| Zugriffsnummer | 18138 |
| Dokumenttyp | Zeitschriftenartikel |
| Sprache | Englisch |
| Titel | Separation of fetal and maternal magnetocardiographic signals in twin pregnancy using Independent Component Analysis (ICA) |
| Autor(in); Institution |
Burghoff, Martin; 8.2, Biosignale, PTB-Berlin
Van Leeuwen, P.; Department of Biomagnetism, Research and Development Center for Microtherapy (EFMT), Bochum, GERMANY
|
| Quelle/Jahr | Neurology and Clinical Neurophysiology:(2004), 2004:39, 1 - 4 |
| ISSN | 1526-8748 ; 1081-1818 |
| URL | |
| Verlag | Cambridge, Mass.: MIT Press |
| Freie Schlagworte | Magnetocardiography ; Fetal magnetocardiography ; Independent Component Analysis |
| Zusammenfassung | The identification of fetal and maternal signals in magnetocardiograms (MCG) is central to data preprocessing and a prerequisite for data analysis and assessment. This is usually done by creating a template of the signal to be identified and marking data segments correlating to this template before averaging. This procedure is not only cumbersome, but may also lead to problems when there are several overlapping signals of interest such as in MCG recording in single or, more so, in twin pregnancy. Independent Component Analysis (ICA), which uses higher order statistics to decompose the signal into statistical independent components, has already been used in single pregnancies to distinguish between maternal and fetal signals. We applied the ICA algorithm TDSEP to 9 data sets of twin pregnancies acquired between the 28th and 38th week of pregnancy. Resulting ICA components can be used for further data analysis, e.g., for finding robust triggers or estimating the heart rate and its variability of the twins. The results showed that the maternal and fetal components can be separated from each other as well as from other sources of noise and artifacts. Differences between averaged ICA time curves and averaged raw data are not significant. Limitations include a concurrence of heart rates and changes in signal morphology due to gross movement. Nonetheless, ICA offers a fast and efficient approach for the preprocessing of MCGs with multiple signals of interest. |
