Validation of beat by beat fetal heart signals acquired from four-channel fetal phonocardiogram with fetal electrocardiogram in healthy late pregnancy

Figure 1. 3D model of the phonogram sensor, from the top to the bottom: top cover and container.

Figure 2. A schematic diagram showing the locations of phonogram sensors. Ch1, Ch2, Ch3 and Ch4 are four phonogram sensors.

Figure 3. locations of fetal ECG electrodes and phonocardiogram sensors for validation experiment.

Figure 4. Four-channel unprocessed fetal phonocardiogram signals with 4 FPCG sensors. Ch 1–4 show maternal breathing movement (as low frequency baseline oscillation), and maternal and fetal heart movements (as high frequency oscillations).

Figure 5. A 20 second example of: extracted fetal heart sound signal (Denoised Fetal Heart Sounds), simultaneously acquired maternal (mECG) and fetal (fECG) ECG signals, extracted maternal heart sound (Denoised Maternal Heart Sounds) and extracted maternal breathing movement (Mother Breathing) signals from 4-channel phonocardiogram sensors, fetal ECG, and maternal ECG.

Figure 6. Beat by beat validation with pulsed Doppler M-mode images.

Figure 7. From a 38-week pregnancy: the instantaneous fetal heart rate estimated from simultaneous fetal ECG signals (black dotted line) and phonocardiogram signals (red line).

Figure 8. (Upper panel) Bland Altman plot of the relationship of the difference between mean fetal heart rate from ECG (FHRECG) and from phonocardiogram signals (FHRPCG) versus their average values from 270 heart beats collected from 15 fetuses (33–40 weeks). Mean bias (—), +2 SD and −2SD lines are shown. SD = standard deviation. (Lower panel) Correlation between FHRECG and FHRPCG (r = 0.96; p < 0.01).