The fetal electrocardiogram contains within it, information regarding the health of the fetus. Currently, fetal ECG is recorded directly from the scalp of the baby during labour. However, it has been shown that fetal ECG can also be measured using surface electrodes attached to a pregnant mother's abdomen. The advantage of this method lies in the fact that fetal ECG can be measured noninvasively before the onset of labour. The difficulty lies in isolating the fetal ECG from extraneous signals that are simultaneously recorded with it. Several signal processing methodologies have been put forth in order to extract the fetal ECG component from a mixture of signals. Two recent techniques that have been put forth include a scheme that has previously been used to nonlinearly reduce noise in deterministically chaotic noise and the other uses a blind source separation technique called independent component analysis. In this thesis, we describe the significance of the fetal electrocardiogram as a diagnostic tool in medicine, a brief overview of the theory behind the nonlinear noise reduction technique and blind source separation, and results from having processed synthetic and real data using both techniques. We find that although the noise reduction technique performs adequately, the blind source separation process performs faster and more robustly against similar data. The two techniques can be used in tandem to arrive at an approximate fetal ECG signal, which can be further analyzed by calculating, for example, the fetal heart rate. / Thesis / Master of Engineering (ME)

This thesis is focused on the analysis of fetal ECG records measured by indirect method from mother´s abdomen. The thesis consists of the theoretical part is focused on fetal, heart development and description of fetal ECG signal. This thesis also offers an overview of fECG signal processing methods used nowadays. The practical part of the thesis deals with the implementation of algorithms based on wavelet transformation and Least Mean Square LMS method in Matlab programming environment. The final part of the thesis consists of the analysis of achieved results.

Les malformations cardiaques congénitales sont la première cause de décès liés à une anomalie congénitale. L’´electrocardiogramme du fœtus (ECGf), qui est censé contenir beaucoup plus d’informations par rapport aux méthodes échographiques conventionnelles, peut ˆêtre mesuré´e par des électrodes sur l’abdomen de la mère. Cependant, il est tr`es faible et mélangé avec plusieurs sources de bruit et interférence y compris l’ECG de la mère (ECGm) dont le niveau est très fort. Dans les études précédentes, plusieurs méthodes ont été proposées pour l’extraction de l’ECGf à partir des signaux enregistrés par des électrodes placées à la surface du corps de la mère. Cependant, ces méthodes nécessitent un nombre de capteurs important, et s’avèrent inefficaces avec un ou deux capteurs. Dans cette étude trois approches innovantes reposant sur une paramétrisation algébrique, statistique ou par variables d’état sont proposées. Ces trois méthodes mettent en œuvre des modélisations différentes de la quasi-périodicité du signal cardiaque. Dans la première approche, le signal cardiaque et sa variabilité sont modélisés par un filtre de Kalman. Dans la seconde approche, le signal est découpé en fenêtres selon les battements, et l’empilage constitue un tenseur dont on cherchera la décomposition. Dans la troisième approche, le signal n’est pas modélisé directement, mais il est considéré comme un processus Gaussien, caractérisé par ses statistiques à l’ordre deux. Dans les différentes modèles, contrairement aux études précédentes, l’ECGm et le (ou les) ECGf sont modélisés explicitement. Les performances des méthodes proposées, qui utilisent un nombre minimum de capteurs, sont évaluées sur des données synthétiques et des enregistrements réels, y compris les signaux cardiaques des fœtus jumeaux. / Congenital heart defects are the leading cause of birth defect-related deaths. The fetal electrocardiogram (fECG), which is believed to contain much more information as compared with conventional sonographic methods, can be measured by placing electrodes on the mother’s abdomen. However, it has very low power and is mixed with several sources of noise and interference, including the strong maternal ECG (mECG). In previous studies, several methods have been proposed for the extraction of fECG signals recorded from the maternal body surface. However, these methods require a large number of sensors, and are ineffective with only one or two sensors. In this study, state modeling, statistical and deterministic approaches are proposed for capturing weak traces of fetal cardiac signals. These three methods implement different models of the quasi-periodicity of the cardiac signal. In the first approach, the heart rate and its variability are modeled by a Kalman filter. In the second approach, the signal is divided into windows according to the beats. Stacking the windows constructs a tensor that is then decomposed. In a third approach, the signal is not directly modeled, but it is considered as a Gaussian process characterized by its second order statistics. In all the different proposed methods, unlike previous studies, mECG and fECG(s) are explicitly modeled. The performances of the proposed methods, which utilize a minimal number of electrodes, are assessed on synthetic data and actual recordings including twin fetal cardiac signals.

ECG is used clinically to detect a multitude of medical conditions, such as heart-problems like arrhythmias and heart failure, and to give a good general image of the function of the heart with a quick and harmless exam. In many clinical cases, normal ECG measurements cannot be taken, such as with fetuses where ECG signals from the mother’s own body hinder the measurement. This paper examines using machine learning algorithms to be able to simulate ECG graphs from ultrasound data alone. These algorithms are trained on ultrasound and ECG data acquired from the same patient simultaneously. The data used in the training of the algorithms is taken from samples acquired from 100 adult patients. The results found using this method to simulate an ECG indicate good possibilities for future usefulness, where machine learning to acquire simulated ECG can help facilitate clinicians in evaluating fetal heart function, as well as in other cases where ECG cannot be measured normally. / EKG används kliniskt för att upptäcka en mängd olika åkommor, så som hjärtsvikt och arytmier, men också för att ge en generell bild av hjärtfunktionen med en snabb och harmlös undersökning. I många kliniska fall kan dock inte normal EKG mätning ske, så som för foster då EKG signaler från moderns egna kropp hindrar EKG-mätningen. I detta papper undersöks användandet av maskininlärningsalgoritmer för att kunna simulera EKG grafer från enbart ultraljuds data. Dessa algoritmer är tränade på ultraljud och EKG data som simultant fåtts från samma undersökning av en patient. I detta papper har ultraljudsdatan som använts kommit från 100 mätningar från olika vuxna patienter. Resultaten funna från undersökningen av EKG simulerings metoden indikerar goda möjligheter för framtida användbarhet, då maskininlärningsalgoritmer för att simulera EKG kan underlätta när kliniker ska utvärdera hjärtfunktionen hos foster, eller i andra fall då EKG inte kan mätas normalt.

The non-invasive fetal ECG (NIFECG), derived from abdominal surface electrodes, offers novel diagnostic possibilities for prenatal medicine. Despite its straightforward applicability, NIFECG signals are usually corrupted by many interfering sources. Most significantly, by the maternal ECG (MECG), whose amplitude usually exceeds that of the fetal ECG (FECG) by multiple times. The presence of additional noise sources (e.g. muscular/uterine noise, electrode motion, etc.) further affects the signal-to-noise ratio (SNR) of the FECG. These interfering sources, which typically show a strong non-stationary behavior, render the FECG extraction and fetal QRS (FQRS) detection demanding signal processing tasks. In this thesis, several of the challenges regarding NIFECG signal analysis were addressed. In order to improve NIFECG extraction, the dynamic model of a Kalman filter approach was extended, thus, providing a more adequate representation of the mixture of FECG, MECG, and noise. In addition, aiming at the FECG signal quality assessment, novel metrics were proposed and evaluated. Further, these quality metrics were applied in improving FQRS detection and fetal heart rate estimation based on an innovative evolutionary algorithm and Kalman filtering signal fusion, respectively. The elaborated methods were characterized in depth using both simulated and clinical data, produced throughout this thesis. To stress-test extraction algorithms under ideal circumstances, a comprehensive benchmark protocol was created and contributed to an extensively improved NIFECG simulation toolbox. The developed toolbox and a large simulated dataset were released under an open-source license, allowing researchers to compare results in a reproducible manner. Furthermore, to validate the developed approaches under more realistic and challenging situations, a clinical trial was performed in collaboration with the University Hospital of Leipzig. Aside from serving as a test set for the developed algorithms, the clinical trial enabled an exploratory research. This enables a better understanding about the pathophysiological variables and measurement setup configurations that lead to changes in the abdominal signal's SNR. With such broad scope, this dissertation addresses many of the current aspects of NIFECG analysis and provides future suggestions to establish NIFECG in clinical settings.

The non-invasive fetal ECG (NIFECG), derived from abdominal surface electrodes, offers novel diagnostic possibilities for prenatal medicine. Despite its straightforward applicability, NIFECG signals are usually corrupted by many interfering sources. Most significantly, by the maternal ECG (MECG), whose amplitude usually exceeds that of the fetal ECG (FECG) by multiple times. The presence of additional noise sources (e.g. muscular/uterine noise, electrode motion, etc.) further affects the signal-to-noise ratio (SNR) of the FECG. These interfering sources, which typically show a strong non-stationary behavior, render the FECG extraction and fetal QRS (FQRS) detection demanding signal processing tasks. In this thesis, several of the challenges regarding NIFECG signal analysis were addressed. In order to improve NIFECG extraction, the dynamic model of a Kalman filter approach was extended, thus, providing a more adequate representation of the mixture of FECG, MECG, and noise. In addition, aiming at the FECG signal quality assessment, novel metrics were proposed and evaluated. Further, these quality metrics were applied in improving FQRS detection and fetal heart rate estimation based on an innovative evolutionary algorithm and Kalman filtering signal fusion, respectively. The elaborated methods were characterized in depth using both simulated and clinical data, produced throughout this thesis. To stress-test extraction algorithms under ideal circumstances, a comprehensive benchmark protocol was created and contributed to an extensively improved NIFECG simulation toolbox. The developed toolbox and a large simulated dataset were released under an open-source license, allowing researchers to compare results in a reproducible manner. Furthermore, to validate the developed approaches under more realistic and challenging situations, a clinical trial was performed in collaboration with the University Hospital of Leipzig. Aside from serving as a test set for the developed algorithms, the clinical trial enabled an exploratory research. This enables a better understanding about the pathophysiological variables and measurement setup configurations that lead to changes in the abdominal signal's SNR. With such broad scope, this dissertation addresses many of the current aspects of NIFECG analysis and provides future suggestions to establish NIFECG in clinical settings.:Abstract Acknowledgment Contents List of Figures List of Tables List of Abbreviations List of Symbols (1)Introduction 1.1)Background and Motivation 1.2)Aim of this Work 1.3)Dissertation Outline 1.4)Collaborators and Conflicts of Interest (2)Clinical Background 2.1)Physiology 2.1.1)Changes in the maternal circulatory system 2.1.2)Intrauterine structures and feto-maternal connection 2.1.3)Fetal growth and presentation 2.1.4)Fetal circulatory system 2.1.5)Fetal autonomic nervous system 2.1.6)Fetal heart activity and underlying factors 2.2)Pathology 2.2.1)Premature rupture of membrane 2.2.2)Intrauterine growth restriction 2.2.3)Fetal anemia 2.3)Interpretation of Fetal Heart Activity 2.3.1)Summary of clinical studies on FHR/FHRV 2.3.2)Summary of studies on heart conduction 2.4)Chapter Summary (3)Technical State of the Art 3.1)Prenatal Diagnostic and Measuring Technique 3.1.1)Fetal heart monitoring 3.1.2)Related metrics 3.2)Non-Invasive Fetal ECG Acquisition 3.2.1)Overview 3.2.2)Commercial equipment 3.2.3)Electrode configurations 3.2.4)Available NIFECG databases 3.2.5)Validity and usability of the non-invasive fetal ECG 3.3)Non-Invasive Fetal ECG Extraction Methods 3.3.1)Overview on the non-invasive fetal ECG extraction methods 3.3.2)Kalman filtering basics 3.3.3)Nonlinear Kalman filtering 3.3.4)Extended Kalman filter for FECG estimation 3.4)Fetal QRS Detection 3.4.1)Merging multichannel fetal QRS detections 3.4.2)Detection performance 3.5)Fetal Heart Rate Estimation 3.5.1)Preprocessing the fetal heart rate 3.5.2)Fetal heart rate statistics 3.6)Fetal ECG Morphological Analysis 3.7)Problem Description 3.8)Chapter Summary (4)Novel Approaches for Fetal ECG Analysis 4.1)Preliminary Considerations 4.2)Fetal ECG Extraction by means of Kalman Filtering 4.2.1)Optimized Gaussian approximation 4.2.2)Time-varying covariance matrices 4.2.3)Extended Kalman filter with unknown inputs 4.2.4)Filter calibration 4.3)Accurate Fetal QRS and Heart Rate Detection 4.3.1)Multichannel evolutionary QRS correction 4.3.2)Multichannel fetal heart rate estimation using Kalman filters 4.4)Chapter Summary (5)Data Material 5.1)Simulated Data 5.1.1)The FECG Synthetic Generator (FECGSYN) 5.1.2)The FECG Synthetic Database (FECGSYNDB) 5.2)Clinical Data 5.2.1)Clinical NIFECG recording 5.2.2)Scope and limitations of this study 5.2.3)Data annotation: signal quality and fetal amplitude 5.2.4)Data annotation: fetal QRS annotation 5.3)Chapter Summary (6)Results for Data Analysis 6.1)Simulated Data 6.1.1)Fetal QRS detection 6.1.2)Morphological analysis 6.2)Own Clinical Data 6.2.1)FQRS correction using the evolutionary algorithm 6.2.2)FHR correction by means of Kalman filtering (7)Discussion and Prospective 7.1)Data Availability 7.1.1)New measurement protocol 7.2)Signal Quality 7.3)Extraction Methods 7.4)FQRS and FHR Correction Algorithms (8)Conclusion References (A)Appendix A - Signal Quality Annotation (B)Appendix B - Fetal QRS Annotation (C)Appendix C - Data Recording GUI