(A) to (G): ROC curves and the corresponding AUC values from 7-fold cross-validation of the results along 6 hours of 1024Hz data across 7 preterm fetal sheep (42 hours total) using 11, 9, 7 and 5 layers in the proposed WF-CNN classifier. The data for each WF-CNN classifier are presented as mean ± SD in the boxplot (H) demonstrating reduced accuracy and increased variability with fewer layers.