Key result
BeatFlow-ECG cuts PPG-to-ECG reconstruction error ~27% compared with Conditional DDPM-1D.
Why the study?
Continuous cardiac monitoring outside clinical settings requires informative and practical signals, as wearable PPG is easier to acquire continuously than ECG but provides only an indirect measurement that is highly sensitive to motion.
Does BeatFlow-ECG improve single-channel ECG waveform reconstruction and beat-timing accuracy from wearable PPG and IMU signals compared with existing deterministic, adversarial, and diffusion-based models?
Does BeatFlow-ECG improve single-channel ECG waveform reconstruction and beat-timing accuracy from wearable PPG and IMU signals compared with existing deterministic, adversarial, and diffusion-based models?
BeatFlow-ECG demonstrates that conditional rectified-flow modeling incorporating inertial motion cues significantly improves single-channel ECG reconstruction from noisy wearable wrist PPG signals during daily activities, though clinical validation in patient cohorts with arrhythmias remains necessary.
Improves wearable PPG-to-ECG reconstruction in benchmark datasets; leaves open arrhythmia validation and clinical translation.
Continuous cardiac monitoring outside clinical settings requires signals that are both informative and practical to collect during daily life. Electrocardiography (ECG) provides rich information about cardiac rhythm and waveform morphology, while wearable photoplethysmography (PPG) is easier to acquire continuously but is only an indirect cardiovascular measurement and is highly sensitive to motion. We present BeatFlow-ECG, a conditional rectified-flow model for reconstructing single-channel ECG from synchronized PPG and inertial measurements. BeatFlow-ECG models reconstruction as conditional transport from noise to ECG using a convolutional encoder-decoder with a transformer bottleneck and explicit flow-time conditioning. Motion information is incorporated through IMU-derived conditioning features, motion-dependent loss weighting, and an easy-to-hard training curriculum. We evaluate the model under leave-one-subject-out protocols on PPG-DaLiA and WESAD. BeatFlow-ECG achieves the best results among the evaluated deterministic, adversarial, and diffusion-based baselines across all reported waveform and beat-timing metrics, with Pearson correlations of 0.983 and 0.986 and R-peak F1 scores of 0.946 and 0.955, respectively. Compared with Conditional DDPM-1D, L1 error decreases from 0.085 to 0.062 on PPG-DaLiA and from 0.074 to 0.055 on WESAD. Additional analyses on PPG-DaLiA show higher correlation in fixed R-peak-relative waveform regions and lower reconstruction error across low-, medium-, and high-motion subsets.
No takes yet. Share an insight, caveat, or question.
Kamel et al. (2026) studied Continuous ambulatory cardiac monitoring in healthy subjects. BeatFlow-ECG (conditional rectified-flow model for single-channel ECG reconstruction from synchronized PPG and IMU signals) vs. Deterministic, adversarial, and diffusion-based reconstruction baselines (AttnSeq, AttnSeq-Large, UNet1D, UNet1D-Large, Conditional DDPM-1D, and CardioGAN) was evaluated on ECG waveform reconstruction L1 error on PPG-DaLiA under leave-one-subject-out (LOSO) evaluation (95% CI 0.051-0.072, p=<0.01). BeatFlow-ECG reduced ECG reconstruction L1 error to 0.062 vs 0.085 with Conditional DDPM-1D (p<0.01) and achieved Pearson r of 0.983 on PPG-DaLiA.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: