Separation of fetal and maternal ECG complexes from a mixed signal using an algorithm based on linear regression

Area of Science:

• Biomedical Engineering
• Signal Processing
• Obstetrics

Background:

• Fetal heart rate (FHR) monitoring is crucial during labor.
• Current methods like fetal scalp ECG and Doppler ultrasound have limitations.
• Utilizing intrauterine ECG signals for FHR monitoring is challenging due to signal complexity and maternal ECG interference.

Purpose of the Study:

• To develop and present an algorithm for separating fetal and maternal electrocardiogram (ECG) signals.
• To enable accurate fetal heart rate measurement from intrauterine ECG recordings.
• To overcome the limitations of existing FHR monitoring techniques.

Main Methods:

• An algorithm was developed to detect and differentiate between fetal and maternal ECG complexes using intrauterine electrode data.
• Linear regression functions were employed to compare detected ECG complexes against predefined templates.
• The algorithm processes data in 10-second segments and includes artifact elimination for coincident maternal and fetal complexes.

Main Results:

• The algorithm successfully separates fetal and maternal ECG signals from intrauterine recordings.
• Two distinct FHR signals are generated for measurement.
• Artifacts caused by coincident maternal and fetal complexes are effectively eliminated.
• The method requires no additional maternal ECG input, relying only on the presence of two distinct ECG complex shapes.

Conclusions:

• The presented algorithm offers a novel approach to fetal and maternal ECG signal separation.
• It enhances the accuracy of fetal heart rate monitoring during labor using intrauterine signals.
• The algorithm's independence from extra maternal ECG input and its adaptability to abdominal electrodes present significant advantages.