P2Es: Physiology-Conditioned Diffusion for 12-Lead ECG Generation from PPG

Area of Science:

• Biomedical Engineering
• Cardiovascular Monitoring
• Artificial Intelligence in Healthcare

Background:

• The 12-lead electrocardiogram (ECG) is the gold standard for cardiovascular assessment but requires cumbersome setups.
• Photoplethysmography (PPG) offers a convenient alternative but cannot reconstruct multi-lead ECGs due to limitations in capturing spatial-temporal data.
• Existing PPG methods lack the necessary constraints to accurately model inter-lead dependencies for ECG reconstruction.

Purpose of the Study:

• To develop a novel framework, P2Es, for reconstructing high-fidelity 12-lead ECGs from PPG signals.
• To address the limitations of current PPG-based methods in capturing the spatial-temporal dynamics required for ECG reconstruction.
• To enable scalable, longitudinal cardiac screening and pre-clinical risk stratification using ambulatory monitoring.

Main Methods:

• Introduced P2Es, a physiology-conditioned diffusion framework integrating GroupFinder for PPG-ECG latent space alignment.
• Implemented a forward process with frequency blurring and temporal noise to simulate real-world distortions.
• Utilized a hierarchical multi-scale generation module with frequency deblurring in the reverse process for detailed signal recovery.
• Developed a resource-efficient edge architecture for on-device inference.

Main Results:

• P2Es significantly outperforms state-of-the-art baselines in reconstructing 12-lead ECGs from PPG.
• The GroupFinder module effectively aligns PPG latent spaces with ECG-derived clusters, resolving physiological information gaps.
• The hierarchical generation module successfully recovers fine-grained diagnostic details, overcoming frequency blurring and noise.
• The edge architecture facilitates rapid, on-device inference, enabling practical ambulatory monitoring.

Conclusions:

• P2Es presents a scalable and effective solution for reconstructing 12-lead ECGs from PPG signals.
• This framework overcomes previous limitations, enabling detailed cardiac monitoring in ambulatory settings.
• P2Es holds significant potential for longitudinal cardiac screening and pre-clinical risk stratification.