MIMIC-IV-ECG: Diagnostic Electrocardiogram Matched Subset. PhysioNet. 2023."},{"key":"2026072014400554000_2026.07.15.26358182v1.14","doi-asserted-by":"crossref","first-page":"154","DOI":"10.1038\/s41597-020-0495-6","article-title":"PTB-XL, a large publicly available electrocardiography dataset","volume":"7","year":"2020","journal-title":"Scientific Data"}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.07.15.26358182","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,20]],"date-time":"2026-07-20T21:40:28Z","timestamp":1784583628000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.64898\/2026.07.15.26358182"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,17]]},"references-count":14,"URL":"https:\/\/doi.org\/10.64898\/2026.07.15.26358182","relation":{},"subject":[],"published":{"date-parts":[[2026,7,17]]},"subtype":"preprint"}}