Skip to content
Licensed Unlicensed Requires Authentication Published by De Gruyter March 10, 2015

Segmented independent component analysis for improved separation of fetal cardiac signals from nonstationary fetal magnetocardiograms

Luiz O. Murta, Mauro G. Guzo, Eder R. Moraes, Oswaldo Baffa, Ronald T. Wakai and Silvia Comani

Abstract

Fetal magnetocardiograms (fMCGs) have been successfully processed with independent component analysis (ICA) to separate the fetal cardiac signals, but ICA effectiveness can be limited by signal nonstationarities due to fetal movements. We propose an ICA-based method to improve the quality of fetal signals separated from fMCG affected by fetal movements. This technique (SegICA) includes a procedure to detect signal nonstationarities, according to which the fMCG recordings are divided in stationary segments that are then processed with ICA. The first and second statistical moments and the signal polarity reversal were used at different threshold levels to detect signal transients. SegICA effectiveness was assessed in two fMCG datasets (with and without fetal movements) by comparing the signal-to-noise ratio (SNR) of the signals extracted with ICA and with SegICA. Results showed that the SNR of fetal signals affected by fetal movements improved with SegICA, whereas the SNR gain was negligible elsewhere. The best measure to detect signal nonstationarities of physiological origin was signal polarity reversal at threshold level 0.9. The first statistical moment also provided good results at threshold level 0.6. SegICA seems a promising method to separate fetal cardiac signals of improved quality from nonstationary fMCG recordings affected by fetal movements.


Corresponding author: Silvia Comani, BIND-Behavioral Imaging and Neural Dynamics Center, University “G. d’Annunzio” of Chieti-Pescara, Via dei Vestini 33, 66100 Chieti, Italy, Phone: +39 0871 3556925, Fax: +39 0871 3556930, E-mail: ; and Department of Neuroscience, Imaging and Clinical Sciences, University “G. d’Annunzio” of Chieti-Pescara, Via dei Vestini 33, 66100 Chieti, Italy

Acknowledgments

This research was supported mainly by the Italian Ministry for Research, INTERLINK Project II04CD8G5A, “New methods to reconstruct and analyze the fetal cardiac signals recorded by magnetocardiography”, and by partial funds from the National Institutes of Health (grant R01 HL063174), CNPq (Brazil, grants 308604/2013-0 and 307941/2010-8), and FAPESP (Brazil, grant 2010/09921-3).

References

[1] Bell AJ, Sejnowski TJ. An information maximization approach to blind separation and blind deconvolution. Neural Comput 1995; 7: 1129–1159.10.1162/neco.1995.7.6.1129Search in Google Scholar

[2] Blaschke T, Wiskott L. CuBICA: independent component analysis by simultaneous third- and fourth-order cumulant diagonalization. IEEE Trans Signal Process 2004; 52: 1250–1256.10.1109/TSP.2004.826173Search in Google Scholar

[3] Burghoff M, Van Leeuwen P. Separation of fetal and maternal magnetocardiographic signals in twin pregnancy using independent component analysis (ICA). Neurol Clin Neurophysiol 2004; 2004: 39.Search in Google Scholar

[4] Cardoso JF, Souloumiac A. Jacobi angles for simultaneous diagonalization. Siam J Matrix Anal Appl 1996; 17: 161–164.10.1137/S0895479893259546Search in Google Scholar

[5] Comani S, Liberati M, Mantini D, et al. Characterization of fetal arrhythmias by means of fetal magnetocardiography in three cases of difficult ultrasonographic imaging. Pacing Clin Electrophysiol 2004; 27: 1647–1655.10.1111/j.1540-8159.2004.00699.xSearch in Google Scholar

[6] Comani S, Mantini D, Alleva G, Di Luzio S, Romani GL. Optimal filter design for shielded and unshielded ambient noise reduction in fetal magnetocardiography. Phys Med Biol 2005; 50: 5509–5521.10.1088/0031-9155/50/23/006Search in Google Scholar

[7] Comani S, Mantini D, Alleva G, Gabriele E, Liberati M, Romani GL. Simultaneous monitoring of separate fetal magnetocardiographic signals in twin pregnancy. Physiol Meas 2005; 26: 193–201.10.1088/0967-3334/26/3/005Search in Google Scholar

[8] Comani S, Mantini D, Lagatta A, Esposito F, Di Luzio S, Romani GL. Time course reconstruction of fetal cardiac signals from fMCG: independent component analysis versus adaptive maternal beat subtraction. Physiol Meas 2004; 25: 1305–1321.10.1088/0967-3334/25/5/019Search in Google Scholar

[9] Comon P. Independent component analysis, a new concept. Signal Process 1994; 36: 287–314.10.1016/0165-1684(94)90029-9Search in Google Scholar

[10] Crato N, Linhares RR, Lopes SRC. Statistical properties of detrended fluctuation analysis. J Stat Comput Simul 2010; 80: 625–641.10.1080/00949650902755152Search in Google Scholar

[11] Cuneo BF, Strasburger JF, Yu S, et al. In utero diagnosis of long QT syndrome by magnetocardiography. Circulation 2013; 128: 2183–2191.10.1161/CIRCULATIONAHA.113.004840Search in Google Scholar

[12] Grimm B, Kaehler C, Schleussner E, Schneider U, Haueisen J, Seewald HJ. Influence of intrauterine growth restriction on cardiac time intervals evaluated by fetal magnetocardiography. Early Hum Dev 2003; 74: 1–11.10.1016/S0378-3782(03)00079-3Search in Google Scholar

[13] Hild KE, Alleva G, Nagarajan S, Comani S. Performance comparison of six independent components analysis algorithms for fetal signal extraction from real fMCG data. Phys Med Biol 2007; 52: 449–462.10.1088/0031-9155/52/2/010Search in Google Scholar

[14] Hild KE, Erdogmus D, Principe JC. Blind source separation of time-varying, instantaneous mixtures using an on-line algorithm. IEEE International Conference On Acoustics, Speech, and Signal Processing 2002; Vol I: 993–996.10.1109/ICASSP.2002.1005909Search in Google Scholar

[15] Horigome H, Shiono J, Shigemitsu S, et al. Detection of cardiac hypertrophy in the fetus by approximation of the current dipole using magnetocardiography. Pediatr Res 2001; 50: 242–245.10.1203/00006450-200108000-00013Search in Google Scholar

[16] Hyvarinen A. Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans Neural Netw 1999; 10: 626–634.10.1109/72.761722Search in Google Scholar

[17] Kahler C, Grimm B, Schleussner E, et al. The application of fetal magnetocardiography to investigate fetal arrhythmias and congenital heart defects. Prenat Diagn 2001; 21: 176–182.10.1002/1097-0223(200103)21:3<176::AID-PD22>3.0.CO;2-WSearch in Google Scholar

[18] Li ZM, Strasburger JF, Cuneo BF, Gotteiner NL, Wakai RT. Giant fetal magnetocardiogram P waves in congenital atrioventricular block – a marker of cardiovascular compensation? Circulation 2004; 110: 2097–2101.10.1161/01.CIR.0000144302.30928.AASearch in Google Scholar

[19] Mantini D, Hild KE, Alleva G, Comani S. Performance comparison of independent component analysis algorithms for fetal cardiac signal reconstruction: a study on synthetic fMCG data. Phys Med Biol 2006; 51: 1033–1046.10.1088/0031-9155/51/4/018Search in Google Scholar

[20] Menendez T, Achenbach S, Beinder E, et al. Usefulness of magnetocardiography for the investigation of fetal arrhythmias. Am J Cardiol 2001; 88: 334–336.10.1016/S0002-9149(01)01658-7Search in Google Scholar

[21] Mensah-Brown NA, Lutter WJ, Comani S, Strasburger JF, Wakai RT. Independent component analysis of normal and abnormal rhythm in twin pregnancies. Physiol Meas 2011; 32: 51–64.10.1088/0967-3334/32/1/004Search in Google Scholar

[22] Moraes ER, Murta LO, Baffa O, Wakai RT, Comani S. Linear and nonlinear measures of fetal heart rate patterns evaluated on very short fetal magnetocardiograms. Physiol Meas 2012; 33: 1563–1583.10.1088/0967-3334/33/10/1563Search in Google Scholar

[23] Schneider U, Fiedler A, Liehr M, Kähler C, Schleussner E. Fetal heart rate variability in growth restricted fetuses. Biomed Tech (Berl) 2006; 51: 248–250.10.1515/BMT.2006.048Search in Google Scholar

[24] Sörnmo L and Laguna P. Bioelectrical signal processing in cardiac and neurological applications. Amsterdam: Academic Press Series in Biomedical Engineering, Elsevier, 2005 (ISBN 0124375529).10.1016/B978-012437552-9/50003-9Search in Google Scholar

[25] Strasburger JF, Wakai RT. Fetal cardiac arrhythmia detection and in utero therapy. Nat Rev Cardiol 2010; 7: 277–290.10.1038/nrcardio.2010.32Search in Google Scholar

[26] Taylor MJ, Smith MJ, Thomas M, et al. Non-invasive fetal electrocardiography in singleton and multiple pregnancies. Br J Obstet Gynecol 2003; 110: 668–678.10.1046/j.1471-0528.2003.02005.xSearch in Google Scholar

[27] Van Leeuwen P, Hailer B, Bader W, Geissler J, Trowitzsch E, Gronemeyer DHW. Magnetocardiography in the diagnosis of fetal arrhythmia. Br J Obstet Gynaecol 1999; 106: 1200–1208.10.1111/j.1471-0528.1999.tb08149.xSearch in Google Scholar

[28] Wakai RT. Assessment of fetal neurodevelopment via fetal magnetocardiography. Exp Neurol 2004; 190: S65–S71.10.1016/j.expneurol.2004.04.019Search in Google Scholar

[29] Wakai RT, Lengle JM, Leuthold AC. Transmission of electric and magnetic foetal cardiac signals in a case of ectopia cordis: the dominant role of the vernix caseosa. Phys Med Biol 2000; 45: 1989–1995.10.1088/0031-9155/45/7/320Search in Google Scholar

[30] Wakai RT, Strasburger JF, Li Z, Deal BJ, Gotteiner NL. Magnetocardiographic rhythm patterns at initiation and termination of fetal supraventricular tachycardia. Circulation 2003; 107: 307–312.10.1161/01.CIR.0000043801.92580.79Search in Google Scholar

[31] Zhao H, Cuneo BF, Strasburger JF, Huhta JC, Gotteiner NL, Wakai RT. Electrophysiological characteristics of fetal atrioventricular block. J Am Coll Cardiol 2008; 51: 77–84.10.1016/j.jacc.2007.06.060Search in Google Scholar

[32] Ziehe A, Muller KR, Nolte G, Mackert BM, Curio G. Artifact reduction in magnetoneurography based on time-delayed second-order correlations. IEEE Trans Biomed Eng 2000; 47: 75–87.10.1109/10.817622Search in Google Scholar

Received: 2014-9-27
Accepted: 2015-1-27
Published Online: 2015-3-10
Published in Print: 2015-6-1

©2015 by De Gruyter