Novel Bayesian vectorcardiographic loop alignment for improved monitoring of ECG and fetal movement

Research output: Contribution to journalArticleAcademicpeer-review

21 Citations (Scopus)
2 Downloads (Pure)

Abstract

The continuous analysis of electrocardiographic (ECG) signals is complicated by morphological variability in the ECG due to movement of the heart. By aligning vectorcardiographic loops, movement-induced ECG variations can be partly corrected for. Existing methods for loop alignment can account for loop rotation, scaling, and time delays, but they lack the possibility to include a priori information on any of these transformations, and they are unreliable in case of low-quality signals, such as fetal ECG signals. The inclusion of a priori information might aid in the robustness of loop alignment and is, hence, proposed in this paper. We provide a generic Bayesian framework to derive our loop alignment method. In this framework, existing methods can be readily derived as well, as a simplification of our method. The loop alignment is evaluated by comparing its performance in loop alignment to two existing methods, for both adult and fetal ECG recordings. For the adult ECG recordings, a quantitative performance assessment shows that the developed method outperforms the existing method in terms of robustness. For the fetal ECG recordings, it is demonstrated that the developed method can be used to correct ECG signals for movement-induced morphology changes (enabling diagnostics) and that the method is capable of classifying recorded ECG signals to periods of fetal movement or rest (p
Original languageEnglish
Pages (from-to)1580-1588
Number of pages9
JournalIEEE Transactions on Biomedical Engineering
Volume60
Issue number6
DOIs
Publication statusPublished - 2013

Fingerprint Dive into the research topics of 'Novel Bayesian vectorcardiographic loop alignment for improved monitoring of ECG and fetal movement'. Together they form a unique fingerprint.

Cite this