Related Experiment Video
Updated: Jul 8, 2026
Electromyometrial Imaging of Uterine Contractions in Pregnant Women
Published on: May 26, 2023
Analysis of electrohysterogram signals for predicting obstetric outcome using machine learning methods: a scoping
Rubana H Chowdhury1,2, Roma Sultana3, Mithila Arman4
1Department of Electrical and Electronic Engineering, Chittagong University of Engineering & Technology, Chattogram, 4349, Bangladesh. rubanachy@gmail.com.
Purpose:
Efficiently detecting obstetric outcomes, such as preterm birth or mode of delivery, is crucial for enhancing mother and newborn health. Proactive screening, identification, and prevention in asymptomatic pregnant women exhibiting risk factors for preterm birth or c-section delivery can mitigate incidence and fatality rates.
Methods:
This study thoroughly reviewed prediction models for obstetric outcomes, described the EHG signal acquisition protocols, pre-processing methods, feature extraction from EHG, and model properties, and compared their quality to establish the most effective prediction model for clinical decision-making. For this study, the biomedical databases (PubMed, Scopus, Web of Science, Embase) of published publications were searched from December 2000 to February 2025. In addition to electrohysterography, other search terms to consider are electrohysterogram, uterine electromyography, Term-preterm labor, birth delivery mode, and EHG in machine learning.
Results:
Based on a literature review, the prevailing recording technique for acquiring EHG signals across various applications, including pregnancy monitoring, preterm risk evaluation, and birth delivery mode detection, commonly employs four bipolar electrodes. A bandpass filter of minimum bandwidth of 0.1 to 4 Hz is most commonly used for pre-process the EHG signal. High discriminative performance was reported in 5 studies, with an Area Under the Curve (AUC) ranging from 0.93 to 0.99. A single classifier may suffice for predicting obstetric outcomes using an EHG signal, eliminating the need for a combined classifier. A total of 98.5% of the studies exhibited a high risk of bias in the analysis domain, primarily due to the limited sample size and the absence of external validation.
Conclusion:
This review will familiarize academics and obstetricians with the comprehensive EHG analytical process and its prospective uses in clinical decision support systems.
