SIGMA-PPG: Statistical-prior Informed Generative Masking Architecture for PPG Foundation Model

TL;DR

SIGMA-PPG introduces a novel generative model for PPG signals that leverages statistical priors and semantic consistency to improve robustness and performance across multiple tasks.

Contribution

The paper presents a new generative foundation model with a prior-guided adversarial masking mechanism and semantic constraints, addressing noise and redundancy issues in PPG modeling.

Findings

Outperforms five state-of-the-art baselines on 12 tasks

Pre-trained on over 120,000 hours of data

Enhances codebook semantic density and robustness

Abstract

Current foundation model for photoplethysmography (PPG) signals is challenged by the intrinsic redundancy and noise of the signal. Standard masked modeling often yields trivial solutions while contrastive methods lack morphological precision. To address these limitations, we propose a Statistical-prior Informed Generative Masking Architecture (SIGMA-PPG), a generative foundation model featuring a Prior-Guided Adversarial Masking mechanism, where a reinforcement learning-driven teacher leverages statistical priors to create challenging learning paths that prevent overfitting to noise. We also incorporate a semantic consistency constraint via vector quantization to ensure that physiologically identical waveforms (even those altered by recording artifacts or minor perturbations) map to shared indices. This enhances codebook semantic density and eliminates redundant feature structures. Pre-trained on over 120,000 hours of data, SIGMA-PPG achieves superior average performance compared to five state-of-the-art baselines across 12 diverse downstream tasks. The code is available at https://github.com/ZonghengGuo/SigmaPPG.