This research focuses on the development of algorithms to extract diagnostic information from the ECG signal, which can be used to improve automatic detection systems and home monitoring solutions. In the first part of this work, a generically applicable algorithm for model selection in kernel principal component analysis is presented, which was inspired by the derivation of respiratory information from the ECG signal. This method not only solves a problem in biomedical signal processing, but more importantly offers a solution to a long-standing problem in the field of machine learning. Next, a methodology to quantify the level of contamination in a segment of ECG is proposed. This level is used to detect artifacts, and to improve the performance of different classifiers, by removing these artifacts from the training set. Furthermore, an evaluation of three different methodologies to compute the ECG-derived ...

Obstructive sleep apnea (OSA) is the most prevalent sleep related breathing disorder, nevertheless subjects suffering from it often remain undiagnosed due to the cumbersome diagnosis procedure. Moreover, the prevalence of OSA is increasing and a better phenotyping of patients is needed in order to prioritize treatment. The goal of this thesis was to tackle those challenges in OSA diagnosis. Additionally, two main algorithmic contributions which are generally applicable were proposed within this thesis. The binary interval coded scoring algorithm was extended to multilevel problems and novel monotonicity constraints were introduced. Moreover, improvements to the random-forest based feature selection were proposed including the use of the Cohen’s kappa value, patient independent validation, and further feature pruning steered by the correlation between features. These novel methods were applied together with classification and feature selection methods from the literature to improve the OSA ...

The present doctoral thesis aims towards the development of new long-term, multi-channel, audio-visual processing techniques for the analysis of bioacoustics phenomena. The effort is focused on the study of the physiology of the gastrointestinal system, aiming at the support of medical research for the discovery of gastrointestinal motility patterns and the diagnosis of functional disorders. The term "processing" in this case is quite broad, incorporating the procedures of signal processing, content description, manipulation and analysis, that are applied to all the recorded bioacoustics signals, the auxiliary audio-visual surveillance information (for the monitoring of experiments and the subjects' status), and the extracted audio-video sequences describing the abdominal sound-field alterations. The thesis outline is as follows. The main objective of the thesis, which is the technological support of medical research, is presented in the first chapter. A quick problem definition is initially ...

The electrocardiogram or ECG is a relatively easy-to-record signal that contains an enormous amount of potentially useful information. It is currently mostly being used for screening purposes. For example, pre-participation cardiovascular screening of young athletes has been endorsed by both scientific organisations and sporting governing bodies. A typical cardiac examination is taken in a hospital environment and lasts 10 seconds. This is often sufficient to detect major pathologies, yet this small sample size of the heart’s functioning can be deceptive when used to evaluate one’s general condition. A solution for this problem is to monitor the patient outside of the hospital, during a longer period of time. Due to the extension of the analysis period, the detection rate of cardiac events can be highly increased, compared to the cardiac exam in the hospital. However, it also increases the likelihood of ...

The electrocardiogram (ECG) signals bear fundamental information for medical experts to make decisions about heart diseases. Therefore, in the past decades the scientific community has made great efforts to develop methods for the heartbeat features extraction via ECG records with the highest accuracy and efficiency using different strategies. It should be noted that noise and artifacts induced by external factors make it difficult to learn specific patterns of ECG signals, which play an important role to find abnormalities. Using filtering techniques such as the unbiased finite impulse response FIR (UFIR) filtering approach promises better results. Aimed at extracting the features with the highest accuracy, in this dissertation, we have designed and applied to ECG signals the adaptive UFIR filter and smoother. We also compared the proposed technique with the traditional method such as UFIR predictors, standard filters (e.g. low-pass filter), ...

Electrocardiography (ECG) is the standard method for assessing the state of the cardiovascular system non-invasively. In the context of magnetic resonance imaging (MRI) the ECG signal is used for cardiac monitoring and triggering, i.e., the acquisition of images synchronized to the cardiac cycle. However, ECG acquisition is impeded by the static and dynamic magnetic fields which alter the measured voltages and may reduce signal-to-noise ratio (SNR), leading to false alarms during cardiac monitoring or to image artifacts during cardiac triggering. A major source of noise is the magnetohydrodynamic (MHD) effect as it is proportional to field strength and represents a key challenge in application of ultra-high-field (UHF) MRI >=7 T. In this work, two approaches for overcoming these limitations are proposed: i) Development of a hardware and software system based on the principal of photoplethysmography imaging (PPGi) as an optical ...

This thesis focuses on the design of various automatic signal processing algorithms to extract information from physiological signals of preterm infants. Overall, the aim was to improve the neurodevelopmental outcome of the neonate. More specifically, three main research objectives were carried out. The first objective was to describe the maturation of neonates during their stay in the neonatal intensive care unit. The second objective was to assess the stress and pain in premature infants and their impact on the development of neonates. The third objective was to predict developmental disabilities, such as autism. The first part of this thesis presents an extensive overview of various developmental models to describe the maturation of premature infants. Three main strategies were proposed. The first strategy proposed an investigation of EEG connectivity networks. A variety of functional and effective connectivity methods were combined with ...

The first part of the presented work is focused on measuring of QT intervals. QT interval can be an indicator of the cardiovascular health of the patient and detect any potential abnormalities. The QT interval is measured from the onset of the QRS complex to the end of the T wave. However, measurements for the end of the T wave are often highly subjective and the corresponding verification is difficult. Here we propose two methods of QT interval measuring – wavelet based and template matching method. Methods are compared with each other and tested on standard QT database. The second part of the presented work is focused on modelling of arrhythmias using McSharry’s model followed with classification using an artificial neural network. The proposed method uses pre-processing of signals with Linear Approximation Distance Thresholding method and Line Segment Clustering method ...

This thesis focuses on wearables for health status monitoring, covering applications aimed at emergency solutions to the COVID-19 pandemic and aging society. The methods of ambient assisted living (AAL) are presented for the neurodegenerative disease Parkinson’s disease (PD), facilitating ’aging in place’ thanks to machine learning and around wearables - solutions of mHealth. Furthermore, the approaches using machine learning and wearables are discussed for early-stage COVID-19 detection, with encouraging accuracy. Firstly, a publicly available dataset containing COVID-19, influenza, and healthy control data was reused for research purposes. The solution presented in this thesis is considering the classification problem and outperformed the state-of-the-art methods, whereas the original paper introduced just anomaly detection and not shown the specificity of the created models. The proposed model in the thesis for early detection of COVID-19 achieved 78 % for the k-NN classifier. Moreover, a ...

Objective diagnosis and therapy evaluation are still challenging tasks for many neurological disorders. This is highly related to the diversity of cases and the variety of treatment modalities available. Especially in the case of epilepsy, which is a complex disorder not well-explained at the biochemical and physiological levels, there is the need for investigations for novel features, which can be extracted and quantified from electrophysiological signals in clinical practice. Neurotherapy is a complementary treatment applied in various disorders of the central nervous system, including epilepsy. The method is subsumed under behavioral medicine and is considered an operant conditioning in psychological terms. Although the application areas of this promising unconventional approach are rapidly increasing, the method is strongly debated, since the neurophysiological underpinnings of the process are not yet well understood. Therefore, verification of the efficacy of the treatment is one ...

Over the last decade there has been an increasing interest in solutions for the continuous monitoring of health status with wireless, and in particular, wearable devices that provide remote analysis of physiological data. The use of wireless technologies have introduced new problems such as the transmission of a huge amount of data within the constraint of limited battery life devices. The design of an accurate and energy efficient telemonitoring system can be achieved by reducing the amount of data that should be transmitted, which is still a challenging task on devices with both computational and energy constraints. Furthermore, it is not sufficient merely to collect and transmit data, and algorithms that provide real-time analysis are needed. In this thesis, we address the problems of compression and analysis of physiological data using the emerging frameworks of Compressive Sensing (CS) and sparse ...

The rate at which our heart beats, is a dynamical process enabling adaptive changes according to the demands of our body. These variations in heart rate are widely studied in so-called heart rate variability (HRV) analyses, as they contain much information about the activity of our autonomic nervous system. Variability in the heart rate arises from several processes, such as thermoregulation, hormones, arterial blood pressure, respiration, etc. One of the main short-term modulators of the heart rate is respiration. This phenomenon is called respiratory sinus arrhythmia (RSA) and comprises the rhythmic fluctuation of the heart rate at respiratory frequency. It has also widely been used as an index of vagal outflow. However, this has been widely debated as some studies have shown that the magnitude of RSA changes with respiratory rate and the depth of breathing, independently of parasympathetic activity. ...

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia encountered in clinical practice, and one of the main causes of ictus and strokes. Despite the advances in the comprehension of its mechanisms, its thorough characterization and the quantification of its effects on the human heart are still an open issue. In particular, the choice of the most appropriate therapy is frequently a hard task. Radiofrequency catheter ablation (CA) is becoming one of the most popular solutions for the treatment of the disease. Yet, very little is known about its impact on heart substrate during AF, thus leading to an inaccurate selection of positive responders to therapy and a low success rate; hence, the need for advanced signal processing tools able to quantify AF impact on heart substrate and assess the effectiveness of the CA therapy in an objective and ...

Advances in wearable electrocardiogram (ECG) monitoring devices have allowed for new cardiovascular applications to emerge beyond diagnostics, such as stress and fatigue detection, athletic performance assessment, sleep disorder characterization, mood recognition, activity surveillance, biometrics, and fitness tracking, to name a few. Such devices, however, are prone to artifacts, particularly due to movement, thus hampering heart rate and heart rate variability measurement and posing a serious threat to cardiac monitoring applications. To address these issues, this thesis proposes the use of a spectro-temporal signal representation called “modulation spectrum”, which is shown to accurately separate cardiac and noise components from the ECG signals, thus opening doors for noise-robust ECG signal processing tools and applications. First, an innovative ECG quality index based on the modulation spectral signal representation is proposed. The representation quantifies the rate-of-change of ECG spectral components, which are shown to ...

Chapter 2 contains a short introduction to the fundamentals of compressed sensing theory, which is the larger context of this thesis. We start with introducing the key concepts of sparsity and sparse representations of signals. We discuss the central problem of compressed sensing, i.e. how to adequately recover sparse signals from a small number of measurements, as well as the multiple formulations of the reconstruction problem. A large part of the chapter is devoted to some of the most important conditions necessary and/or sufficient to guarantee accurate recovery. The aim is to introduce the reader to the basic results, without the burden of detailed proofs. In addition, we also present a few of the popular reconstruction and optimization algorithms that we use throughout the thesis. Chapter 3 presents an alternative sparsity model known as analysis sparsity, that offers similar recovery ...