Fetal magnetocardiography (fMCG) has been shown to augment fetal ultrasound evaluation for high-risk conditions, but the clinical utility of fMCG depends on the reliability of the cardiac traces reconstructed. We performed a methodological study to examine the influence of gestational age on the properties of the fetal magnetocardiograms extracted with two methods of signal reconstruction: the template matching technique (TMT), which extracts the maternal components from the signal using only temporal information, and independent component analysis (ICA), which separates the fetal signals by using information on the spatial distribution of the mixed source signals in addition to higher order temporal statistics. Efficiency and accuracy were evaluated in terms of fetal beat detection, signal characteristics and duration of cardiac time intervals (CTI) on the averaged traces. ICA outperformed TMT with respect to beat detection and signal-to-noise ratio. The timing of the heartbeats and the duration of the CTI were essentially the same, whereas some alterations in signal morphology were observed in the ICA traces. We conclude that ICA may be useful in early gestation when the signals are noisy, while TMT should be preferred when accurate beat morphology is required for diagnostic purposes.
Influence of gestational age on the effectiveness of spatial and temporal methods for the reconstruction of the fetal magnetocardiogram
COMANI, Silvia;
2009-01-01
Abstract
Fetal magnetocardiography (fMCG) has been shown to augment fetal ultrasound evaluation for high-risk conditions, but the clinical utility of fMCG depends on the reliability of the cardiac traces reconstructed. We performed a methodological study to examine the influence of gestational age on the properties of the fetal magnetocardiograms extracted with two methods of signal reconstruction: the template matching technique (TMT), which extracts the maternal components from the signal using only temporal information, and independent component analysis (ICA), which separates the fetal signals by using information on the spatial distribution of the mixed source signals in addition to higher order temporal statistics. Efficiency and accuracy were evaluated in terms of fetal beat detection, signal characteristics and duration of cardiac time intervals (CTI) on the averaged traces. ICA outperformed TMT with respect to beat detection and signal-to-noise ratio. The timing of the heartbeats and the duration of the CTI were essentially the same, whereas some alterations in signal morphology were observed in the ICA traces. We conclude that ICA may be useful in early gestation when the signals are noisy, while TMT should be preferred when accurate beat morphology is required for diagnostic purposes.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
