cs.LGApr 9, 2026

PRISM-CTG: A Foundation Model for Cardiotocography Analysis with Multi-View SSL

Authors: Sheng Wong, Ravi Shankar, Beth Albert, Hao Fei, Lin Li, Imane Ben M'Barek, Manu Vatish, Gabriel Davis Jones

Organizations: Nuffield Department of Women’s & Reproductive Health, University of Oxford, United Kingdom · OXDHL, University of Oxford, Oxford, United Kingdom · OATML, Department of Computer Science, University of Oxford, United Kingdom · Université Paris Cité, Paris, France

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Supervised deep learning models for automated CTG analysis are typically constrained by narrowly curated labelled datasets and limited patient cohorts, leaving substantial volumes of physiologically informative clinical recordings untapped. To address this limitation, we propose Physiology-aware Representation Learning via Integrated Self-supervision and Metadata for CTG (PRISM-CTG), a clinically grounded self-supervised foundation model (FM) for CTG that leverages large-scale unlabelled recordings to learn transferable domain-level representations. PRISM-CTG is pretrained using a multi-view self-supervised framework that jointly optimises 3 complementary pretext objectives: random-projected guided masked signal reconstruction, clinical variable prediction, and feature classification. Each objective is associated with a dedicated task-specific token, enabling specialised representation learning, while controlled cross-attention facilitates information exchange across clinical context. By reframing patient metadata and domain knowledge, which are often underutilised in conventional training as prediction targets, Prism-CTG transforms readily available clinical information into additional supervisory targets that guide clinically meaningful representation learning. Extensive experiments across 7 downstream CTG tasks in both antepartum and intrapartum domains demonstrated that PRISM-CTG consistently outperforms in-domain and SSL baselines. Notably, PRISM-CTG demonstrated strong generalisation under external validation on 2 datasets, while achieving comparable performance to studies trained on substantially larger, privately labelled datasets. To our knowledge, this is the first study to introduce large-scale FM for CTG that learns domain-level representations.

Figures & tables

Figure 1: Overview of a major limitation in conventional supervised CTG training. Supervised models typically select temporally aligned recordings for a single predefined task, discarding CTGs acquired earlier in the gestational period (top). In contrast, PRISM-CTG leverages CTGs from the full gestational period during pretraining to learn general representations that are transferable across multiple downstream tasks.
Figure 2: PRISM-CTG architecture. The model is pretrained with 3 pretext objectives and evaluated via linear probing by freezing the encoder and training a task-specific linear head on 6 downstream tasks (four antepartum, two intrapartum*).
  AntepartumIntrapartum 
ModelPre- trainingT1T2T3T4T5T6T7Average
∙\bullet Supervised Learning
ResNet✗0.6640.6840.6700.7720.8190.8020.8080.746
TimesNet✗0.5360.6070.6220.7080.7030.7610.8070.678
TimeMixer✗0.6400.6680.6630.7680.8290.8100.8650.749
CTGNet✗0.6290.6310.6280.7110.7810.7680.8490.714
Table 1: Comparison of AUC performance across antepartum and intrapartum CTG tasks. The best AUC for each task is shown in bold , and the second-highest is underlined .
Figure 3: (Left) Average performance across all tasks under different data regimes with the top performing SSL and supervised models. (Right) Ablation study of PRISM-CTG showing average AUC for the full model and its variants.
Figure 4: Average performance grouped by proportion of signal dropout
Figure 5: Average performance on external validation on APHP and CTU-UHB.
More in this topic

Self-Supervised Time Series Representation Learning

Self-supervised time series representation learning uses unlabeled sequences to learn features that transfer to forecasting, classification, and anomaly detection.

See all 142 papers

Explore similar work

CardsList
Apr 2, 2026

Learning ECG Image Representations via Dual Physiological-Aware Alignments

Electrocardiograms (ECGs) are among the most widely used diagnostic tools for cardiovascular diseases, and a large amount of ECG data worldwide appears only in image form. However, most existing automated ECG analysis methods rely on access to raw signal recordings, limiting their applicability in real-world and resource-constrained settings. In this paper, we present ECG-Scan, a self-supervised framework for learning clinically generalized representations from ECG images through dual physiological-aware alignments: 1) Our approach optimizes image representation learning using multimodal contrastive alignment between image and gold-standard signal-text modalities. 2) We further integrate domain knowledge via soft-lead constraints, regularizing the reconstruction process and improving signal lead inter-consistency. Extensive benchmarking across multiple datasets and downstream tasks demonstrates that our image-based model achieves superior performance compared to existing image baselines and notably narrows the gap between ECG image and signal analysis. These results highlight the potential of self-supervised image modeling to unlock large-scale legacy ECG data and broaden access to automated cardiovascular diagnostics.

Hung Manh Pham, Jialu Tang, Aaqib Saeed +3

University of Cambridge · Singapore Management University · Eindhoven University of Technology, Netherlands

Electrocardiogram Representation LearningMedical Image Representation LearningElectrocardiographySelf-Supervised Visual Representation LearningMultimodal Contrastive Learning
AskPDF
Jun 27, 2026

Learning from Acquisition: Metadata-driven Multimodal Pre-training for Cardiac MRI

Cardiac magnetic resonance imaging (CMR) routinely records structured acquisition metadata, yet most CMR foundation models rely primarily on image-only pre-training and leave this naturally available source of weak semantic supervision largely underexplored. We propose MetaCLIP-CMR, a metadata-driven framework based on Contrastive Language--Image Pre-training (CLIP), which converts imaging modality, anatomical view, scanner vendor, field strength, and scanner model into textual supervision for CMR representation learning. The pretrained image encoder is evaluated on imaging modality classification, cine view classification, and cardiac segmentation. MetaCLIP-CMR achieves 86.8% modality accuracy and 86.5% cine view accuracy, clearly outperforming ImageNet and masked reconstruction initialisations. For downstream cardiac segmentation, MetaCLIP-CMR consistently obtains the highest Dice score across the evaluated ACDC and M&Ms cine short-axis (SAX) settings under both full-data and 20% fine-tuning regimes. Compared with recent image-focused large-scale CMR pre-training models, MetaCLIP-CMR achieves comparable ACDC segmentation performance, while requiring less than 1% of their pre-training image scale. These results suggest that metadata learning offers a natural and easy-to-use strategy for transforming routinely recorded acquisition information into effective supervision for foundation-level CMR representation learning, highlighting the promise of metadata-driven multimodal pre-training.

Xueyi Fu, Liwei Hu, Zi Wang +1

Department of Surgery & Cancer, Imperial College London, London, UK · Bioengineering Department and Imperial-X, Imperial College London, London, UK · National Heart and Lung Institute, Imperial College London, London, UK +2

Medical Image SegmentationMultimodal Contrastive LearningMedical Image Representation LearningCardiac MRI
AskPDF
May 12, 2026

Pretraining Strategies and Scaling for ECG Foundation Models: A Systematic Study

Specialized foundation models are beginning to emerge in various medical subdomains, but pretraining methodologies and parametric scaling with the size of the pretraining dataset are rarely assessed systematically and in a like-for-like manner. This work focuses on foundation models for electrocardiography (ECG) data, one of the most widely captured physiological time series world-wide. We present a comprehensive assessment of pretraining methodologies, covering five different contrastive and non-contrastive self-supervised learning objectives for ECG foundation models, and investigate their scaling behavior with pretraining dataset sizes up to 11M input samples, exclusively from publicly available sources. Pretraining strategy has a meaningful and consistent impact on downstream performance, with contrastive predictive coding (slightly ahead of JEPA) yielding the most transferable representations across diverse clinical tasks. Scaling pretraining data continues to yield meaningful improvements up to 11M samples for most objectives. We also compare model architectures across all pretraining methodologies and find evidence for a clear superiority of structured state space models compared to transformers and CNN models. We hypothesize that the strong inductive biases of structured state space models, rather than pretraining scale alone, are the primary driver of effective ECG representation learning, with important implications for future foundation model development in this and potentially other physiological signal domains.

M A Al-Masud, Nils Strodthoff

AI4Health Division Carl von Ossietzky Universität Oldenburg Oldenburg, Lower Saxony, Germany

Electrocardiogram Foundation ModelsSelf-Supervised Time Series Representation LearningState Space ModelsElectrocardiographyTime-Series Foundation ModelsAI Scaling Laws+1
AskPDF