A Null Space-Based Blind Source Separation for Fetal Electrocardiogram Signals

Area of Science:

• Biomedical Engineering
• Signal Processing
• Maternal-Fetal Medicine

Background:

• Extracting fetal electrocardiogram (FECG) signals non-invasively is crucial for prenatal monitoring.
• Existing methods often struggle with maternal electrocardiogram (MECG) interference and require complex processing.

Purpose of the Study:

• To introduce a new deterministic algorithm for non-invasive FECG signal extraction.
• To evaluate the performance of the proposed algorithm against established methods.

Main Methods:

• Development of a null space idempotent transformation matrix (NSITM) for signal separation.
• Utilizing MECG and FECG peak detection, control logic, and adaptive comb filtering for noise removal.
• Validation using real (Daisy, Physionet) and synthesized ECG databases.

Main Results:

• The NSITM algorithm effectively extracts FECG signals, comparable to PCA, FastICA, and PLP.
• Demonstrated significant performance improvements in synthesized data across various signal-to-noise ratios (SNRs).
• Showcased enhanced sensitivity (SE), accuracy (ACC), and positive predictive value (PPV) for FECG detection.

Conclusions:

• The NSITM algorithm presents a feasible and effective approach for non-invasive FECG extraction.
• The method offers superior performance, particularly in challenging signal conditions.
• This technique holds promise for improved prenatal diagnostic capabilities.