Pulseflow: PPG Counterfactual Generation Via Latent Transport

Hung Manh Pham; Dong Ma; Bin Zhu; Pan Zhou

cs.LGSep 27, 2026

Pulseflow: PPG Counterfactual Generation Via Latent Transport

Authors: Hung Manh Pham, Dong Ma, Bin Zhu, Pan Zhou

Organizations: University of Cambridge, United Kingdom

Abstract

Photoplethysmography (PPG) has become an important modality for continuous cardiovascular monitoring, including atrial fibrillation (AF) detection. However, labeled AF recordings remain limited in many clinical settings, making model adaptation difficult when only limited target data are available. Generative modeling offers a natural way to alleviate this scarcity by synthesizing additional AF signals. Existing approaches, however, mainly generate samples that match the target condition without explicitly modeling how an observed source recording should be transformed, making it difficult to leverage abundant source recordings from a specific population or cohort for targeted augmentation. We introduce PulseFlow, a source-conditioned counterfactual generation framework that combines conditional representation learning with invertible latent transport to edit cardiac rhythm while retaining information from the source. Experiments across two clinical cohorts demonstrate effective rhythm transformation, measurable source correspondence, and improved AF classification under limited labels.

Figures & tables

	Rhythm	Realism				RR match
Method	AF conv	FD	FD	SDNN	RMSSD	$W_{1}$
	% $\to$	1 $\downarrow$	2 $\downarrow$	ms $\to$	ms $\to$	ms $\downarrow$
Real data	92.6	0.49	0.14	120.5	166.7	–
PPG-VAE [ 5 ]	69.0	216.7	278.9	127.1	186.7	133.5
Diff-SCM [ 16 ]	61.6	83.2	29.5	100.7	142.2	166.7
Rectified flow [ 7 ]	95.3	58.2	20.8	142.1	189.7	126.3

Table 1: Rhythm editing quality. Arrows mark proximity to real data (

\to

) or lower is better (

\downarrow

Augmentation	Setting 1	Setting 2	Setting 3
Real only	67.6 $\pm$ 8.9	76.2 $\pm$ 6.5	81.4 $\pm$ 5.1
PPG-VAE [ 5 ]	68.3 $\pm$ 9.6	77.4 $\pm$ 5.5	85.2 $\pm$ 5.7
Diff-SCM [ 16 ]	72.6 $\pm$ 8.7	79.5 $\pm$ 5.8	85.5 $\pm$ 3.5
Rectified flow [ 7 ]	76.2 $\pm$ 5.4	82.0 $\pm$ 2.6	87.0 $\pm$ 2.6
WGAN-GP [ 4 ]	73.7 $\pm$ 7.9	80.1 $\pm$ 3.6	85.8 $\pm$ 5.5
Latent DDPM [ 19 ]	77.2 $\pm$ 7.2	82.4 $\pm$ 3.5	86.9 $\pm$ 3.9

Table 2: VitalDB augmentation under scarce AF labels. AUROC (%), mean

\pm

SD over 5 partitions.

Variant	AF conv	FD	SDNN	RMSSD	$W_{1}$	Aug
	% $\to$	1 $\downarrow$	ms $\to$	ms $\to$	ms $\downarrow$	% $\uparrow$
w/o Stage 2	5.5	178.5	28.2	35.0	162.6	68.4
w/o $\bm{\varepsilon}$	93.6	47.5	115.0	155.1	112.8	75.0
w/o $\mathcal{L}_{\mathrm{cf}},\bm{\varepsilon}$	87.9	52.8	128.8	176.6	142.3	72.0
w/o $\mathcal{L}_{\mathrm{vital}}$	92.1	67.4	126.5	175.7	124.4	81.0
PulseFlow	93.1	45.5	117.0	158.5	113.6	83.6

Table 3: Ablation experiments. Arrows indicate proximity to real data (

\to

), lower is better (

\downarrow

); Aug denotes Setting 1 augmentation.

Explore similar work

PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation
May 9, 2026Xiaoda Wang, Minxiao Wang, Kaiqiao Han +10Cardiac Physiological Data
Department of Computer Science, Emory University · Department of Computer Science, University of California, Los Angeles · Nell Hodgson Woodruff School of Nursing, Emory University +2
VAMP-Diff: VampPrior Latent Diffusion for Photoplethysmography Modeling
May 17, 2026Fatemeh Ghasemi Balouei, Nathan Willemsen, Mahesh Banavar +1Remote Photoplethysmography Identity-Conditioned Latent Diffusion Models
Department of Electrical and Computer Engineering, Clarkson University, Potsdam, USA · Data Science, Worcester Polytechnic Institute, Worcester, USA · Department of Computer Science, Data Science, and Artificial Intelligence, Worcester Polytechnic Institute, Worcester, USA
ECGFlowCMR: Pretraining with ECG-Generated Cine CMR Helps Cardiac Disease Classification and Phenotype Prediction
Jan 28, 2026Xiaocheng Fang, Zhengyao Ding, Guangkun Nie +9Cardiac Magnetic Resonance Imaging Cardiac
State Key Laboratory of General Artificial Intelligence, Peking University · School of Intelligence Science and Technology, Peking University, Beijing, China · Polytechnic Institute of Zhejiang University, Zhejiang University, Hangzhou, China +4

May 9, 2026, eess.SPJ/K move · Enter open · S save

PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation

Xiaoda Wang, Minxiao Wang, Kaiqiao Han, Defu Cao+9

Department of Computer Science, Emory University · Department of Computer Science, University of California, Los Angeles · Nell Hodgson Woodruff School of Nursing, Emory University+2

Cardiac Physiological Data Rectified Flow

Open paper AskPDF arXiv

Electrocardiography (ECG) is the clinical standard for cardiac assessment but requires dedicated hardware that does not scale to daily-life monitoring. Photoplethysmography (PPG) is ubiquitous in wearables but lacks ECG-specific diagnostic morphology and is corrupted by motion and sensor noise. PPG-to-ECG generation aims to bridge this gap by recovering electrical morphology and timing from peripheral pulse signals. However, existing methods largely rely on statistical alignment and data-driven generation. They fail to explicitly structure the latent space around physiology-aware electro-hemodynamic factors and lack constraints from forward physiological dynamics. To address these challenges, we propose PG-LRF, a physiology-guided latent rectified flow framework. PG-LRF introduces an electro-hemodynamic simulator that co-models ECG and PPG through shared cardiac phase dynamics. Guided by this simulator, a Physiology-Aware AutoEncoder learns a structured electro-hemodynamic latent space. Then we integrate this simulator guidance into a PPG-conditioned latent rectified flow, enforcing ECG-side morphology consistency and ECG-to-PPG forward hemodynamic consistency during generative transport. Experiments on the large-scale MC-MED dataset demonstrate that PG-LRF significantly improves PPG-to-ECG generation and downstream cardiovascular disease classification, proving its ability to generate ECGs that are both signal-faithful and physiologically plausible under the ECG-to-PPG hemodynamic pathway

Abstract

Explore similar work

PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation

VAMP-Diff: VampPrior Latent Diffusion for Photoplethysmography Modeling

ECGFlowCMR: Pretraining with ECG-Generated Cine CMR Helps Cardiac Disease Classification and Phenotype Prediction