PulseBound: Future-Beat State Forecasting Under an Explicit Information Boundary

2026-10-08

AI Summary

This study addresses the issue of future information leakage caused by normalization or transformation in PPG representation learning by proposing a causality-safe predictive learning framework. Methodologically, it integrates physiologically structured future beat prediction with explicit window boundary memorization, and designs content-agnostic truncation alongside prefix-specific normalization mechanisms to enforce strict suffix invariance for isolating causal violations. Furthermore, the model incorporates a shared horizon-conditioned head, element-wise masking, and optional ECG supervision. Experimental results demonstrate that the proposed method significantly reduces prediction errors on the MIMIC and VitalDB datasets, achieves state-of-the-art performance across multiple downstream tasks, and empirically verifies zero suffix dependency at the gradient level.

Abstract

Predictive representation learning from photoplethysmography (PPG) can violate causal information access even with causal attention, as normalization, nonlocal transforms, or companion views may depend on withheld samples. We introduce PulseBound, a PPG representation learner combining physiologically structured future-beat prediction with an explicit stored-window information boundary. A content-independent cutoff separates the visible prefix from the prediction target. Prefix-only normalization, suffix replacement before derived-view construction, and aligned masking ensure that encoder inputs depend only on the visible prefix and cutoff. This yields stored-suffix invariance: with fixed model state, randomness, prefix, and cutoff, changing the stored suffix cannot change the forecast context. A shared horizon-conditioned head predicts nine rhythm and morphology descriptors for up to four extractor-valid future beats, using elementwise validity masks; optional ECG-derived pulse-arrival-time supervision is restricted to training. On MIMIC and VitalDB groups held out from PulseBound backbone pretraining, PulseBound reduces nine-state transformed-space MAE relative to last-visible-beat persistence by 28.06% and 22.22%, respectively, with gains in MAE, MAE-Skill, and Spearman correlation across all 40 source-cutoff-horizon cells. In a separate comparison of seven models on 13 downstream tasks, PulseBound achieves the best mean on nine frozen linear-probe and seven full-fine-tuning tasks. Stored-suffix interventions cause zero recorded changes in forecast contexts or predictions, with zero suffix-input gradients at audited precision under the stored-window interface. These findings separate three testable aspects of predictive physiological representation learning: information access, supervised future structure, and transfer.

Problem

• Photoplethysmography
• Causal information leakage
• Predictive representation learning
• Future-beat forecasting
• Information boundary

Innovation

• Predictive Representation Learning
• Information Boundary
• Stored-Suffix Invariance
• Photoplethysmography
• Future-Beat Forecasting