Related Experiment Video
Updated: Jan 3, 2026
Author Spotlight: Advancing Labor Management Through Electromyometrial Imaging for Understanding Uterine Contractions
Published on: May 26, 2023
Preliminary Study on the Efficient Electrohysterogram Segments for Recognizing Uterine Contractions with
Jin Peng1, Dongmei Hao1, Haipeng Liu2
1College of Life Science and Bioengineering, Beijing University of Technology, Intelligent Physiological Measurement and Clinical Translation, Beijing International Platform for Scientific and Technological Cooperation, Beijing 100024, China.
This study used convolutional neural networks (CNNs) to analyze electrohysterogram (EHG) signals for monitoring uterine contractions (UCs). Shorter EHG segments around the TOCO peak improved UC recognition accuracy, showing potential for non-invasive pregnancy monitoring.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Obstetrics
Background:
- Uterine contractions (UCs) are crucial indicators of pregnancy progression.
- Electrohysterogram (EHG) signals reflect uterine electrical activity, offering a non-invasive method for UC monitoring.
- Convolutional Neural Networks (CNNs) show promise for analyzing complex biological signals like EHG.
Purpose of the Study:
- To evaluate the effectiveness of different electrohysterogram (EHG) signal segments for recognizing uterine contractions (UCs) using a convolutional neural network (CNN).
- To determine optimal EHG segment durations and locations for accurate UC detection.
- To assess the performance of a CNN model in classifying UC and non-UC EHG segments.
Main Methods:
- Utilized an open-access Icelandic 16-electrode EHG database comprising 122 recordings from 45 pregnant women.
- Developed a CNN model using 7136 UC and 7136 non-UC EHG segments (60s duration) from 107 recordings.
- Evaluated the CNN model using fivefold cross-validation, measuring sensitivity (SE), specificity (SP), and accuracy (ACC).
- Tested the model's performance on shorter EHG segments (10s, 20s, 30s) around the TOCO peak.
Main Results:
- The CNN model achieved an average SE of 0.82, SP of 0.93, and ACC of 0.88 for 60s EHG segments.
- Shorter EHG segments (10s, 20s, 30s) around the TOCO peak demonstrated higher SE and ACC compared to other segments of similar duration.
- Specifically, 20s EHG segments around the TOCO peak yielded superior SE compared to 10s and 30s segments on the same side of the TOCO peak.
Conclusions:
- The developed CNN-based method effectively identifies efficient EHG segments for recognizing uterine contractions (UCs).
- This approach offers a promising tool for non-invasive UC monitoring during pregnancy.
- Optimizing EHG segment selection enhances the accuracy of CNN-based UC detection systems.
