Artificial Intelligence–Based Analysis Of Early Subtle Changes In Fetal Movements For Predicting High-Risk Pregnancy Outcomes: A Systematic Review

Abstract

Background: Early detection of fetal movement (FM) irregularities is a critical marker for fetal well-being and can predict adverse pregnancy outcomes. Recent advances in artificial intelligence (AI) and machine learning (ML) have transformed traditional fetal monitoring, enabling continuous, objective, and predictive fetal health assessment.

Objective: This systematic review synthesizes evidence on AI-driven techniques used for early fetal movement detection and their capacity to predict high-risk pregnancies.

Methods: Following PRISMA 2020 guidelines, 11 peer-reviewed empirical studies published between 2018 and 2025 were analyzed. Studies were identified through PubMed, Scopus, IEEE Xplore, and Web of Science. Data extraction included AI models, modalities, performance metrics, and clinical correlations.

Results: Studies applied deep learning, convolutional neural networks (CNNs), ensemble algorithms, and wearable sensor-based models. Detection accuracy ranged from 82% to 95%, with models such as CatBoost and multilayer perceptrons achieving the highest predictive power. MRI- and fMRI-based approaches demonstrated biological insights linking neural maturation to motor behavior, while wearable accelerometer systems offered real-time, noninvasive monitoring with >90% accuracy.

AI-assisted ultrasound improved diagnostic precision and reduced operator workload.
Conclusion: AI technologies effectively quantify subtle fetal movements, correlate them with developmental and maternal variables, and enhance prediction of high-risk pregnancies. Integration of AI-based systems in obstetrics could revolutionize early fetal assessment and preventive maternal care.