Recurrence quantification analysis applied to fetal heart rate variability with fetal magnetocardiography

This work presents the application of recurrence quantification analysis (RQA) for fetal heart rate variability (FHRV) data. RQA is non-linear tool that has several advantage compared with the preferred analysis based on time and frequency domain. Fetal cardiac signals were measured with fetal magnetocardiography (fMCG) which is a technique with high temporal resolution that enables precise detection of the R-R intervals. We applied the tool to 20 fMCG recordings from pregnant women between 29 to 38 weeks gestation. We calculated different parameters of the RQA (REC, DET, LAM, Lmax, Vmax, and MDL). We performed an analysis to determinate the appropriate values of the embedding dimension (m), time delay (tau) and cut-off distance (epsilon) based on all recordings. We demonstrated the applicability of the RQA for a clinical case of a fetus with supraventricular tachycardia (SVT) and found higher values on Lmax and Vmax metrics compared with a normal fetus. We suggested that m = 2, tau = 20, epsilon = 10% are appropriate settings for RQA, but validity of those settings by comparing the RQA measures of randomly shuffled data may be required. In future, we plan to expand this approach to recordings with cardiac anomalies and perform a comparison with the values obtained from the preferred analysis of FHRV.