FRACTAL ANGLE BASED DIFFERENTIATION OF TERM PREGNANCIES USING UTERINE ELECTROMYOGRAPHIC SIGNALS

Abstract

Uterine Electromyography (uEMG) is a non-invasive technique that provides quantitative measure of uterine activity from the abdominal surface. In this work, an attempt has been made to investigate Term (gestational age > 37 weeks) uEMG signals using Adaptive Fractal Analysis (AFA). For this, the signals obtained in second and third trimesters are considered and subjected to AFA. The fluctuation function is computed and the corresponding linear scaling regions are identified based on Chi-square statistic, standard error of slope, and coefficient of determination. Angle-based features from multiple scaling regions namely, inter-fractal angle and, short- and long-term fractal angles are extracted and are used for further analysis. The obtained results demonstrates that AFA approach can characterize the Term signals during varied gestational ages. All features show significant differences (p < 0.05) in both groups. Feature values suggest that the third trimester signals possess more correlated and smoother fluctuations when compared to second trimester signals. This is attributed to the increased coordination of uterine contractions as delivery approaches. Hence, it appears that the proposed adaptive angle-based fractal features could be potential biomarkers in analyzing the muscle contractions associated with Term pregnancies.