Cross-Sectional Observational Study of Typical in utero Fetal Movements Using Machine Learning

Abstract

Early variations of fetal movements are the hallmark of a healthy developing central nervous system. However, there are no automatic methods to quantify the complex 3D motion of the developing fetus in utero. The aim of this prospective study was to use machine learning (ML) on in utero MRI to perform quantitative kinematic analysis of fetal limb movement, assessing the impact of maternal, placental, and fetal factors. In this cross-sectional, observational study, we used 76 sets of fetal (24-40 gestational weeks [GW]) blood oxygenation level-dependent (BOLD) MRI scans of 52 women (18-45 years old) during typical pregnancies. Pregnant women were scanned for 5-10 min while breathing room air (21% O2) and for 5-10 min while breathing 100% FiO2 in supine and/or lateral position. BOLD acquisition time was 20 min in total with effective temporal resolution approximately 3 s. To quantify upper and lower limb kinematics, we used a 3D convolutional neural network previously trained to track fetal key points (wrists, elbows, shoulders, ankles, knees, hips) on similar BOLD time series. Tracking was visually assessed, errors were manually corrected, and the absolute movement time (AMT) for each joint was calculated. To identify variables that had a significant association with AMT, we constructed a mixed-model ANOVA with interaction terms. Fetuses showed significantly longer duration of limb movements during maternal hyperoxia. We also found a significant centrifugal increase of AMT across limbs and significantly longer AMT of upper extremities <31 GW and longer AMT of lower extremities >35 GW. In conclusion, using ML we successfully quantified complex 3D fetal limb motion in utero and across gestation, showing maternal factors (hyperoxia) and fetal factors (gestational age, joint) that impact movement. Quantification of fetal motion on MRI is a potential new biomarker of fetal health and neuromuscular development.

Keywords: Fetal movement; In utero fetal magnetic resonance imaging; Maternal hyperoxia.

Conflict of interest statement

The authors have no conflicts of interest to declare.

The Author(s). Published by S. Karger AG, Basel.

Figures

Fig. 1

Flowchart of our study sample shows inclusion and exclusion.

Fig. 2

Difference in AMT between lower and upper extremities before 30 GW. a An example slice of the MR volume. b The fetal pose generated from labeled key points. c 3D masks of the fetal body and uterus. d The change of velocity over time, shaded regions indicate velocities are greater than the threshold.

Fig. 3

Overview of major histogenic events (a), scatterplots showing the mean AMT (upper row) and difference in AMT between upper and lower extremities for gestational age (bottom row) (b), and boxplots showing the difference in AMT between hyperoxia and normoxia (c). a Neurogenic events relevant for movement. b Upper row: mean AMT for upper and lower extremities, by fetal age group (rounded to the nearest week), averaged across all subjects, joints, and other covariates. Lower row: ratio between the movements of upper and lower extremities (difference in %) and 95% confidence intervals for the fetal age group, for a given subject, joint, and other covariate values. Shaded vertical bars mark periods during which significant differences between movements of upper and lower extremities were found, as indicated by non-overlap of confidence interval with dashed horizontal line at ratio 1.0. c The data points are adjusted for the variables in the model. Middle quartile/median (the horizontal line through the box); quartile boundaries (top and the bottom of the box). “Whiskers” extend to the farthest data point within the 1.5 × interquartile range of the box. For a given subject, joint, and other covariate values, the mean absolute movement time was 20.7% higher in the hyperoxic state (95% CI 13.9–27.9%, p < 0.0001).