{"paper":{"title":"Extending Pretrained 10-Second ECG Foundation Models to Longer Horizons","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A lightweight plug-in module guided by a frozen 10-second ECG model can process longer and variable-length recordings without retraining the backbone.","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anshul Thakur, David A. Clifton, Fredrik K. Gustafsson, Jean-Michel Morel, Jinpei Han, Kangning Cui, Lei Clifton, Mattia Carletti, Patitapaban Palo, Raymond H. Chan, Shreyank N Gowda, Wei Tang, Xiao Gu","submitted_at":"2026-05-16T12:52:23Z","abstract_excerpt":"Electrocardiogram (ECG) foundation models pretrained on typical diagnostic 10-second ECG segments, have demonstrated strong transferability across a range of clinical applications. However, many real-world applications produce recordings that are typically longer, and are varied in duration during inference time. These 10-second models have no built-in way to combine information across time. Extending them to longer horizons introduces two challenges: structural incompatibilities arising from input-length disparities, and semantic challenges that limit meaningful temporal aggregation. We propo"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on multiple long-horizon ECG tasks, datasets, and foundation model backbones demonstrate that our method enables robust long-horizon extension from pretrained snapshot models, consistently outperforming sliding-window and pooling-based baselines with strong parameter efficiency.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That a lightweight plug-in module guided by a frozen 10-second pretrained model can achieve semantically informed temporal aggregation for variable-length ECGs without significant loss of clinically relevant information from the original pretraining.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A parameter-efficient plug-in framework adds structurally compatible long-sequence processing and semantically informed temporal modeling to extend pretrained 10-second ECG foundation models to longer variable-length inputs.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A lightweight plug-in module guided by a frozen 10-second ECG model can process longer and variable-length recordings without retraining the backbone.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2a9152b6272a910a28466b028a9228867defefca626c919945cea5e98eb41d7c"},"source":{"id":"2605.16975","kind":"arxiv","version":1},"verdict":{"id":"75601b65-d297-4dc5-b20e-3dc7ba2d091f","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T20:59:04.114150Z","strongest_claim":"Experiments on multiple long-horizon ECG tasks, datasets, and foundation model backbones demonstrate that our method enables robust long-horizon extension from pretrained snapshot models, consistently outperforming sliding-window and pooling-based baselines with strong parameter efficiency.","one_line_summary":"A parameter-efficient plug-in framework adds structurally compatible long-sequence processing and semantically informed temporal modeling to extend pretrained 10-second ECG foundation models to longer variable-length inputs.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That a lightweight plug-in module guided by a frozen 10-second pretrained model can achieve semantically informed temporal aggregation for variable-length ECGs without significant loss of clinically relevant information from the original pretraining.","pith_extraction_headline":"A lightweight plug-in module guided by a frozen 10-second ECG model can process longer and variable-length recordings without retraining the backbone."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16975/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T21:31:19.114325Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T21:10:54.372049Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T19:51:56.700714Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T19:50:07.200163Z","status":"skipped","version":"0.1.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.219543Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.306544Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"78722c12007c6089bab9b622bcfa42e5045ea72b26844497071e6bf43cc71c61"},"references":{"count":38,"sample":[{"doi":"","year":2021,"title":"Artificial intelligence-enhanced electrocardiography in cardiovascular disease management.Nature Reviews Cardiology, 18(7):465–478, 2021","work_id":"fa6a2cb1-f57f-4b81-9a08-cf4c829c06f0","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"Deep learning for ecg analysis: Benchmarks and insights from ptb-xl.IEEE journal of biomedical and health informatics, 25(5):1519–1528, 2020","work_id":"b5977990-42c2-4872-83ba-672510c0787b","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"An electrocardiogram foundation model built on over 10 million recordings with external evaluation across multiple domains","work_id":"904af5fc-5ba9-47e9-801e-4b86d616bfd7","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Aldo Faisal, and David A","work_id":"b4c94084-eb62-4de6-9f94-bcd585f75ec8","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Ecg-fm: An open electrocardiogram foundation model.JAMIA open, 8(5):ooaf122, 2025","work_id":"8c5680c2-6757-4509-aa0a-0ca9b1e91401","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":38,"snapshot_sha256":"36dd25df82d76ecfa334922a346d831bbaaee90f083fa4c1277943b32500a253","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"2b4ba3dc3cf4f425c7df9fcdd75d6562b3e87ee64f31feeae50464cb4d904816"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}