Robust PPG-Exclusive Heart Rate Estimation During Maximal Exercise Using Population-Based Priors

Q: What is the clinical evidence from this study?

Study design: Observational. Population: Healthy subjects performing maximal cycloergometer exercise (n=21). Intervention: PPG-Prior (PPG-exclusive heart rate estimation using population-based priors) vs. BeliefPPG (PPG + accelerometer benchmark). Primary outcome: Median Mean Absolute Error (MAE) in BPM relative to ground-truth ECG during maximal exercise (Median MAE reduction of 63.3% (IQR: 47.3%-74.6%)).

Q: What does this research mean for the field?

Incorporating population-based physiological priors and dynamic time warping alignment enables accurate heart rate estimation from motion-corrupted photoplethysmography alone during maximal exercise, reducing mean absolute error by a median of 63.3% relative to a benchmark combining photoplethysmography and accelerometer data. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

Elisabetta De Giovanni; Iñigo Urteaga; Matteo Croci; Tomás Teijeiro

doi:10.48550/arxiv.2610.37132

October 27, 2026Open Access

Robust PPG-Exclusive Heart Rate Estimation During Maximal Exercise Using Population-Based Priors

View Full Paper

Ask AI

Bookmark

Share

Key result

PPG-Prior reduces heart rate error ~63% vs. BeliefPPG during maximal exercise.

Median MAE reduction of 63.3% (IQR: 47.3%-74.6%)
n=21

Why the study?

Motion artifacts during intense physical activity reduce the reliability of photoplethysmography (PPG) heart rate monitoring due to spectral overlap with the cardiac signal, and it was unclear whether severely contaminated PPG signals could yield reliable heart rate estimates without auxiliary sensors.

Does a PPG-exclusive heart rate estimation algorithm leveraging population-based physiological priors reduce heart rate estimation error compared to a PPG plus accelerometer benchmark in healthy adults undergoing maximal and submaximal exercise testing?

Key Points

To determine whether photoplethysmography (PPG) signals heavily contaminated by motion artifacts during intense exercise can yield accurate heart rate estimates without auxiliary sensors by leveraging population-based physiological priors.
Simultaneous PPG, accelerometer, and ground-truth ECG data were collected from 21 healthy subjects during maximal cycloergometer tests with 3-minute, 30-watt increments.
Population-based heart rate priors were constructed from normalized ECG data using Leave-One-Subject-Out cross-validation and aligned via dynamic time warping (DTW) to compute a barycenter average.
Photoplethysmography power spectral density was combined with the prior to generate a posterior distribution, from which candidate trajectories were ranked and selected.
The proposed PPG-Prior approach achieved a median mean absolute error (MAE) of 8.88 BPM (IQR: 2.92 to 11.94 BPM), compared to 21.35 BPM for the PPG+accelerometer BeliefPPG benchmark.
The method reduced MAE relative to ground-truth ECG by a median of 63.3% (IQR: 47.3% to 74.6%) compared to the PPG+accelerometer benchmark.
The Top-Traj scoring framework achieved an MAE below 3 BPM in 28.6% of cases and below 5 BPM in 38.1% of cases, whereas the benchmark never dropped below 3 BPM and was under 5 BPM in 14.3% of cases.

Study Design

Type

Observational (n=21)

Multicenter

Yes

Structured PICO

Does a PPG-exclusive heart rate estimation algorithm leveraging population-based physiological priors reduce heart rate estimation error compared to a PPG plus accelerometer benchmark in healthy adults undergoing maximal and submaximal exercise testing?

P

Population

21 healthy adults aged 20-37 (33.3% female) engaged in regular physical activity undergoing maximal cycloergometer exercise testing, plus 5 volunteers aged 26-35 in a submaximal validation cohort.

I

Intervention

PPG-Prior algorithm: a PPG-exclusive heart rate (HR) estimation method combining PPG Power Spectral Density (PSD) with population-based HR evolution priors (derived from ground-truth ECG using Leave-One-Subject-Out cross-validation and soft Dynamic Time Warping alignment) and iterative Viterbi search (Top-Traj or Min-DTW trajectory selection) without auxiliary motion sensors.

C

Comparator

BeliefPPG deep learning algorithm using PPG and 3-axis accelerometer (PPG+ACC) data (and PPG-only configuration), evaluated against single-lead chest ECG (leads I-II), Hexoskin ProShirt, or Shimmer ECG as ground truth.

O

Outcome

Mean absolute error (MAE) in beats per minute (BPM) calculated second-by-second relative to ground-truth ECG, and percentage MAE reduction compared to the BeliefPPG (PPG+ACC) benchmark.surrogate

Main Result

Effect estimate: Median MAE reduction of 63.3% (IQR: 47.3%-74.6%)

Absolute Event Rate: 8.88% vs 21.35%

Limitations

Trajectory selection lacks a segment-level adaptive strategy to dynamically evaluate posterior reliability.
Validation is restricted to incremental maximal and submaximal exercise protocols rather than non-incremental real-world sports training.
Prior empirical learning is restricted to specific available datasets, which may miss relevant data gaps in the estimation.
Lacks personalized priors for tailored continuous monitoring.
Current trajectory selection lacks a segment-level adaptive strategy to dynamically evaluate posterior reliability in noisy segments.
Validation is restricted to incremental maximal and submaximal cycloergometer exercise protocols rather than non-incremental real-world sports training sessions.
Prior empirical learning is restricted to specific available datasets, which may miss relevant data gaps in the estimation that require physiological trend modeling and forecasting.

Implication

Leveraging population-based physiological priors and Dynamic Time Warping alignment enables accurate, PPG-exclusive wearable heart rate monitoring during high-intensity exercise, outperforming deep learning benchmarks that require auxiliary accelerometer sensors.

Abstract

PPG is the main technology for wearable heart rate (HR) monitoring. However, motion artifacts during intense physical activity reduce its reliability due to spectral overlap with the cardiac signal. This study investigates whether severely contaminated PPG signals can yield reliable HR estimates without auxiliary sensors by leveraging prior knowledge of HR dynamics. We acquired simultaneous PPG and ECG signals from 21 subjects during a maximal cycloergometer test. Using Leave-One-Subject-Out cross-validation, we built HR evolution priors from the ground-truth ECG and applied Dynamic Time Warping (DTW) to align normalized HR curves and estimate a barycenter average. DTW aligns normalized HR curves with minimal error across all subjects, validating robust population-based priors. For each test subject, we combined the PPG Power Spectral Density (PSD) with the prior distribution to generate a "posterior PSD". We identified and ranked candidate HR trajectories using heuristic search within the "posterior PSD" and selected the most likely path. Our approach reduced mean absolute error (MAE) in BPM relative to the ground-truth ECG by a median of 63.3% (IQR: 47.3%-74.6%) compared to a PPG+accelerometer benchmark. These results show that physiological priors and DTW-based alignment enable robust, PPG-exclusive HR monitoring during high-intensity exercise.

Discussion

No takes yet. Share an insight, caveat, or question.

Cite This Study

Giovanni et al. (2026) conducted an observational in Healthy subjects performing maximal cycloergometer exercise (n=21). PPG-Prior (PPG-exclusive heart rate estimation using population-based priors) vs. BeliefPPG (PPG + accelerometer benchmark) was evaluated on Median Mean Absolute Error (MAE) in BPM relative to ground-truth ECG during maximal exercise (Median MAE reduction of 63.3% (IQR: 47.3%-74.6%)). PPG-Prior reduced heart rate mean absolute error by a median of 63.3% (8.88 vs 21.35 BPM) compared to the BeliefPPG (PPG+accelerometer) benchmark during maximal exercise.

synapsesocial.com/papers/6ae4339ed9418cd7c8f46851 https://doi.org/10.48550/arxiv.2610.37132

View Full Paper

Ask AI

Bookmark

Share

Robust PPG-Exclusive Heart Rate Estimation During Maximal Exercise Using Population-Based Priors

Key Points

Study Design

Structured PICO

Main Result

Limitations

Implication

Abstract

Discussion

Cite This Study

Robust PPG-Exclusive Heart Rate Estimation During Maximal Exercise Using Population-Based Priors

Key Points

Study Design

Structured PICO

Main Result

Limitations

Implication

Abstract

Discussion

Cite This Study

Keep reading

Keep reading