Fetal heart rate (FHR) monitoring is done for accessing fetal wellbeing during antepartum and intrapartum phases. Although noninvasive fetal electrocardiogram (NIfECG) is a potential data acquisition method for FHR, extraction of fetal electrocardiogram (ECG) from the abdominal ECG (aECG) is one of the major challenging research areas. This chapter proposed and assessed a method suitable for single channel based on principal component analysis (PCA) for extracting fetal ECG. Maternal R peaks and fetal R peaks were detected using Pan Tomkins algorithm (PTA) and improved Pan Tomkins algorithm (IPTA), respectively. Performance of fetal QRS detection is assessed using two open-access databases available online. The method shows satisfactory performance when compared with similar methods and makes it suitable for using a single channel system.
TopIntroduction
According to World Health Organization’s (WHO) health statistics published in 2018, globally 2.5 million neonatal deaths occurred in during year. Preterm births, intrapartum-related complications (such as birth asphyxia) are some of major causes of these deaths (World Health Statistics, 2018). Fetal monitoring performed during antepartum and intrapartum phase plays important role to determine fetal wellbeing which ultimately aims towards accurate and timely diagnosis of hypoxia which is considered to be early stage of asphyxia. Fetal Heart Rate (FHR) and heart rate variability are considered to be significant parameters and also critical to diagnose various pathological conditions such as fetal distress, fetal asphyxia, fetal arrhythmias, bradycardia, and oxygen deficiency (Van Geijn et al., 1991), (Kennedy, 1998) and (Sameni R. et al., 2010).
Cardiotocograph (CTG) is the most common noninvasive fetal monitoring device widely used in clinics which captures FHR and uterine contractions. Although a CTG is normally used for FHR detection during antepartum and intrapartum phase (Rogers et al., 2010), repetition and long-time monitoring are not advisable due to ultrasound irradiation exposure is not completely safe for the fetus (Barnett et al., 2001). Another disadvantage of intrapartum FHR monitoring using the CTG is more maternal heart rate and FHR ambiguity compared to the electrocardiogram (ECG) (Reinhard, 2013).
FHR monitoring based on ECG can be alternative to CTG which can be done using invasive way which uses an invasive electrode attached to fetal scalp which is inserted through the dilated cervix. Although fetal ECG captured through scalp electrode provides a very good accuracy, it has apparent limitation that it requires dilated cervix therefore limited use during intrapartum phase and it also increases the chances of infection to both the fetus and the mother.
Alternative noninvasive method to detect fetal ECG is from maternal abdominal signal and is most suitable for long term monitoring of fetal health (Behar, 2016). We have also proposed Noninvasive fetal ECG (NIfECG) as potential data acquisition method for FHR. (Kurtadikar et al., 2021). Major components of abdominal ECG (aECG) are fetal ECG and maternal ECG (mECG), however fetal and maternal movements, mothers abdominal muscle artefacts, uterine contractions, maternal respiration, electrical noise from the device and power line affects quality of aECG. Extracting fetal ECG from aECG is challenging due to the fact that fetal ECG and mECG overlaps temporally as well as in the frequency domain as shown in Figure 1(Behar, 2016). The amplitude of mECG is 2-10 times that of fetal ECG, and the frequency of the QRS band of the ECG overlaps for the fetus and the mother which makes it difficult to separate these signals with simple signal processing techniques (Mujumdar et al., 2019).
Figure 1. Power spectral density distribution (Burg method, order 20) for 5 min of scalp electrode ECG and 5 min of adult ECG. Notice the frequency overlap between the adult and fetal scalp ECG signals particularly in the frequency band of the QRS (Behar, 2016)