Data distribution impacts the performance and generalisability of contrastive learning-based foundation models of electrocardiograms
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911151516811264 |
|---|---|
| author | Khattak, Gul Rukh Patlatzoglou, Konstantinos Barker, Joseph Pastika, Libor Zeidaabadi, Boroumand El-Medany, Ahmed Aggour, Hesham Liang, Yixiu Ribeiro, Antonio H. Annis, Jeffrey Ribeiro, Antonio Luiz Pinho Ge, Junbo Kramer, Daniel B. Waks, Jonathan W. Brittain, Evan Peters, Nicholas Ng, Fu Siong Sau, Arunashis |
| author_facet | Khattak, Gul Rukh Patlatzoglou, Konstantinos Barker, Joseph Pastika, Libor Zeidaabadi, Boroumand El-Medany, Ahmed Aggour, Hesham Liang, Yixiu Ribeiro, Antonio H. Annis, Jeffrey Ribeiro, Antonio Luiz Pinho Ge, Junbo Kramer, Daniel B. Waks, Jonathan W. Brittain, Evan Peters, Nicholas Ng, Fu Siong Sau, Arunashis |
| contents | Contrastive learning is a widely adopted self-supervised pretraining strategy, yet its dependence on cohort composition remains underexplored. We present Contrasting by Patient Augmented Electrocardiograms (CAPE) foundation model and pretrain on four cohorts (n = 5,203,352), from diverse populations across three continents (North America, South America, Asia). We systematically assess how cohort demographics, health status, and population diversity influence the downstream performance for prediction tasks also including two additional cohorts from another continent (Europe). We find that downstream performance depends on the distributional properties of the pretraining cohort, including demographics and health status. Moreover, while pretraining with a multi-centre, demographically diverse cohort improves in-distribution accuracy, it reduces out-of-distribution (OOD) generalisation of our contrastive approach by encoding cohort-specific artifacts. To address this, we propose the In-Distribution Batch (IDB) strategy, which preserves intra-cohort consistency during pretraining and enhances OOD robustness. This work provides important insights for developing clinically fair and generalisable foundation models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_10369 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Data distribution impacts the performance and generalisability of contrastive learning-based foundation models of electrocardiograms Khattak, Gul Rukh Patlatzoglou, Konstantinos Barker, Joseph Pastika, Libor Zeidaabadi, Boroumand El-Medany, Ahmed Aggour, Hesham Liang, Yixiu Ribeiro, Antonio H. Annis, Jeffrey Ribeiro, Antonio Luiz Pinho Ge, Junbo Kramer, Daniel B. Waks, Jonathan W. Brittain, Evan Peters, Nicholas Ng, Fu Siong Sau, Arunashis Machine Learning Artificial Intelligence Signal Processing Tissues and Organs Contrastive learning is a widely adopted self-supervised pretraining strategy, yet its dependence on cohort composition remains underexplored. We present Contrasting by Patient Augmented Electrocardiograms (CAPE) foundation model and pretrain on four cohorts (n = 5,203,352), from diverse populations across three continents (North America, South America, Asia). We systematically assess how cohort demographics, health status, and population diversity influence the downstream performance for prediction tasks also including two additional cohorts from another continent (Europe). We find that downstream performance depends on the distributional properties of the pretraining cohort, including demographics and health status. Moreover, while pretraining with a multi-centre, demographically diverse cohort improves in-distribution accuracy, it reduces out-of-distribution (OOD) generalisation of our contrastive approach by encoding cohort-specific artifacts. To address this, we propose the In-Distribution Batch (IDB) strategy, which preserves intra-cohort consistency during pretraining and enhances OOD robustness. This work provides important insights for developing clinically fair and generalisable foundation models. |
| title | Data distribution impacts the performance and generalisability of contrastive learning-based foundation models of electrocardiograms |
| topic | Machine Learning Artificial Intelligence Signal Processing Tissues and Organs |
| url | https://arxiv.org/abs/2509.10369 |