An optimized hybrid methodology for non-invasive fetal electrocardiogram signal extraction and monitoring

Background and objective: Electronic fetal heart monitoring is currently used during pregnancy throughout most of the developed world to detect risk conditions for both the mother and the fetus. Non-invasive fetal electrocardiogram (NI-fECG), recorded in the maternal abdomen, represents an alternative to cardiotocography, which could provide a more accurate estimate of fetal heart rate. Different methodologies, with varying advantages and disadvantages, have been developed for NI-fECG signal detection and processing. Methods: In this context, we propose a hybrid methodology, combining independent component analysis, signal quality indices, empirical mode decomposition, wavelet thresholding and correlation analysis for NI-fECG optimized signal extraction, denoising, enhancement and addressing the intrinsic mode function selection problem. Results: The methodology has been applied in four different datasets, and the obtained results indicate that our method can produce accurate fetal heart rate (FHR) estimations when tested against different datasets of variable quality and acquisition protocols, on the FECGDARHA dataset our method achieved average values of Sensitivity = 98.55%, Positive Predictive Value = 91.73%, F1 = 94.92%, Accuracy = 90.91%, while on the ARDNIFECG dataset it achieved average values of Sensitivity = 92.96%, Positive Predictive Value = 91.66%, F1 = 93.60%, Accuracy = 90.45%. Conclusions: The proposed methodology is completely unsupervised, has been proven robust in different signal-to-noise ratio scenarios and abdominal signals, and could potentially be applied to the development of real-time fetal monitoring systems.