A novel algorithm based on ensemble empirical mode decomposition for non-invasive fetal ECG extraction.

Area of Science:

• Biomedical Engineering
• Maternal-Fetal Medicine
• Signal Processing

Background:

• Non-invasive fetal electrocardiography (fECG) is a crucial technique for monitoring fetal well-being during pregnancy and labor.
• Abdominal fECG recordings are contaminated by maternal signals, primarily the maternal electrocardiogram (mECG), posing a significant challenge for accurate fetal assessment.
• Effective filtering methods are essential to isolate the fetal signal for reliable analysis.

Purpose of the Study:

• To develop and evaluate a novel signal processing method for extracting fetal electrocardiographic signals from abdominal recordings.
• To suppress maternal electrocardiographic interference and improve the accuracy of fetal QRS complex detection and heart rate determination.
• To assess the performance of the proposed method on established fECG databases.

Main Methods:

• A hybrid approach combining Independent Component Analysis (ICA), Recursive Least Squares (RLS), and Ensemble Empirical Mode Decomposition (EEMD) was employed.
• The method was validated using two distinct datasets: Fetal Electrocardiograms, Direct and Abdominal with Reference Heartbeats Annotations, and the PhysioNet Challenge 2013 database.
• Performance was evaluated using standard statistical metrics including accuracy, sensitivity, positive predictive value, and F1-score.

Main Results:

• On the Fetal Electrocardiograms, Direct and Abdominal with Reference Heartbeats Annotations database, the method achieved an average accuracy of 92.75%, with 11 out of 12 recordings exceeding 80% accuracy.
• Testing on the PhysioNet Challenge 2013 database yielded an average accuracy of 78.24%, with 17 out of 25 recordings achieving over 80% accuracy.
• The method also demonstrated high accuracy in non-invasive ST segment analysis on selected records.

Conclusions:

• The proposed combination of ICA, RLS, and EEMD offers a robust and effective solution for extracting fetal electrocardiographic signals from abdominal recordings.
• This advanced signal processing technique significantly enhances the accuracy of fetal heart rate monitoring and QRS complex detection, outperforming previous methods.
• The method holds promise for improving non-invasive fetal monitoring, aiding in better clinical decision-making during pregnancy and delivery.