SCTD-ICA: A ICA-Based Approach for Fetal ECG Extraction From Single Channel Abdominal ECG.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Fetal electrocardiogram (FECG) signals are crucial for assessing fetal cardiac health.
• Maternal abdominal ECG (AECG) signals are complex, containing maternal ECG (MECG) and other noise, making FECG extraction difficult.
• Traditional Independent Component Analysis (ICA) requires multi-channel data, limiting its application in single-channel FECG analysis.

Purpose of the Study:

• To develop and validate a novel pipeline for extracting FECG signals from single-channel abdominal recordings.
• To overcome the multi-channel data limitation of ICA for FECG extraction.
• To enable automated identification and analysis of fetal cardiac activity from non-invasive recordings.

Main Methods:

• Proposed a Single Channel Time Delay ICA (SCTD-ICA) pipeline.
• Utilized time-delay mapping to transform single-channel data into a multi-dimensional format suitable for ICA.
• Applied ICA to the transformed data for fetal component separation.
• Employed power spectrum analysis for automated fetal component identification.

Main Results:

• Achieved high F1 metrics for fetal QRS detection: 96.14% on the ADFECGDB dataset and 95.76% on the PCDB dataset.
• The SCTD-ICA pipeline demonstrated performance comparable to state-of-the-art deep learning methods.
• Successfully extracted FECG signals from single-channel abdominal recordings.

Conclusions:

• The proposed SCTD-ICA pipeline offers an effective solution for FECG extraction from single-channel data.
• This method is suitable for continuous maternal and fetal health monitoring.
• The approach provides a viable alternative to multi-channel systems and complex deep learning models.