Deep Convolutional Long Short-Term Memory Network for Fetal Heart Rate Extraction.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Fetal electrocardiography (ECG) offers an alternative to standard fetal monitoring.
• Extracting fetal heart rate (HR) from abdominal ECG signals requires accurate R-peak detection.
• Low signal-to-noise ratio periods, often due to uterine activity, challenge HR detection.

Purpose of the Study:

• To develop a deep learning model for direct fetal HR prediction from multichannel fetal ECG.
• To improve the accuracy and robustness of fetal HR detection, especially in challenging conditions.

Main Methods:

• A hybrid deep learning model combining convolutional neural networks (CNNs) and long short-term memory (LSTM) networks was developed.
• The model directly predicts fetal HR from multichannel fetal ECG signals.
• The network was trained on labor-recorded data and evaluated on independent test datasets.

Main Results:

• The proposed CNN-LSTM algorithm achieved a positive percent agreement of 92.1% on a test dataset.
• It achieved 98.1% positive percent agreement on the 2013 Physionet /Computing in Cardiology Challenge set-A.
• The method outperformed a leading state-of-the-art signal processing algorithm.

Conclusions:

• The combined CNN-LSTM network provides an effective method for direct fetal HR prediction.
• This deep learning approach enhances fetal monitoring accuracy, particularly during noisy signal conditions.
• The algorithm demonstrates superior performance compared to existing signal processing techniques.