Fetal heart rate estimation from a single phonocardiogram signal using non-negative matrix factorization

Area of Science:

• Biomedical Engineering
• Maternal-Fetal Medicine
• Signal Processing

Background:

• Fetal heart rate (FHR) monitoring is crucial for assessing fetal well-being.
• The cardiotocogram (CTG) is the standard non-invasive method but has limitations.
• Fetal phonocardiogram (fetal PCG) offers an alternative, but requires complex sound identification.

Purpose of the Study:

• To develop a new methodology for FHR estimation from fetal PCG.
• To eliminate the need for distinguishing S1 and S2 heart sounds in fetal PCG analysis.
• To utilize a single cardio-microphone for simplified data acquisition.

Main Methods:

• Non-negative Matrix Factorization (NMF) applied to the spectrogram of fetal PCG signals.
• Fetal PCG signals treated as a source-filter model.
• Analysis performed using a single cardio-microphone without explicit heart sound labeling.

Main Results:

• The proposed NMF-based method demonstrated satisfactory performance on a preliminary dataset.
• FHR estimations derived from fetal PCG showed a high correlation (around 90%) with the reference CTG.
• The method successfully estimated FHR without the need to distinguish specific cardiac sound events.

Conclusions:

• The novel NMF approach provides a viable and simplified method for FHR estimation from fetal PCG.
• This technique overcomes the limitations of traditional fetal PCG analysis by not requiring heart sound identification.
• The high correlation with CTG suggests potential for improved, less complex fetal well-being monitoring.