AI Model Enhances Fetal Heart Rate Monitoring and Variability Assessment

Researchers have developed an AI-based FHrCTG model for fetal heart rate (FHR) monitoring that reconstructs signals and assesses variability. The model was pre-trained on 558,412 unlabeled data points and refined with 7,266 expert-reviewed entries. It uses an Intersection Overlapping Labels (IOL) approach to transform rate analysis into categorical judgments. Testing showed high sensitivity and specificity for detecting critical FHR decelerations (89.13% and 87.78%) and accelerations (62.5% and 92.04%). The model addresses limitations of traditional methods, including equipment performance, data transmission, and subjective assessments.