Related Experiment Video
Updated: Aug 14, 2026
06:30
A Multi-Electrode Array Platform for Modeling Epilepsy Using Human Pluripotent Stem Cell-Derived Brain Assembloids
Published on: September 27, 2024
Synthesis of Multi-Channel Fetal ECG using Generative Modeling
Swedha Sankaranarayanan1, Marta Regis2, Judith O E H van Laar3
1Electrical Engineering, Eindhoven University of Technology, Groene Loper 3, Eindhoven, Eindhoven, Noord-Brabant, 5612 AE, Netherlands.
Physiological Measurement
|August 12, 2026
Summary
Synthesizing realistic multi-channel fetal electrocardiography (fECG) signals addresses data scarcity for improved fetal monitoring. This method generates accurate fECG data, aiding algorithm development and enhancing non-invasive fetal heart rate assessment.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Maternal-Fetal Medicine
Background:
- Fetal electrocardiography (fECG) offers beat-to-beat cardiac information for assessing fetal response to uterine activity.
- Accurate fECG-derived cardiotocograms (CTGs) require high signal quality, which is challenging with non-invasive methods due to low signal-to-noise ratio and maternal interference.
- Invasive scalp electrodes provide reliable fECG but are only usable during labor, highlighting the need for non-invasive data.
Purpose of the Study:
- To address the scarcity of large datasets with simultaneous non-invasive multi-channel fECG and reference scalp electrode fECG.
- To propose a novel method for synthesizing realistic multi-channel fECG signals to augment existing datasets.
- To enable the development and validation of advanced fetal monitoring algorithms.
Main Methods:
- A generative model was employed to synthesize multi-channel fECG signals by modifying inter-beat intervals (TP segments) based on a target fetal heart rate (fHR).
- The duration of TP segments was adjusted to match target fHR inter-beat intervals, preserving the morphology of fetal QRS complexes.
- Signal quality indices (SQIs) and fHR estimates were used to evaluate the fidelity of the synthesized fECG signals.
Main Results:
- High correlations (r > 0.9) were observed between input and synthesized signals for most SQIs, indicating low mean absolute differences.
- The synthesized fHR standard deviation showed a low error of ±3.35 BPM compared to the target fHR.
- Morphology-based SQIs confirmed the preservation of fetal QRS complex morphology, validating the synthesis approach.
Conclusions:
- The proposed method successfully generates realistic multi-channel fECG signals that accurately reflect target fHR patterns.
- This approach provides a valuable tool for data augmentation and algorithm development in non-invasive fetal monitoring.
- The synthesized data can enhance the robustness of data-driven and source separation methods for improved fetal health assessment.