Towards phase-rectified signal averaging technique for fetal heart-rate signal processing

Pregnancy is a truly delicate condition, both for the woman and for the foetus, which are subjected to a series of risks from the embryogenesis and the foetus formation, up to the childbirth. Each of these stages needs a precise set of tests, analysis and monitoring procedures, to limit the risks and check that the physiological parameters are maintained. In this present study, the focus is on the labour, so what nearly precedes the delivery itself, since this phase could be crucial for the positive outcome of the birth. Problems that could present in this time, as a reduced oxygenation of the foetus, could, in fact, bring to pathological states as hypoxia or acidosis, which could be extremely negative events for the child wellbeing. Monitoring the foetus’s fetal heart rate, together with the uterine contractions recording during labour, using a cardiotocograph, is a valid solution to keep the FHR parameters of the foetus under control and to determine the occurrence of eventual worrying features of the tracks. However, this procedure is still based on the personal interpretation of the medical personnel, which follows the FIGO guidelines for CTG feature identification, such as the presence of accelerations and decelerations, but keeps a certain level of subjective evaluation, based, for example, on the different experience or formation of the doctor. To overcome this subjectivity, the aim of the research is to find an automatic method to correlate the occurrence of specific CTG features to certain risk factors, to allow clinicians to identify and predict critical conditions, in which it is necessary to intervene with a Cesarean cut. This study, in a particular way, wants to determine the values of the deceleration reserve (DR), which is an index calculated from the sum of acceleration (AC) and deceleration capacities (DC). These two parameters are derived by the application of a methodology called Phase Rectified Signal Averaging (PRSA), which allows to extract and average intervals of signal, in our case foetal RR series, obtained from the CTG recordings of the FHR. Once DR is derived for each subject, it is correlated to the three neonatal post-partum parameters, pH, Base Excess and Apgar score, which provide information about the health state of the child as soon after the birth accomplishment. R and p-values are obtained for each couple DR/pH, DR/BE, and DR/Apgar, showing, respectively, values of 0.05 for the correlation DR/pH, -0.04 for DR/BE, and 0.14 for the R-value and of 0.09 for DR/pH, 0.14 for DR/BE, and < 10-5 for DR/Apgar for the p-value. These results express a lack of significant and reliable correlations between DR and any of the three neonatal post-partum parameters, meaning that the aim of the study seems not to be accomplished, using this specific approach. Further evaluations could be done, for example considering a different index instead of DR, as well as the possibility to use different parameters to correlate. The possibility to correlate FHR features, analysed through CTG recordings, to neonatal post-partum parameters is therefore not completely excluded but left to further and different approaches. A future perspective could be, for example, the possibility of considering introducing in the method the recording of uterine contractions, which were not evaluated in this study