Variability analysis of fetal heart rate signals as obtained from abdominal electrocardiographic recordings

The present paper introduces an original method of digital signal processing for an automatic analysis of non-invasive abdominal ECG recordings on pregnant women starting from the 25th week of gestation. The procedure has been implemented on a DEC-VAX 750 digital computer at the Department of Electrical Engineering, Polytechnic of Milano and the signals are recorded at the Department of Obstetrics and Gynecology "L. Mangiagalli", University of Milano, Italy. The experimental results presented in here are still preliminary as only few cases have been considered up to now (about 20) and the goal of the paper is mainly focused on the algorithmic aspects of the whole procedure implemented in the computer and on the approach of heart rate variability (HRV) signal analysis both in the mother and in the fetus. Abdominal ECG lead processing is illustrated starting from the step of maternal (M) and fetal (F) QRS recognitions through linear digital filtering (derivative and low-pass FIR filter, Weber-Cappellini window) and weighted averaging techniques synchronized with maternal QRS's. Figure 1 a shows the original abdominal lead; figure 1 b the filtered signal for MQRS recognitions; figure 2 a the template of maternal cardiac cycle as obtained after the averaging operation synchronized with the instants of MQRS occurrence. The subtraction of the template results in the abdominal lead shown in figure 1 c in which the contribution of MECG is practically entirely reduced even in the case of MQRS and FQRS overlapping.(ABSTRACT TRUNCATED AT 250 WORDS)