Raw fetal PCG dataset contaminated with Mother's PCG

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Fetal phonocardiogram (fPCG) signals are crucial for monitoring fetal well-being.
• Maternal heart sounds often contaminate fPCG recordings, posing a significant challenge for analysis.
• Existing datasets may lack sufficient real-world noise and variability for robust algorithm development.

Purpose of the Study:

• To present a novel, openly available dataset of raw fetal phonocardiogram (fPCG) signals.
• To provide fPCG data naturally contaminated with maternal heart sounds for algorithm testing.
• To facilitate research in fetal monitoring and biomedical signal processing.

Main Methods:

• Recordings of fetal heart sounds using a non-invasive stethoscope and microphone setup.
• Simultaneous acquisition of reference fetal and maternal heart rates via Doppler ultrasound and pulse oximeter.
• Data collected from eight pregnant participants at different gestational stages (32-36+ weeks).
• Signal storage in standard formats (MATLAB .mat, CSV) with accompanying metadata.

Main Results:

• A comprehensive dataset of raw fPCG signals with inherent maternal sound contamination.
• Inclusion of demographic details and stethoscope placement notes for enhanced usability.
• Data availability in standard formats for widespread accessibility and application.

Conclusions:

• The presented fPCG dataset addresses the challenge of maternal signal overlap and background noise.
• This resource is valuable for developing and validating signal separation and denoising algorithms.
• It supports advancements in non-invasive fetal monitoring and biomedical signal processing research.