Teaching PPG How not Who: Fixed-Effects Distillation from ECG

2026-10-07

AI Summary

This study addresses the tendency of models in ECG-to-PPG knowledge distillation to over-memorize static individual identity features while neglecting dynamic physiological state variations. To mitigate this, we propose a fixed-effect-based distillation framework that precisely disentangles individual traits by subtracting recording-level means, thereby eliminating identity bias and guiding the PPG model to focus on intra-individual physiological dynamics for multimodal signal alignment and debiasing. This approach more than doubles state consistency and significantly improves intra-individual label prediction performance. Extensive experiments across diverse backbone architectures and datasets validate its generalization capability, establishing a novel paradigm for precise health monitoring via wearable devices.

Abstract

ECG is widely used to teach PPG-only models, yet what it teaches is unexamined. Wearables are valued for tracking how a person's cardiovascular state changes, but ECG-to-PPG distillation mostly learns who the person is. A per-recording mean, the trait, holds 40-59% of a frozen ECG teacher's target, and pooled students memorise it without carrying it to new recordings. The raw alignment cosine misses this, since a constant predictor scores 0.793. Across 34 runs, the more identity a student memorises, the less state it learns. Fixed-effects distillation subtracts each recording's mean from prediction and target, so the trait cancels exactly, while a pooled anchor keeps it. State agreement more than doubles, within-person labels improve while age and sex do not, and the gain holds on two backbones and two further databases. Conditioning on the recording turns distillation toward the within-person changes that wearables monitor.

Problem

• ECG-to-PPG distillation
• identity memorization
• cardiovascular state tracking
• wearables
• fixed-effects

Innovation

• Fixed-effects distillation
• ECG-to-PPG
• Knowledge distillation
• Within-person state tracking
• Trait cancellation