An advanced algorithm for fetal heart rate estimation from non-invasive low electrode density recordings

Area of Science:

• Biomedical Engineering
• Signal Processing
• Obstetrics

Background:

• Non-invasive fetal electrocardiography (fECG) for monitoring fetal well-being remains a significant research challenge.
• An annotated fECG dataset spurred an international challenge, highlighting the need for advanced detection algorithms.
• Previous algorithms achieved top-ten performance but required further refinement for improved accuracy.

Purpose of the Study:

• To present an improved algorithm for identifying fetal QRS complexes from non-invasive abdominal ECG signals.
• To enhance the accuracy and reliability of fetal cardiac monitoring using advanced signal processing techniques.
• To evaluate the algorithm's performance against established metrics on multiple datasets.

Main Methods:

• Developed an algorithm based on maternal QRS complex subtraction via synchronized averaging.
• Filtered maternal P and T waves and applied independent component analysis (ICA) for fetal QRS enhancement.
• Analyzed fetal and maternal RR time series to correct potential annotation errors and improve pseudoperiodicity.

Main Results:

• The improved algorithm achieved high median sensitivity (0.982) and positive predictivity (0.976) on dataset A.
• Performance evaluation on datasets B and C demonstrated consistent accuracy across different patient cohorts.
• The algorithm ranked 5th and 4th in the challenge for events 1 and 2, outperforming many open-source entries.

Conclusions:

• The enhanced non-invasive fetal ECG algorithm significantly improves the detection of fetal QRS complexes.
• The method offers a robust and accurate approach for fetal cardiac monitoring, particularly in challenging signal conditions.
• This advancement holds promise for more reliable prenatal diagnostics and improved obstetric care.