Fetal cardiac time intervals: validation of an automatic tool for beat-to-beat detection on fetal magnetocardiograms

Fetal magnetocardiography (fMCG) allows the non - invasive registration of fetal cardiac activity. This technique, combined with the use of independent component analysis (ICA) for signal processing, allows reconstructing of reliable fetal cardia c traces. Low noise fetal signals can be used to evaluate fetal cardiac time intervals (fCTI) , useful to monitor fetal heart function. In this work we present a method for the automatic detection of cardiac waves (ACWD); it was validated on 45 fMCG data se ts of normal fetuses with gestational age from 22 to 37 weeks. The outcomes of the automatic procedure were compared with those of a manual procedure performed by three independent operators on rhythm strips of 100 consecutive cardiac cycles for each data set. Distances between the wave boundaries detected with the two methods were statistically estimated using confidence intervals: differences were always comparable to those that could be obtained from different investigators' estimates. Statistical correl ation between fCTI quantified with ACWD and with a manual procedure was assessed using the parametric two - tailed Pearson's correlation test, significance level at α = 0.01. The automatic procedure showed a computation time decrease in the ratio of approxim ately 1:600 with respect to the manual procedure performed on the same number of beats.