JNU-CTG: A Large-Scale Continuous Cardiotocography Dataset for Fetal Well-being Assessment

Cardiotocography (CTG) is a widely used clinical tool for antepartum fetal surveillance, but the development and validation of automated analysis methods have been limited by the scarcity of large-scale, publicly accessible CTG waveform datasets. Here, we present the Jinan University CTG (JNU-CTG) dataset, a large-scale, multimodal CTG resource comprising 20,769 continuous 30-minute recordings sampled at 4 Hz from 12,606 unique pregnancies collected during routine antenatal monitoring. The dataset includes raw fetal heart rate and uterine contraction waveforms, expert-annotated CTG morphological patterns, and clinically relevant neonatal outcomes, including 1-, 5-, and 10-minute Apgar scores and neonatal asphyxia diagnoses. In addition, JNU-CTG provides maternal-fetal demographic characteristics, pregnancy-related complications, and more than 100 pre-computed features covering temporal, frequency-domain, morphological, and deep learning representations. Benchmark experiments using a one-dimensional convolutional neural network further validate the utility of JNU-CTG for developing and evaluating automated CTG interpretation models. JNU-CTG establishes an open data resource to facilitate reproducible research in fetal monitoring, obstetric artificial intelligence, and maternal–newborn health.