Published In
Icces 2026 Conference Proceedings of 2026 the 2nd International Conference Computing and Emerging Sciences
Document Type
Conference Proceeding
Publication Date
4-20-2026
Subjects
Applied computing, Life and medical sciences, Healthcare information systems, Computing methodologies, Machine learning, Machine learning approaches, Neuralnetworks
Abstract
Wearable and bedside sensors continuously generate electrocardiograms (ECG), photoplethysmograms (PPG), and related physiological waveforms that could enable earlier detection of deterioration and more personalized care. However, current deep learning pipelines in biomedical signal processing often remain taskand device-specific, degrade under domain shift (new hospitals, sensors, skin tones, motion), and provide limited uncertainty information for safety-critical decisions. We propose PhysioBridge, a foundation-model approach that learns a shared representation space for ECG and PPG via self-supervised pretraining and explicit physiology constraints, then supports downstream adaptation with distribution-free risk control. PhysioBridge introduces (i) multi-rate patch tokenization that preserves clinically meaningful morphology across heterogeneous sampling rates; (ii) coupled masked modeling and cross-modal contrastive alignment to bind ECG and PPGdynamics while remaining label-efficient; (iii) differentiable physiology regularizers that enforce heart-rate consistency and pulsetransit-time plausibility; and (iv) calibrated conformal prediction wrappers that provide coverage guarantees for classification and regression under minimal assumptions. We specify a reproducible evaluation protocol on public waveform resources and ECG corpora, emphasizing cross-device generalization, out-of-distribution detection, and fairness-aware reporting. PhysioBridge is positioned as a practical step toward trustworthy AI foundations for healthcare signal intelligence—supporting scalable reuse, transparent risk management, and safer deployment pathways.
Rights
Copyright (c) 2026 The Authors Creative Commons License This work is licensed under a Creative Commons Attribution 4.0 International License.
DOI
10.1145/3797491.3797535
Persistent Identifier
https://archives.pdx.edu/ds/psu/44701
Citation Details
Alzubaidi, A., Al-shuwaili, A., & Al-bayaty, A. (2026). PhysioBridge: Physiology-Constrained Self-Supervised Foundation Model for Cross-Device ECG–PPG Learning with Conformal Risk Control. In (Editor), Proceedings of the 2026 2nd International Conference on Computing and Emerging Sciences.