Analysis of Pregnancy Progression in Term Condition Using Propagation Features and Uterine EMG Measurements

Pregnancy involves various anatomical and physiological changes, with weak and inefficient uterine contractions during early stages. As labor approaches, uterine activity becomes more coordinated due to enhanced synchronization of electrical signals. Monitoring this progression is critical for improving maternal and fetal health through early detection of complications and timely interventions. Uterine electromyography (uEMG) is a noninvasive technique for assessing uterine electrical activity and has emerged as a promising tool for tracking pregnancy progression. This study aims to evaluate synchronization measures derived from multi-channel uEMG signals to characterize pregnancy progression. uEMG signals obtained from the publicly available databases are considered during before (T1) and after (T2) 26 weeks of gestation from term delivery. Signals from three bipolar channels, with a fourth derived channel are preprocessed using a four-pole Butterworth bandpass filter (0.3-3 Hz) to mitigate noise. Propagation features, including maximum cross-correlation, mean coherence, peak coherence, imaginary coherence, and phase locking value (PLV) are extracted across all channel pairs. Random forest (RF)-based feature importance ranking is employed for feature selection, and machine learning classifiers such as RF, Adaptive Boosting, and Support Vector Machine are used for classification. Results show that propagation features are able to characterize the pregnancy progression. Coherence-based measures decrease with increasing gestation, whereas PLV consistently increases in both inter- and intra-subject analyses, indicating enhanced localized phase alignment. The RF model achieves 73.3% accuracy in intra-subject and 67.4% accuracy in inter-subject analyses. These findings suggest that propagation features derived from uEMG can effectively characterize pregnancy progression and may aid in monitoring uterine physiological changes throughout gestation.