Dissertations / Theses on the topic 'Physiological signal processing'

Orthostatic vasovagal syncope is the sudden loss of consciousness resulting from a temporary impairment of cerebral blood flow, within approximately an hour of standing. Patients who suffer from this problem have "vasovagal syndrome". The purpose of this thesis was to devise a method to detect the syndrome following the assumption of upright position. Data from 106 syncopal patients undergoing head-up tilt table testing (HUT) were acquired, including electrical activity of the heart (electrocardiogram), blood pressure, oxygen saturation, and cerebral perfusion parameters from near-infrared spectroscopy (NIRS). The data set was examined with the aim of generating automatic diagnoses. Comparison of the rate-pressure product (blood pressure multiplied by heart rate) during the time of syncope with a recommended threshold, in addition to comparison with monitoring the fall of systolic blood pressure during prolonged tilt, yielded an 84% accuracy rate for vasovagal syndrome. The thesis reviewed the techniques used on the aforementioned time series by previous researchers, emphasising the concepts underlying "time-frequency analysis", a method for analysing nonstationary signals. Since even healthy patients experience time-varying frequency information in their haemodynamics, a transform known as the Smoothed Pseudo-Wigner Ville Distribution (SPWVD) is well suited to their analysis. This distribution was applied to RR tachograms, plots of heart period against time. After the smoothing parameters of the SPWVD were chosen based on artificial data, the optimised transform was then applied to a second artificial tachogram to calculate the LF/HF (low- to high-frequency) ratio, an indicator of heart rate variability. The computed LF/HF ratio tracked the expected value within an error margin of 3.6%. Finally, by applying the same transform to clinical data, it was proved to offer better resolution than an alternative known as the Lomb periodogram. Classical techniques from the literature predicting vasovagal syncope were found to fail on the current data set: out of 29 tests, only two yielded statistically significant differences between the two patient groups. These were compared with the author's time-frequency analysis of RR tachograms, linear regression of heart rate, and examination of NIRS oscillations and changes on tilt. Of these, the ICFV during time period P3 was found to perform best (negative predictive value: 0.86). A linear classifier was used to combine the best four predictors; it achieved an overall accuracy of 0.88. Following the data-driven approach, an analytical modelling approach was undertaken. In order to define an appropriate model that traded off simplicity with comprehensiveness, the mechanisms of vasovagal syncope were reviewed. A model of orthostasis was developed, validated, and used toward parameter estimation from patient data. Three parameters (baroreceptor operating point, cardiac effectiveness, and baroreflex gain) were gleaned from the supine baseline recording to "normalise" the model for a given patient, before four new parameters (sympathetic and parasympathetic gains at the sino-atrial node, peripheral vasoconstriction gain, and total blood volume) were estimated from the data collected in the upright position. The expectation was that this approach would improve feature extraction (and hence prediction accuracy) as well as the clinical interpretation of the results. However, the modelling approach was found to offer no significant improvement upon the data-driven signal processing results: a linear classifier on the four post-tilt parameters yielded a negative predictive value of just 0.69. This result may have been due to inaccuracies in the time series data owing to instrumentation error. It is also possible that the modelling approach was not able to provide the quality of feature extraction necessary for predicting vasovagal syncope in the elderly. Finally, methods to predict syncope during mid- to late HUT were examined. Using information derived from heart rate and baroreflex sensitivity, a technique was developed to ease patient comfort by terminating the test approximately 2 minutes before syncope was expected to occur.

In today’s high paced, hi-tech and high stress environment, with extended work hours, long to-do lists and neglected personal health, sleep deprivation has become common in modern culture. Coupled with these factors is the inherent repetitious and tedious nature of certain occupations and daily routines, which all add up to an undesirable fluctuation in individuals’ cognitive attention and capacity. Given certain critical professions, a momentary or prolonged lapse in attention level can be catastrophic and sometimes deadly. This research proposes to develop a real-time monitoring system which uses fundamental physiological signals such as the Electrocardiograph (ECG), to analyze and predict the presence or lack of cognitive attention in individuals during task execution. The primary focus of this study is to identify the correlation between fluctuating level of attention and its implications on the physiological parameters of the body. The system is designed using only those physiological signals that can be collected easily with small, wearable, portable and non-invasive monitors and thereby being able to predict well in advance, an individual’s potential loss of attention and ingression of sleepiness. Several advanced signal processing techniques have been implemented and investigated to derive multiple clandestine and informative features. These features are then applied to machine learning algorithms to produce classification models that are capable of differentiating between the cases of a person being attentive and the person not being attentive. Furthermore, Electroencephalograph (EEG) signals are also analyzed and classified for use as a benchmark for comparison with ECG analysis. For the study, ECG signals and EEG signals of volunteer subjects are acquired in a controlled experiment. The experiment is designed to inculcate and sustain cognitive attention for a period of time following which an attempt is made to reduce cognitive attention of volunteer subjects. The data acquired during the experiment is decomposed and analyzed for feature extraction and classification. The presented results show that to a fairly reasonable accuracy it is possible to detect the presence or lack of attention in individuals with just their ECG signal, especially in comparison with analysis done on EEG signals. The continual work of this research includes other physiological signals such as Galvanic Skin Response, Heat Flux, Skin Temperature and video based facial feature analysis.

This thesis investigates the mechanisms underlying drug-induced arrhythmia and pro- poses a new approach for the automated analysis of the electrocardiogram (ECG). The current method of assessing the cardiac safety of new drugs in clinical trials is by the measurement and analysis of the QT interval. However, the sensitivity and specificity of the QT interval has been questioned and alternative biomarkers based on T-wave mor- phology have been proposed in the literature. The mechanisms underlying drug effects on T-wave morphology are not clearly understood. Therefore, a combined approach of for- ward cardiac modelling and inverse ECG analysis is adopted to investigate the effects of sotalol, a compound known to have pro-arrhythmic effects, on ventricular repolarisation. A computational model of sotalol and IKr, an ion channel that plays a critical role in ventricular repolarisation, was developed. This model was incorporated into a model of the human ventricular myocyte, and subsequently arranged in a 1-D fibre model of 200 cells. The model was used to assess the effect of sotalol on IKr, action potential duration and biomarkers of ventricular repolarisation derived from the simulated ECG. In parallel, an automated ECG analysis method based on machine learning, signal processing and time-frequency analysis is developed to identify a number of fiducial points in ECG waveforms so that timing intervals and a smooth T-wave segment can be extracted for morphology analysis. The approach is to train a hidden Markov model (HMM) using a data set of ECG waveforms and the corresponding expert annotations. The signal is first encoded using the undecimated wavelet transform (UWT). The UWT coefficients are used for R-peak detection, signal encoding for the HMM and a wavelet de-noising procedure. Using the Viterbi algorithm, the trained HMM is then applied to a subset of the ECG signal to infer the fiducial points for each heart beat. Furthermore, a method for deriving a confidence measure based on the trained HMM is implemented so that a level of confidence can be associated with the automated annotations. Finally, the T-wave segment is extracted from the de-noised ECG signal for morphology characterisation. This thesis contributes to the literature on automated characterisation of drug ef- fects on ventricular repolarisation in three different ways. Firstly, it investigates the mechanisms underlying the effects of drug inhibition of IKr on ventricular repolarisation as captured by the simulated ECG signal. Secondly, it shows how the combination of UWT encoding and HMM inference can be effectively used to segment 24-hour Holter ECG recordings. Evaluation of the segmentation algorithm on a clinical ECG data set demonstrates the ability of the algorithm to overcome problems associated with existing automated systems, and hence provide a more robust analysis of ECG signals. Finally, the thesis provides insight into the drug effects of sotalol on ventricular repolarisation as captured by biomarkers extracted from the surface ECG.

User-centered computer based learning is an emerging field of interdisciplinary research. Research in diverse areas such as psychology, computer science, neuroscience and signal processing is making contributions the promise to take this field to the next level. Learning systems built using contributions from these fields could be used in actual training and education instead of just laboratory proof-of-concept. One of the important advances in this research is the detection and assessment of the cognitive and emotional state of the learner using such systems. This capability moves development beyond the use of traditional user performance metrics to include system intelligence measures that are based on current neuroscience theories. These advances are of paramount importance in the success and wide spread use of learning systems that are automated and intelligent. Emotion is considered an important aspect of how learning occurs, and yet estimating it and making adaptive adjustments are not part of most learning systems. In this research we focus on one specific aspect of constructing an adaptive and intelligent learning system, that is, estimation of the emotion of the learner as he/she is using the automated training system. The challenge starts with the definition of the emotion and the utility of it in human life. The next challenge is to measure the co-varying factors of the emotions in a non-invasive way, and find consistent features from these measures that are valid across wide population. In this research we use four physiological sensors that are non-invasive, and establish a methodology of utilizing the data from these sensors using different signal processing tools. A validated set of visual stimuli used worldwide in the research of emotion and attention, called International Affective Picture System (IAPS), is used. A dataset is collected from the sensors in an experiment designed to elicit emotions from these validated visual stimuli. We describe a novel wavelet method to calculate hemispheric asymmetry metric using electroencephalography data. This method is tested against typically used power spectral density method. We show overall improvement in accuracy in classifying specific emotions using the novel method. We also show distinctions between different discrete emotions from the autonomic nervous system activity using electrocardiography, electrodermal activity and pupil diameter changes. Findings from different features from these sensors are used to give guidelines to use each of the individual sensors in the adaptive learning environment.
Ph.D.
School of Electrical Engineering and Computer Science
Engineering and Computer Science
Electrical Engineering PhD

Over the last century, physiological signals have been broadly analyzed and processed not only to assess the function of the human physiology, but also to better diagnose illnesses or injuries and provide treatment options for patients. In particular, Electrocardiogram (ECG), blood pressure (BP) and impedance are among the most important biomedical signals processed and analyzed. The majority of studies that utilize these signals attempt to diagnose important irregularities such as arrhythmia or blood loss by processing one of these signals. However, the relationship between them is not yet fully studied using computational methods. Therefore, a system that extract and combine features from all physiological signals representative of states such as arrhythmia and loss of blood volume to predict the presence and the severity of such complications is of paramount importance for care givers. This will not only enhance diagnostic methods, but also enable physicians to make more accurate decisions; thereby the overall quality of care provided to patients will improve significantly. In the first part of the dissertation, analysis and processing of ECG signal to detect the most important waves i.e. P, QRS, and T, are described. A wavelet-based method is implemented to facilitate and enhance the detection process. The method not only provides high detection accuracy, but also efficient in regards to memory and execution time. In addition, the method is robust against noise and baseline drift, as supported by the results. The second part outlines a method that extract features from ECG signal in order to classify and predict the severity of arrhythmia. Arrhythmia can be life-threatening or benign. Several methods exist to detect abnormal heartbeats. However, a clear criterion to identify whether the detected arrhythmia is malignant or benign still an open problem. The method discussed in this dissertation will address a novel solution to this important issue. In the third part, a classification model that predicts the severity of loss of blood volume by incorporating multiple physiological signals is elaborated. The features are extracted in time and frequency domains after transforming the signals with Wavelet Transformation (WT). The results support the desirable reliability and accuracy of the system.

The stapedius muscle, which is located in the middle ear, goes into contraction when the ear is exposed to high sound intensities. This muscle activation is called ‘the acoustic reflex’. Measurement of the acoustic reflex is clinically of importance since it can reveal diagnostic information about the middle ear’s pathologies. Moreover, this muscle-activation alters the acoustic characteristics of the middle ear (i.e. the acoustic impedance and the power reflectance), which in turn, can significantly manipulate one’s perception of sounds. In the present study, these acoustic characteristics are measured in the ear canal by means of absorbance measures using equivalent Thevenin circuit theory. The quantities are then compared to form the shift responses between the baseline (before the activation) and the post-activator effect. This project investigates the shifts in power reflectance and admittance of the middle ear caused by the stapedius-muscle contraction. The wideband characterization (0.1- 8 kHz) of these acoustic reflex-induced shifts is achieved using chirp signals as a probe and through ipsilateral broadband noise activator. The data acquisition and signal processing of the project are carried out using MATLAB software. The hardware consists of National Instruments USB-6212 data acquisition interface and low noise microphone system Etymotic Research ER-10B+. A group of 10 adults including 5 males and 5 females are recruited as the participants for the project. The measurements of the reflectance shifts indicate that the most robust frequency region affected by the acoustic reflex is up to 4 kHz whereas for the admittance shifts, this region is up to 2 kHz. In addition, it is shown that the stapedius-muscle contraction leads to the attenuation of the lowfrequency transmission into the middle ear (less than 1 kHz) consistent with a stiffnesscontrolled system in this range of frequencies. In contrast, the results imply that the activation of the stapedius muscle leads to a slight enhancement of the frequency transmission in the range of 1-4 kHz (corresponding to the speech frequency band). These findings suggest a beneficial role for the stapedius-muscle contraction in the perception of speech during vocalization. Furthermore, the implemented methods in this project  can be useful in better understanding the effect of the stapedius-muscle contraction on the speech perception both in normal hearing and hearing impaired persons.

Abstract The rationale for automatically monitoring anesthetic drug effects on the central nervous system (CNS) is to improve possibilities to gain objective information on a patient's state and to adjust the medication individually. Although monitors have shown their usefulness in practice, there are still a number of unclear issues, especially with respect to the scientific foundations and validity of CNS monitoring techniques, and in monitoring the light hypnotic levels. Current monitors are, for example, often based on heuristics and ad hoc solutions. However, a quantitative index for anesthetic drug effect should have a sound relationship with observations and with the selected control variable. The research objectives are: (1) to explore propofol anesthetic related neurophysiological phenomena that can be applied in the automatic assessment of CNS suppression; (2) to develop a valid control variable for this purpose; (3) by means of digital signal processing and mathematical modeling, to design and to evaluate the performance of an index that correlates with the control variable. This dissertation introduces potentially useful neurophysiological phenomena, such as changes in phase synchronization between different EEG channels due to anesthesia, and painful stimulus evoked responses during the burst suppression. Furthermore, it refines the progression of the time-frequency patterns during the induction of anesthesia and shows their relation to the instant of unresponsiveness. The presented spontaneous and evoked EEG phenomena provide complementary information about the CNS functional suppression. Most significantly, the dissertation proposes a continuous and observation based control variable (r scale) and the means to predict its values by using EEG data. The definition of the scale provides a basis for anticipating the instant of the loss of consciousness. Additionally, the phase synchronization index as an indicator of drug effect is introduced. The approximate entropy descriptor performance is evaluated and optimised with a non-stationary signal recorded during the induction of anesthesia. The results open up opportunities to improve the preciseness, scientific validity and the interpretation of information on the anesthetic effects on CNS, and therefore, to increase the reliability of the anesthesia monitoring. Further work is needed to extend and verify the results in deep anesthesia.

This work encompasses several objectives, but is primarily concerned with an experiment where 33 participants were shown 32 slides in order to create ‗weakly induced emotions‘. Recordings of the participants‘ physiological state were taken as well as a self report of their emotional state. We then used an assortment of classifiers to predict emotional state from the recorded physiological signals, a process known as Physiological Pattern Recognition (PPR). We investigated techniques for recording, processing and extracting features from six different physiological signals: Electrocardiogram (ECG), Blood Volume Pulse (BVP), Galvanic Skin Response (GSR), Electromyography (EMG), for the corrugator muscle, skin temperature for the finger and respiratory rate. Improvements to the state of PPR emotion detection were made by allowing for 9 different weakly induced emotional states to be detected at nearly 65% accuracy. This is an improvement in the number of states readily detectable. The work presents many investigations into numerical feature extraction from physiological signals and has a chapter dedicated to collating and trialing facial electromyography techniques. There is also a hardware device we created to collect participant self reported emotional states which showed several improvements to experimental procedure.

Falling asleep while operating a moving vehicle is a contributing factor to the statistics of road related accidents. It has been estimated that 20% of all accidents where a vehicle has been involved are due to sleepiness behind the wheel. To prevent accidents and to save lives are of uttermost importance. In this thesis, given the world’s largest dataset of driver participants, two methods of evaluating driver sleepiness have been evaluated. The first method was based on the creation of epochs from lane departures and KSS, whilst the second method was based solely on the creation of epochs based on KSS. From the epochs, a number of features were extracted from both physiological signals and the car’s controller area network. The most important features were selected via a feature selection step, using sequential forward floating selection. The selected features were trained and evaluated on linear SVM, Gaussian SVM, KNN, random forest and adaboost. The random forest classifier was chosen in all cases when classifying previously unseen data.The results shows that method 1 was prone to overfit. Method 2 proved to be considerably better, and did not suffer from overfitting. The test results regarding method 2 were as follows; sensitivity = 80.3%, specificity = 96.3% and accuracy = 93.5%.The most prominent features overall were found in the EEG and EOG domain together with the sleep/wake predictor feature. However indications have been made that complexities might contribute to the detection of sleepiness as well, especially the Higuchi’s fractal dimension.

This thesis presents three main contributions in the context of modeling and simulation of physiological systems. The first one is a formalization of the methodology involved in multi-formalism and multi-resolution modeling. The second one is the presentation and improvement of a modeling and simulation framework integrating a range of tools that help the definition, analysis, usage and sharing of complex mathematical models. The third contribution is the application of this modeling framework to improve diagnostic and therapeutic strategies for clinical applications involving the cardiovascular system: hypertension-based heart failure (HF) and coronary artery disease (CAD). A prospective application in cardiac resynchronization therapy (CRT) is also presented, which also includes a model of the therapy. Finally, a final application is presented for the study of the baroreflex responses in the newborn lamb. These case studies include the integration of a pulsatile heart into a global cardiovascular model that captures the short and long term regulation of the cardiovascular system with the representation of heart failure, the analysis of coronary hemodynamics and collateral circulation of patients with triple-vessel disease enduring a coronary artery bypass graft surgery, the construction of a coupled electrical and mechanical cardiac model for the optimization of atrio ventricular and intraventricular delays of a biventricular pacemaker, and a model-based estimation of sympathetic and vagal responses of premature newborn lambs.

Les nouveaux-nés prématurés sont vulnérables à des complications comme l’hyperbilirubinémie néonatale et le sepsis tardif (LOS), posant des défis importants dans les unités de soins intensifs néonatals (USIN). Malgré les avancées en matière de soins, la détection précoce et la gestion efficace de ces affections restent complexes. Cette thèse, basée sur l’étude CARESS-Premi (NCT01611740), vise à développer des techniques avancées de traitement des données et des modèles interprétables d’apprentissage automatique afin d’améliorer la prise de décision en USIN, via des systèmes de surveillance non invasifs, continus et en temps réel. Les principales contributions comprennent : (i) une chaîne optimisée de traitement des signaux pour l’analyse ECG en conditions réelles, adaptée aux USIN; (ii) un modèle mathématique patient-spécifique pour la caractérisation de la dynamique postnatale de la bilirubine, avec des paramètres comme biomarqueurs potentiels pour détecter les comorbidités associées ; (iii) une estimation non invasive de la bilirubine utilisant des modèles d’apprentissage automatique à effets mixtes intégrant l’analyse de la variabilité de la fréquence cardiaque (HRV) et des informations physiologiques ; (iv) des modèles pour la détection précoce du LOS via l’analyse de la HRV ; (v) la conception, le déploiement et l’évaluation préliminaire d’un système d’aide à la décision clinique (CDSS) on-the-edge, intégrant du traitement des signaux en quasi-temps réel et des modèles d’inférence dans un contexte USIN. Ces résultats démontrent le potentiel du traitement avancé des signaux physiologiques combiné à l’apprentissage automatique pour optimiser les soins néonatals
Preterm infants are highly vulnerable to complications such as neonatal hyperbilirubinemia and late-onset sepsis (LOS), which pose significant challenges in Neonatal Intensive Care Units (NICU). Despite advancements in neonatal care, early detection and effective management of these conditions remain difficult. Based on the CARESS-Premi project (NCT01611740), the dissertation aims to develop advanced data processing techniques and interpretable machine learning (ML) models to enhance NICU decision-making and neonatal outcomes, by leveraging non-invasive, continuous and real-time monitoring systems. The main contributions include: (i) an optimized automatic signal processing pipeline for real-life ECG analysis tailored to NICU; (ii) a patient-specific mathematical model for postnatal bilirubin dynamics characterization in preterm infants, with model parameters serving as potential biomarkers for detecting associated comorbidities; (iii) the knowledge-based non-invasive bilirubin estimation using mixed-effects ML integrating heart rate variability (HRV) analysis and physiological insights; (iv) ML models for LOS early detection using HRV analysis, proving timely alerts before clinical suspicion; (v) the design, deployment and preliminary evaluation of an on-the-edge clinical decision support system (CDSS) integrating quasi-real-time signal processing and ML models in a NICU setting. These results demonstrate the potential of combining advanced physiological signal processing with ML to optimize neonatal care

À l’échelle mondiale, les systèmes de santé ont des coûts croissants et le nombre d’hospitalisations augmente. La télémédecine permet de ramener l’hôpital à la maison et offre aux structures de santé de nouvelles possibilités d’améliorer le parcours de soins des patients. La surveillance physiologique est une condition préalable à l’efficacité des systèmes de télémédecine et est assurée par des dispositifs médicaux connectés qui ne sont pas entièrement automatisés. Les patients doivent les utiliser activement au quotidien : ces inconvénients entraînent soit un désengagement du patient, soit du personnel soignant supplémentaire. Les moniteurs passifs et sans contact des signes vitaux, tels que les ballistocardiographes qui mesurent les activités motrices, respiratoires et cardiaques, peuvent résoudre l’inefficacité de la télémédecine. En outre, ils sont plus confortables et plus sûrs pour les patients que les moniteurs traditionnels, ce qui est crucial pour le développement neurologique néonatal ou dans les cas de dégénérescence mentale, bien qu’ils soient actuellement moins précis. Comment améliorer la précision de la surveillance physiologique en ballistocardiographie pour accroître l’efficacité de la télémédecine ? Dans cette thèse, le matériel est fourni par une instrumentation propriétaire basée sur un accéléromètre, un logiciel dédié, un simulateur de battements cardiaques, et des campagnes de mesure pour les bases de données de ballistocardiogrammes bruts. De nouvelles méthodes d’amplification analogique et de filtrage numérique sont étudiées pour améliorer la précision de la ballistocardiographie. La force ballistocardiographique, provenant de la déformation de la crosse aortique lors de la systole ventriculaire et mesurée sur le côté du lit, est en effet modulée par les activités respiratoires et motrices, et est polluée par les artefacts mécaniques de l’environnement. En outre, la ballistocardiographie n’est pas normalisée et les ballistocardiogrammes présentent des variabilités inter et intra-individuelles élevées, en fonction de la literie, de la position au lit, de la morphologie et de la physiologie du patient. L’amplification analogique est étudiée d’un point de vue mécanique et électronique. Premièrement, en ce qui concerne l’amplification mécanique, un nouveau guide d’ondes, prenant la forme d’un ruban de coton qui encercle le matelas, a été inventé pour concentrer l’énergie de la force ballistocardiographique dans une direction, du thorax jusqu’au capteur. Deuxièmement, en ce qui concerne l’amplification électronique, un circuit de conditionnement hybride a été conçu pour optimiser le compromis entre le gain de l’amplificateur électronique et la durée de saturation après un mouvement. Les méthodes de filtrage numérique visent à séparer les sources de signaux, à éliminer les artefacts puis à détecter les signes vitaux. Trois algorithmes originaux ont été conçus pour reconnaître efficacement les battements de cœur dans les ballistocardiogrammes. Le premier algorithme est la comparaison par déformation temporelle dynamique, où un modèle battement cardiaque est utilisé pour reconnaître les battements cardiaques en utilisant une distance non-linéaire. Le second algorithme modélise les ballistocardiogrammes avec des modèles de Markov cachés périodiques. Le troisième algorithme, le réseau neuronal U-Net, est supervisé et segmente les battements cardiaques en ballistocardiogrammes. Finalement, les ballistocardiogrammes sont amplifiés mécaniquement et électroniquement de 12 dB et 21 dB respectivement, sans saturation après mouvement ; et les algorithmes de filtrage numérique atteignent une précision de 97 % et une sensibilité de 96 % pour la détection des battements cardiaques. Prochainement, le ballistocardiographe conçu sera évalué cliniquement dans une unité de soins intensifs pédiatriques et en télémédecine par rapport à d’autres ballistocardiographes et aux méthodes de référence
Globally, healthcare systems have increasing costs and the number of hospitalizations grows. Telehealth brings hospital at home and provides health structures with new opportunities to improve the patient care pathway. Physiological monitoring is a prerequisite in efficient telehealth systems and is performed by connected medical devices that are not fully automated. Patients need to use them actively on a day-to-day basis: these drawbacks lead either to patient disengagement or to additional caregiver support. Passive contactless vital signs’ monitors, such as ballistocardiograms sleep trackers that measure motor, respiratory and cardiac activities, can solve the telehealth inefficiency. Moreover, they are more comfortable and safer for patients than traditional monitors, which is crucial for neonatal neurological development or in case of mental degeneration, though they are currently less accurate. How to improve physiological monitoring accuracy in ballistocardiography to increase telehealth efficiency? In this thesis, materials are provided by a self-designed accelerometer-based instrumentation, a dedicated software, a heartbeat simulator, and measurement campaigns for raw ballistocardiograms’ databases. Novel analog amplification and digital filtering methods are investigated to improve ballistocardiography accuracy. The ballistocardiographic force, coming from the aortic arch deformation during the ventricular systole and measured on the bedside, is indeed modulated by respiratory and motor activities, and is polluted by environment mechanical artifacts. Furthermore, the ballistocardiography is unstandardized and ballistocardiograms have high inter- and intra-variabilities, depending on the beddings, the position in bed, the morphology and the physiology of the patient. Analog amplification is studied from two perspectives: the mechanical amplification of ballistocardiograms from the patient to the sensor, and the electronic amplification of the analog acceleration signal. First, concerning the mechanical amplification, a novel waveguide bedding, a cotton tape encircling the mattress, was invented to concentrate the strain energy of the ballistocardiographic force in one direction, from the thorax straight to the attached sensor. Second, concerning the electronic amplification, a mixed-signal front-end was conceived to optimize the tradeoff between the electronic amplifier gain and the saturation time after a movement. The conditioning circuit measures the unamplified sensor output, passes it through a digital filter with a sharp transition frequency bandwidth and a proper initialization, and analogically amplifies the difference between this unwanted synthesized signal and the unamplified sensor output using a low noise instrumentation amplifier. Digital filtering methods aims at separating signal sources, removing artifacts then detecting vital signs. Three original algorithms have been designed to efficiently recognize heartbeats in ballistocardiograms. The first algorithm is dynamic time warping template matching, where a heartbeat template is used to match heartbeats using a warping distance. The second algorithm models ballistocardiograms with periodic hidden Markov models. The third algorithm, the U-Net neural network, is supervised and segments heartbeats in ballistocardiograms. Finally, ballistocardiograms are mechanically and electronically amplified by 12 dB and 21 dB respectively, without saturation time; and digital filtering algorithms reach a 97% precision and 96% recall for heartbeats detection. Shortly, the designed ballistocardiograph will be clinically evaluated in a pediatric intensive care unit and in telemedicine against other ballistocardiographs and the gold standard methods

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005.
Includes bibliographical references (p. 119-120).
This report studies physiological signals measured from patients in the Intensive Care Unit (ICU). The signals explored include heart rate, arterial blood pressure, pulmonary artery pressure, and central venous pressure measurements. Following an introduction to these signals, several methods are proposed for visualizing the data using time and frequency domain techniques. By way of a patient case study we motivate a novel method for data clustering based on the singular value decomposition and present some potential applications based on this method for use within the ICU setting.
by Carlos A. Renjifo.
M.Eng.

碩士
亞東技術學院
資訊與通訊工程研究所
101
With the aging population structure, cardiovascular disease is now second among the ten major causes of death, based on a national health survey. A good home health care system can substantially reduce the huge social burden and provide the security obtained from preventive care measures. Previous electrocardiogram (ECG) measurement is done in statistic state with the patients lying still on the bed. However, physiological signal itself possesses significant discreteness, thus it may change greatly if the body system of patient is in a certain state. For some heart diseases such as sudden myocardial infarction and arrhythmia, danger can be reduced only if first-aid is given at the crucial moment, and ECG is regarded as an important indicator for the detection of cardiovascular diseases. The signal is hard to measure and easy to be interfered by noise due to its weakness, so during measurement, quality filter should be used to remove the noise outside the frequency range of the ECG signal, and the ECG measurement must be highly accurate and real-time with continuous monitoring. This study proposed a 24-hour wearable real-time monitoring system that can be used continuously for a long time. This ECG system can transfer ECG signals to an Android phone or tablet computer through the Bluetooth transmission interface for real-time processing. Then, ECG, cardiac rhythm, body temperature and GPS position information will be displayed and synchronously uploaded to cloud database or medical care center for real-time monitoring and query of doctors and medical staff. The 24Hrs real-time physiological measuring system in this study can give medical staff the chance to provide first aid at crucial moments and improve the quality of medical care service.

碩士
國立臺灣大學
電機工程學研究所
92
It has been more than 20 years for the research of psychic phenomena in Asia. The abundant data have already proved the existence of the extraordinary human ability and the possibility to train the psychic power for ordinary persons. In this research, the physiological response of one talented subject who has the greatest gift for finger-reading ability had been measured under finger-reading situations. After analyzing the physiological data of EEG, the onset of brain screen, and skin potential, it is concluded that the rise in skin potential occurred approximately 2.0 seconds before the opening of the brain screen. After the appearance of the skin potential, the brain waves become different , when comparing with eye closed and eye opened situations, their magnitude and distribution ofㄈave change as well as the increasing of the coherence and ApEn. Following a latent period, the brain screen emerges in the brain of the talented subject who percepts the information in the paper holding in her hand. In this study, we also use both linear and nonlinear parameters to discuss the unusual mechanism of the finger-reading as well as the extraordinary phenomena far from the material world.

碩士
中原大學
生物醫學工程研究所
102
Biosignal can provide a lot of usefully physiological information to help the medical professionals to diagnose and identify variant diseases. However, deriving characteristics of these diseases from obtained biosignals need to calculate and signal process. Therefore, biosignal processing can be used to derive these characteristics of biosignal not only in time-domain analysis, but also in frequency-domain analysis. Such as HRV power spectrum density can provide important information for regulatory mechanisms of autonomic nervous system, and the EEG spectrum can be used to diagnose brain diseases. This study proposed a SoC system to analyze ECG signals and focus on spectrum analysis technique. These signal processing produces included several steps: the differentiation, moving average, R-wave detection, and resampling. The equidistant RR-Interval time series were obtained from originally obtained ECG signals, subsequently, these derived signals were inputted into FFT algorithms to transform RR-Interval time series from time-domain to frequency-domain. A critical algorithm of 1024 points FFT was developed. The structure of this algorithm was based on the Radix-2 DIF, and each operator contained the buffer, butterfly algorithm, twiddle factor, memory control, and correct output data clock control. The real-time results were transmitted to the computer by using USB interface, and HRV power spectrum density was presented by Borland C++ Builder program on a personal computer simultaneously. Experimental results presented three kinds of methods to valid. (1)Four kinds of sine wave with different frequencies (1Hz, 10Hz, 50Hz and 125Hz) were given into FFT algorithms, respectively. For example, the spectrum of a 1Hz sine wave should theoretically appear the peak at 1Hz. The practical result was found the peak at 1Hz as we supposed. The others sine wave presented the same results. Preliminary evidence of FFT algorithms was correct. (2) Four different heartbeat frequencies were combined for verification of ECG signal processing. The result showed peaks of the synthesized frequency were located in the theoretical range. (3)The validation of ECG signal in this study has been tested with 10 subjects in two different conditions. One was at resting condition and the other one was at mental athletic condition. At the resting condition, a relatively higher power was appeared at high-frequency range, indicating the parasympathetic was activated. At the mental athletic condition, a relatively higher power was showed at low-frequency range, indicating the sympathetic was activated. From observing the derived HRV power spectrum density, a quantified parameter was provided to descript the activation of the autonomic nervous system. In conclusion, this study presented a real-time physiological signal processing algorithm to achieve real-time measurement and analysis, and results presented an excellent accuracy. In the future, this FFT algorithm can be used in other biosignal processing, such like EEG, blood pressure, and EMG.

碩士
國立中央大學
生物醫學工程研究所
98
This thesis presents a physiological signal processing circuit which can be used to measure many kinds of physiological signals. Based on this circuit, a 3-channel human-computer interface (HCI) system incorporated with a decision rule algorithm is implemented to measure vertical and horizontal eye movements, and alpha waves of brain signals. The 3-channel human-computer interface (HCI) system can be used in three different application domains. First of all, the system is utilized to be a computer interface for the disabled persons. The user can use his or her eye movements to control the mouse and then operate a communication aid for communications, typing, web surfing, and controlling home appliances. Secondly, the system incorporated with an algorithm is utilized to be a tool for recording and detecting the Rapid Eye Movement (REM) events during a sleep period. REM events are detected via the features extracted from the Fast Fourier Transform (FFT), turn counts, and zero-crossing rate (ZCR). The system is also used to control a toy helicopter. The moving directions are controlled by the eye movements and the start/stop is controlled by the alpha waves. Several experiments were designed to evaluate the system. The recognition rate for classifying the eye movements was about 85% ratio correct. Experimental results also shows the system can correctly detect the REM events and control a toy helicopter.

博士
國立臺灣大學
電機工程學研究所
90
The major purpose of this dissertation is to investigate the theories of various time-frequency analysis (TFA) and its capabilities in representing noninvasive physiological signals. The applications of TFAs in cutaneous electrogastrography (EGG) measurement and otoacoustic emissions (OAE) are also demonstrated. Owing to the tiny, noisy and nonstationary characteristics of noninvasive physiological signals, conventional time- and frequency- domain based analysis are not adequate to extract all the information embedded within the original signals. TFAs can effectively decompose the original signals into time-frequency distributions (TFDs) that can provide both time and frequency resolutions. More precise medical diagnosis can thus be achieved. Because TFAs can represent signal features more efficiently, higher performance is accomplished in several biomedical applications, such as signal compressions, and pattern recognitions, by TFA-based signal processing methodologies. The mathematical backgrounds of several commonly used linear and quadratic TFAs are described, and their pros and cons of representing nonstationary signals are discussed by apply simulated signals. Fast algorithms of the digital wavelet transform are introduced and proposed as the appropriate basis for real-time TFA-based signal processing, which are successfully implemented in a digital signal processor (DSP). In the research of cutaneous EGG measurement, a microprocessor-based portable multichannel EGG monitoring system is proposed to record long-term EGG signals. A simulated EGG signal is designed and applied by the TFAs, and we concluded that the short-time Fourier transform (STFT) and Choi-Williams distributions are appropriate for EGG analysis. The slow wave can thus be precisely tracked by these TFAs, and quantitative parameters are proposed. Because it may generate errors by traditional power estimation, TFA-based power estimation, called multibands analysis, is developed in this dissertation. Clinical experiments are also deployed to evaluate the proposed EGG measurement system. In the research of OAE measurement, a DSP-based instrument is developed for OAE monitoring. We used a simulated TEOAE signal to testify that the TFAs can efficiently decompose the original signal, and the results of various TFAs are compared and discussed. The specific feature of how different frequency components vary with time, which is similar to the Cochlear organ, can be successfully extracted by the wavelet transform. Because the acquired TEOAE signals are severely contaminated by environmental white noise, we designed a TFA-based active denoising methodology, called wavelet shrinkage, to suppress the embedded white noise during the measurement. The proposed method is more efficient than traditional statistically averaging method and is implemented in the DSP-based system.

The ability to recognise facial expression is crucial to human social interaction and plays an integral part in most social interaction scenarios. Being able to understand genuine facial expressions is considered a valuable social skill. In this context, a smile is treated as a major facial display that can be recognised easily, but can be confusing. Genuine smiles are usually labelled as both showing and feeling happiness, but posed smiles are far less likely to be understood as feeling happiness than as showing happiness. Previous studies considered either observers’ verbal responses or smilers’ facial features in distinguishing between real and posed smiles. Some of these studies also considered smilers’ physiological signals in this context, but none of the studies considered observers’ physiological responses. The latter concept is considered in this thesis. This thesis outlines a review of relevant physiological signals and related facial expression recognition as well as affective computing literature. This thesis also presents a novel computational approach, called the independent approach, to train the classifiers with a totally independent dataset that uses a ‘leave-one-subject-and-one-stimulus-out’ cross validation technique. Before discussing observers’ physiological reactions to the viewed smile faces, it addresses observers’ pupillary changes to viewed graphical visualisations. Thus, it discusses non-invasive and unobtrusive physiological sensors and relevant computational techniques for discriminating between observed smiles (real versus posed) with a preparatory examination of some of these physiological signals in observed visualisations (radial graphs versus hierarchical graphs). At the beginning of this thesis two graphical visualisations (radial and hierarchical) are chosen to distinguish between observers’ verbal responses and their pupillary responses. The graphical visualisations are snapshots of a kind of data used in checking the degree of compliance with corporate governance best practice. The radial visualisation shows the connections between the board members of BHP Titanium Pty Ltd and ICI Australia Petrochemicals. The hierarchical visualisation exhibits the connections between the board members of the National Australia Bank and Sydney 2001 Olympics. Six, very similar, questions were asked from each participant for each visualisation and found that although observers are not able to distinguish between the radial and hierarchical graphs according to their verbal responses, their pupillary responses can. The analysis of the experiment also shows that observers are verbally 81.0% and physiologically 95.8% correct. The outcomes from the above experiment motivated me to design experiments to analyse the changes of physiological responses to = viewed human facial expressions instead of graphical visualisations. In this regard I have chosen smiles as a human facial expression, because it generally means happiness and/or is used to motivate others. For example, a speaker can be motivated from audience smiles. On the other hand, people can smile from feeling or by acting or posing the smile. Thus discriminating real from posed smiles is important in human centred computing, for computers to ‘understand’ smilers’ mental states. Experiments are designed and conducted to acquire observers’ physiological signals with their verbal responses while watching smile videos. A number of smile videos are collected from the literature and processed to show them to participants/ observers. The processed smile videos are classified as real/ posed according to their elicitation provenance. The physiological signals in the data sets include pupillary response, galvanic skin response (GSR), electrocardiogram (ECG) and/ or blood volume pulse (BVP). Observers’ physiological signals are analysed by developing computer programs via signal processing and machine learning techniques. Several methods are used to develop this computational model such as noise removal, feature extraction, feature selection, fusion, classification, ensemble learning, and so on. In this connection, filtering is considered to remove noise from the recorded signals; normalisation is used to overcome age effects on the signals; interpolation is applied to reconstruct the missing values; principal component analysis (PCA) is employed to eliminate the lighting effects from the recorded signals, etc. Neural networks (NN), support vector machines (SVM), relevance vector machines RVM), k-nearest neighbours (KNN), and ensemble classification techniques are employed to develop the classification model in this thesis. Final results show that participants are verbally 52.7% (on average) to 68.4% (by voting) correct whereas they are physiologically 96.1% correct using an independent approach and ensemble technique. Overall, this thesis contributes a significant dimension to developing a computational model to differentiate between real and posed smiles as well as offering insights into observers’ understanding of radial and hierarchical visualisations based on their peripheral physiology. In other words, this thesis measures observers’ feelings and reactions to the observed images and smiling face visualisations. As physiological signals are not easy to control voluntarily and provide spontaneous and non-conscious outcomes, the outcome of this thesis identifies an authentic and important difference between what observers’ say and feel, that is, reported verbally (say) and reflect physiologically (feel), respectively. Further research suggests including sensor technologies the care-givers or users of avatars to understand human psychology via their recorded physiology.

Listening to a song has been reported to evoke the emotion of the persons. In fact, it acts as a stimulus, which can initiate emotional memories in the person. Recent studies have suggested that the alteration in the emotions while listening tosong is also associated with a change in the physical state of the person. Listening to music has been reported to alter the physiological processes like promoting sleep, reduce anxiety, increase positive post-task attitude towards exercise and/or increase brain activities. The electrical activity of the heart can be analysed non-invasively using electrocardiogram (ECG) signals.Classification of ECG signals has been extensively studied by the researchers for the diagnosis of cardiac diseases as well as for identifying any alteration in the cardiac electrophysiology due to a stimulus.ECG signals were acquired from 18 healthy male volunteers during the pre-and the post-stimulus conditions.The RR intervals (RRIs) were extracted.The processing of RRI time series was performed to understand the autonomic nervous system (ANS) physiology in the volunteers.Recurrence analysis of both the ECG and the RRI signals indicated a higher alteration (acceleration or deceleration) in the heart rate along with the reduction of the causality and patterned behavior of the RRIs when the stimulus was applied.The ECG and the RRI signals were also processed using wavelet packet decomposition (WPD)methods.WPD was conducted on Daubechies wavelet (db04). Statistically important parameters were identified from the extracted parameters using t-test, Classification and Regression Tree (CART), Boosted Tree (BT) and Random Forest (RF) methods. Radial basis function (RBF) and multilayer perceptron (MLP) Artificial Neural Networks(ANNs) were implemented for the classification of the ECG and the RRI signals. In this study, the RBF network proved to be a better classifier than the MLP network, and it resulted in a classification efficiency of ≥80%, suggesting an alteration in the cardiac electrophysiology of the volunteers caused by the stimulus.

The prime objective of this piece of work is to devise novel techniques for computer based classification of Electrocardiogram (ECG) arrhythmias with a focus on less computational time and better accuracy. As an initial stride in this direction, ECG beat classification is achieved by using feature extracting techniques to make a neural network (NN) system more effective. The feature extraction technique used is Wavelet Signal Processing. Coefficients from the discrete wavelet transform were used to represent the ECG diagnostic information and features were extracted using the coefficients and were normalised. These feature sets were then used in the classifier i.e. a simple feed forward back propagation neural network (FFBNN). This paper presents a detail study of the classification accuracy of ECG signal by using these four structures for computationally efficient early diagnosis. Neural network used in this study is a well-known neural network architecture named as multi-Layered perceptron (MLP) with back propagation training algorithm. The ECG signals have been taken from MIT-BIH ECG database, and are used in training to classify 3 different Arrhythmias out of ten arrhythmias. These are normal sinus rhythm, paced beat, left bundle branch block. Before testing, the proposed structures are trained by back propagation algorithm. The results show that the wavelet decomposition method is very effective and efficient for fast computation of ECG signal analysis in conjunction with the classifier.

abstract: Monitoring vital physiological signals, such as heart rate, blood pressure and breathing pattern, are basic requirements in the diagnosis and management of various diseases. Traditionally, these signals are measured only in hospital and clinical settings. An important recent trend is the development of portable devices for tracking these physiological signals non-invasively by using optical methods. These portable devices, when combined with cell phones, tablets or other mobile devices, provide a new opportunity for everyone to monitor one’s vital signs out of clinic. This thesis work develops camera-based systems and algorithms to monitor several physiological waveforms and parameters, without having to bring the sensors in contact with a subject. Based on skin color change, photoplethysmogram (PPG) waveform is recorded, from which heart rate and pulse transit time are obtained. Using a dual-wavelength illumination and triggered camera control system, blood oxygen saturation level is captured. By monitoring shoulder movement using differential imaging processing method, respiratory information is acquired, including breathing rate and breathing volume. Ballistocardiogram (BCG) is obtained based on facial feature detection and motion tracking. Blood pressure is further calculated from simultaneously recorded PPG and BCG, based on the time difference between these two waveforms. The developed methods have been validated by comparisons against reference devices and through pilot studies. All of the aforementioned measurements are conducted without any physical contact between sensors and subjects. The work presented herein provides alternative solutions to track one’s health and wellness under normal living condition.
Dissertation/Thesis
Doctoral Dissertation Electrical Engineering 2016

Physiological signals, electroencephalogram (EEG) and heart rate variability (HRV), are generated by complex self-regulating systems. These signals are extremely inhomogeneous and nonstationary, and fluctuate in an irregular and highly complex manner. These fluctuations are due to underlying dynamics of the system. The synchronous neural activity measured as scalp EEG indicates underlying neural dynamics of the brain. Hence, quantitative EEG analysis has become a very useful tool in interpreting results from physiological experiments. The analysis of HRV provides valuable information to assess the autonomous nervous system (ANS). The HRV can be significantly affected by physiological state changes and many disease states. Hence, HRV analysis is becoming a major experimental and diagnostic tool. In this thesis, we focus on the study of EEG and HRV time series using tools from nonlinear time series analysis with special emphasis on its implications in detecting physiological state changes such as, in diseases like epileptic seizure and schizophrenia, and in altered states of consciousness as in sleep and meditation. The proposed nonlinear techniques are used in discriminating different physiological states from control states. Artifact processing of EEG signal Interferences (artifacts) from various sources unavoidably contaminate EEG recordings. In quantitative analysis, results can differ significantly by these artifacts, which may lead to wrong interpretation of the results. In this part of the thesis, we have devised methods to minimize ocular and muscle artifacts in EEG. The artifact correction methods are based on blind source separation (BSS) techniques such as singular value decomposition (SVD), algorithm for multiple signal extraction (AMUSE), canonical correlation analysis (CCA), information maximization (INFOMAX) independent component analysis (ICA) and joint approximate diagonalization of eigen-matrices (JADE) ICA. We have proposed a method to simulate clean and artifact corrupted EEG data based on the BSS methods. In order to enhance the performance of BSS methods, a technique called wavelet-filtered component inclusion method has been introduced. In addition, second-order statistics (SOS) and higher-order statistics (HOS) based BSS methods have been studied considering less number of EEG channels; and performance comparison of these methods has also been made. We have also addressed the problem of simultaneous correction of ocular and muscle artifacts in EEG recordings using the BSS methods. Irrespective of the BSS methods, the component elimination method has introduced high spectral error in all the bands after reconstruction of clean EEG. However, the wavelet filtered component inclusion method has retained almost all spectral powers of EEG channels in theta, alpha, and beta bands after ocular artifact minimization. When the number of EEG channels is very less, the enhanced CCA (SOS BSS) has given superior artifact minimization results than HOS BSS methods, especially in delta band. The component elimination method is used in muscle artifact minimization, and hence the SVD method cannot be used for this purpose since it leads to large spectral distortion of reconstructed EEG. The AMUSE and CCA methods have given comparable performance in muscle artifact minimization. In addition, the JADE method has introduced less mean spectral error compared to other methods. The CCA method has shown superior performance in simultaneous minimization of ocular and muscle artifacts, and AMUSE and JADE methods have given comparable results. Furthermore, the less computation time of wavelet enhanced SOS BSS methods make them very useful in real clinical environments. Fractal characterization of time series In biomedical signal analysis, fractal dimension (FD) is used as a quantitative measure to estimate complexity of physiological signals. Such analysis helps to study physiological processes of underlying systems. The FD can also be used to study dynamics of transitions between different states of systems like brain and ANS, in various physiological and pathological states. In this part, we have proposed a method to estimate FD of time series, called multiresolution box-counting (MRBC) method. A modification of this method resulted in multiresolution length (MRL) method. The estimation performance of the proposed methods is compared with that of Katz, Sevcik, and Higuchi methods, by simulating mathematically defined fractal signals, and also the computation time is compared between the methods. The MRBC and MRL methods have given comparable performance to that of Higuchi method, in estimating FD of waveforms, with the advantage of less computational time. In addition, various properties of the FD are studied and discussed in connection with classical signal processing concepts such as amplitude, frequency, sampling frequency, effect of noise, band width, correlation, etc. The FD value of signals has increased with number of harmonics, noise variance, band-width, and mid-band frequency, and decreased with degree of correlation in AR signal. An analogy between Katz FD and smoothed Teager energy operator has also been made. Application of fractal analysis to EEG and HRV time series The fluctuation of EEG potentials normally depends upon degree of alertness, and varies in amplitude and frequency. Hence, the EEG is an important clinical tool for studying sleep and sleep related disorders, epileptic seizures, schizophrenia, and meditation. In this part of the thesis, we have used FD which gives signal complexity, and detrended fluctuation analysis (DFA) which gives multiscale exponent of time series to quantify EEG. We have extended the concept of FD to multiscale FD to compute complexity of time series at multiple scales. The main applications of the proposed method are epileptic seizure detection, sleep stage detection, schizophrenia EEG analysis, and analysis of heart rate variability during meditation. For seizure detection, we have used intracranial EEG recordings with seizure-free and seizure intervals. In sleep EEG analysis, whole-night sleep EEG is used and results are compared with the manually scored hypnogram. The schizophrenia symptom is further categorized into positive and negative symptoms and complexity is estimated using FD and DFA. We have also analyzed HRV data of Chi and Kundalini meditation using FD and DFA techniques. In all the applications considered, we have tested for statistical significance of the computed parameters, between the case of interest and corresponding control cases, to discriminate between the physiological states. The ocular artifact has reduced FD while muscle artifact increased FD of EEG. The FD of seizure EEG has shown high value compared to that of seizure-free EEG. In addition, the seizure-free EEG has more DFA exponent-1 than seizure EEG. The value of FD of EEG is decreased with deepening of sleep, wake state having high FD value. The FD of REM state sleep EEG showed value between that of wake and state-1. The DFA exponent-1 has increased with deepening of sleep state, having small value for wake state. The REM state has given exponent-1 value between wake and state-1. The schizophrenia subjects have shown lower FD value than healthy controls in all the EEG channels except the bilateral temporal and occipital regions. The positive symptom sub-group has shown comparatively high FD values than healthy controls as well as overall schizophrenia sample in the bilateral tempero-parietal-occipital region. In addition, the positive symptom sub-group has shown significantly higher regional FD values than negative symptom sub-group especially in right temporal region. The overall schizophrenia samples as well as the positive and negative subgroup have shown least FD values in the bilateral frontal region. The values of DFA exponent-2 have shown significant high value in schizophrenia samples. In addition, the schizophrenia group has shown less DFA exponent-1 in bilateral temporal region than healthy control. The FD, multiscale FD, DFA exponents have shown significant performance in discriminating different physiological states from control states. The FD value of HRV time series during meditation is less compared to pre-meditation state in both Chi and Kundalini meditation. Irrespective of the type of meditation, meditation state has shown significantly high DFA exponent-1 than pre-meditation state, and significantly high DFA exponent-2 in pre-meditation state compared to meditation state. Functional connectivity analysis of brain during meditation In functionally related regions of the brain, even in those regions separated by substantial distances, the EEG fluctuations are synchronous, which is termed as functional connectivity. In this part, a novel application of functional connectivity analysis of brain using graph theoretic approach has been made on the EEG recorded from meditation practitioners. We have used 16 channel EEG data from subjects while performing Raja Yoga meditation. The pre-meditation condition is used as control state, against which meditation state is compared. For finding connectivity between EEG of various channels, we have computed pair-wise linear correlation and mutual information between the EEG channels, to form a connection matrix of size 16x16. Then, various graph parameters, such as average connection density, degree of nodes, characteristic path length, and cluster index, are computed from the connection matrix. The computed parameters are projected on to the scalp to get topographic head maps that give spatial variation of the parameter, and results are compared between meditation and pre-meditation states. The meditation state has shown low average connection density, less characteristic path length, and high average degree in fronto-central and central regions. Furthermore, high cluster index is shown in frontal and central regions than pre-meditation state. The parameters such as complexity, characteristic path length and average connection density are used as features in quadratic discriminant classifier to classify meditation and pre-meditation state, and have given good accuracy performance. Connectivity analysis using mutual information has given high average connection density in meditation state in theta, alpha and beta bands compared to pre-meditation state. The characteristic path length is high in delta, alpha and beta bands in meditation state. In addition, the meditation state has shown high degree and cluster index in theta and beta bands compared to pre-meditation state. Nonlinear dynamical characterization of HRV during meditation The cardiovascular system is influenced by internal dynamics as well as from various external factors, which makes the system more dynamic and nonlinear. In this part of the thesis, a novel application of using HRV data for studying Chi and Kundalini meditation has been made. The HRV time series are embedded into higher dimensional phase-space using Takens’ embedding theorem to reconstruct the attractor. After estimating the minimum embedding dimension to unfold the attractor dynamics, the complexity of the attractor is computed using correlation dimension, Lyapunov exponent, and nonlinearity scores. In all the analyses, the pre-meditation state is used as control state against which meditation state is compared. The statistical significance of the parameters estimated is tested to discriminate meditation state from control state. The HRV time series of both pre-meditation and meditation have shown similar minimum embedding dimensions in both Chi and Kundalini meditation. Irrespective of the type of meditation, the meditation state has shown high correlation dimension, largest Lyapunov exponent, and low nonlinearity score compared to pre-meditation state. Recurrent quantification analysis of HRV during meditation In this part, a novel application of recurrent quantification analysis (RQA) to HRV during meditation is studied. Here, the time series is embedded into a higher dimensional phase-space and Euclidean distance between the embedded vectors is calculated to form a distance matrix. The matrix is converted into binary matrix by applying a suitable threshold, and plotted as image to get recurrence plot. Various parameters are extracted from the recurrence plot such as percent recurrence rate, diagonal parameters (determinism, divergence, entropy, ratio), and vertical or horizontal parameters (laminarity, trapping time, maximal vertical line length). The procedure is applied to HRV data during meditation and pre-meditation (control) to discriminate between the states. The HRV of meditation state has shown more diagonal line structure whereas more black patches are observed in pre-meditation state. In addition, at low embedding dimensions, the meditation state has shown low recurrence rate, high determinism, low divergence, low entropy, high ratio, high laminarity, high trapping time, and less maximal vertical line length compared to pre-meditation state. These RQA parameters have shown superior performance in discriminating meditation state from control state.

Physiological signals, electroencephalogram (EEG) and heart rate variability (HRV), are generated by complex self-regulating systems. These signals are extremely inhomogeneous and nonstationary, and fluctuate in an irregular and highly complex manner. These fluctuations are due to underlying dynamics of the system. The synchronous neural activity measured as scalp EEG indicates underlying neural dynamics of the brain. Hence, quantitative EEG analysis has become a very useful tool in interpreting results from physiological experiments. The analysis of HRV provides valuable information to assess the autonomous nervous system (ANS). The HRV can be significantly affected by physiological state changes and many disease states. Hence, HRV analysis is becoming a major experimental and diagnostic tool. In this thesis, we focus on the study of EEG and HRV time series using tools from nonlinear time series analysis with special emphasis on its implications in detecting physiological state changes such as, in diseases like epileptic seizure and schizophrenia, and in altered states of consciousness as in sleep and meditation. The proposed nonlinear techniques are used in discriminating different physiological states from control states. Artifact processing of EEG signal Interferences (artifacts) from various sources unavoidably contaminate EEG recordings. In quantitative analysis, results can differ significantly by these artifacts, which may lead to wrong interpretation of the results. In this part of the thesis, we have devised methods to minimize ocular and muscle artifacts in EEG. The artifact correction methods are based on blind source separation (BSS) techniques such as singular value decomposition (SVD), algorithm for multiple signal extraction (AMUSE), canonical correlation analysis (CCA), information maximization (INFOMAX) independent component analysis (ICA) and joint approximate diagonalization of eigen-matrices (JADE) ICA. We have proposed a method to simulate clean and artifact corrupted EEG data based on the BSS methods. In order to enhance the performance of BSS methods, a technique called wavelet-filtered component inclusion method has been introduced. In addition, second-order statistics (SOS) and higher-order statistics (HOS) based BSS methods have been studied considering less number of EEG channels; and performance comparison of these methods has also been made. We have also addressed the problem of simultaneous correction of ocular and muscle artifacts in EEG recordings using the BSS methods. Irrespective of the BSS methods, the component elimination method has introduced high spectral error in all the bands after reconstruction of clean EEG. However, the wavelet filtered component inclusion method has retained almost all spectral powers of EEG channels in theta, alpha, and beta bands after ocular artifact minimization. When the number of EEG channels is very less, the enhanced CCA (SOS BSS) has given superior artifact minimization results than HOS BSS methods, especially in delta band. The component elimination method is used in muscle artifact minimization, and hence the SVD method cannot be used for this purpose since it leads to large spectral distortion of reconstructed EEG. The AMUSE and CCA methods have given comparable performance in muscle artifact minimization. In addition, the JADE method has introduced less mean spectral error compared to other methods. The CCA method has shown superior performance in simultaneous minimization of ocular and muscle artifacts, and AMUSE and JADE methods have given comparable results. Furthermore, the less computation time of wavelet enhanced SOS BSS methods make them very useful in real clinical environments. Fractal characterization of time series In biomedical signal analysis, fractal dimension (FD) is used as a quantitative measure to estimate complexity of physiological signals. Such analysis helps to study physiological processes of underlying systems. The FD can also be used to study dynamics of transitions between different states of systems like brain and ANS, in various physiological and pathological states. In this part, we have proposed a method to estimate FD of time series, called multiresolution box-counting (MRBC) method. A modification of this method resulted in multiresolution length (MRL) method. The estimation performance of the proposed methods is compared with that of Katz, Sevcik, and Higuchi methods, by simulating mathematically defined fractal signals, and also the computation time is compared between the methods. The MRBC and MRL methods have given comparable performance to that of Higuchi method, in estimating FD of waveforms, with the advantage of less computational time. In addition, various properties of the FD are studied and discussed in connection with classical signal processing concepts such as amplitude, frequency, sampling frequency, effect of noise, band width, correlation, etc. The FD value of signals has increased with number of harmonics, noise variance, band-width, and mid-band frequency, and decreased with degree of correlation in AR signal. An analogy between Katz FD and smoothed Teager energy operator has also been made. Application of fractal analysis to EEG and HRV time series The fluctuation of EEG potentials normally depends upon degree of alertness, and varies in amplitude and frequency. Hence, the EEG is an important clinical tool for studying sleep and sleep related disorders, epileptic seizures, schizophrenia, and meditation. In this part of the thesis, we have used FD which gives signal complexity, and detrended fluctuation analysis (DFA) which gives multiscale exponent of time series to quantify EEG. We have extended the concept of FD to multiscale FD to compute complexity of time series at multiple scales. The main applications of the proposed method are epileptic seizure detection, sleep stage detection, schizophrenia EEG analysis, and analysis of heart rate variability during meditation. For seizure detection, we have used intracranial EEG recordings with seizure-free and seizure intervals. In sleep EEG analysis, whole-night sleep EEG is used and results are compared with the manually scored hypnogram. The schizophrenia symptom is further categorized into positive and negative symptoms and complexity is estimated using FD and DFA. We have also analyzed HRV data of Chi and Kundalini meditation using FD and DFA techniques. In all the applications considered, we have tested for statistical significance of the computed parameters, between the case of interest and corresponding control cases, to discriminate between the physiological states. The ocular artifact has reduced FD while muscle artifact increased FD of EEG. The FD of seizure EEG has shown high value compared to that of seizure-free EEG. In addition, the seizure-free EEG has more DFA exponent-1 than seizure EEG. The value of FD of EEG is decreased with deepening of sleep, wake state having high FD value. The FD of REM state sleep EEG showed value between that of wake and state-1. The DFA exponent-1 has increased with deepening of sleep state, having small value for wake state. The REM state has given exponent-1 value between wake and state-1. The schizophrenia subjects have shown lower FD value than healthy controls in all the EEG channels except the bilateral temporal and occipital regions. The positive symptom sub-group has shown comparatively high FD values than healthy controls as well as overall schizophrenia sample in the bilateral tempero-parietal-occipital region. In addition, the positive symptom sub-group has shown significantly higher regional FD values than negative symptom sub-group especially in right temporal region. The overall schizophrenia samples as well as the positive and negative subgroup have shown least FD values in the bilateral frontal region. The values of DFA exponent-2 have shown significant high value in schizophrenia samples. In addition, the schizophrenia group has shown less DFA exponent-1 in bilateral temporal region than healthy control. The FD, multiscale FD, DFA exponents have shown significant performance in discriminating different physiological states from control states. The FD value of HRV time series during meditation is less compared to pre-meditation state in both Chi and Kundalini meditation. Irrespective of the type of meditation, meditation state has shown significantly high DFA exponent-1 than pre-meditation state, and significantly high DFA exponent-2 in pre-meditation state compared to meditation state. Functional connectivity analysis of brain during meditation In functionally related regions of the brain, even in those regions separated by substantial distances, the EEG fluctuations are synchronous, which is termed as functional connectivity. In this part, a novel application of functional connectivity analysis of brain using graph theoretic approach has been made on the EEG recorded from meditation practitioners. We have used 16 channel EEG data from subjects while performing Raja Yoga meditation. The pre-meditation condition is used as control state, against which meditation state is compared. For finding connectivity between EEG of various channels, we have computed pair-wise linear correlation and mutual information between the EEG channels, to form a connection matrix of size 16x16. Then, various graph parameters, such as average connection density, degree of nodes, characteristic path length, and cluster index, are computed from the connection matrix. The computed parameters are projected on to the scalp to get topographic head maps that give spatial variation of the parameter, and results are compared between meditation and pre-meditation states. The meditation state has shown low average connection density, less characteristic path length, and high average degree in fronto-central and central regions. Furthermore, high cluster index is shown in frontal and central regions than pre-meditation state. The parameters such as complexity, characteristic path length and average connection density are used as features in quadratic discriminant classifier to classify meditation and pre-meditation state, and have given good accuracy performance. Connectivity analysis using mutual information has given high average connection density in meditation state in theta, alpha and beta bands compared to pre-meditation state. The characteristic path length is high in delta, alpha and beta bands in meditation state. In addition, the meditation state has shown high degree and cluster index in theta and beta bands compared to pre-meditation state. Nonlinear dynamical characterization of HRV during meditation The cardiovascular system is influenced by internal dynamics as well as from various external factors, which makes the system more dynamic and nonlinear. In this part of the thesis, a novel application of using HRV data for studying Chi and Kundalini meditation has been made. The HRV time series are embedded into higher dimensional phase-space using Takens’ embedding theorem to reconstruct the attractor. After estimating the minimum embedding dimension to unfold the attractor dynamics, the complexity of the attractor is computed using correlation dimension, Lyapunov exponent, and nonlinearity scores. In all the analyses, the pre-meditation state is used as control state against which meditation state is compared. The statistical significance of the parameters estimated is tested to discriminate meditation state from control state. The HRV time series of both pre-meditation and meditation have shown similar minimum embedding dimensions in both Chi and Kundalini meditation. Irrespective of the type of meditation, the meditation state has shown high correlation dimension, largest Lyapunov exponent, and low nonlinearity score compared to pre-meditation state. Recurrent quantification analysis of HRV during meditation In this part, a novel application of recurrent quantification analysis (RQA) to HRV during meditation is studied. Here, the time series is embedded into a higher dimensional phase-space and Euclidean distance between the embedded vectors is calculated to form a distance matrix. The matrix is converted into binary matrix by applying a suitable threshold, and plotted as image to get recurrence plot. Various parameters are extracted from the recurrence plot such as percent recurrence rate, diagonal parameters (determinism, divergence, entropy, ratio), and vertical or horizontal parameters (laminarity, trapping time, maximal vertical line length). The procedure is applied to HRV data during meditation and pre-meditation (control) to discriminate between the states. The HRV of meditation state has shown more diagonal line structure whereas more black patches are observed in pre-meditation state. In addition, at low embedding dimensions, the meditation state has shown low recurrence rate, high determinism, low divergence, low entropy, high ratio, high laminarity, high trapping time, and less maximal vertical line length compared to pre-meditation state. These RQA parameters have shown superior performance in discriminating meditation state from control state.

Indiana University-Purdue University Indianapolis (IUPUI)
Highway noise can cause annoyance, affect sleep patterns, and reduce the property value for people in the proximity. Current methods for analyzing the effectiveness of sound barriers only take loudness into consideration. This paper introduces new methods that can be used to analyze the effectiveness of the sound barriers. Our approach uses psychoacoustic measures including sharpness, roughness, fluctuation, strength, and annoyance. Highway noise is non-stationary, therefore each of these metrics are calculated over a short time. Finally analysis is performed the distribution and change over time. We used nth nearest neighbor algorithm to remove sounds that are not a part of the experiment. In the future, this data can be combined with human surveys to see if the change in sound quality due to the presence of sound barriers has a meaningful impact on people's lives.

碩士
國立交通大學
電子工程學系 電子研究所
104
The heavily medical burden caused by population ageing will become a serious challenge for the current and next generation medical care system. There is an urgent need of low-cost disease prevention and home care programs to lower the possible medical burden in the future. The cardiovascular diseases have been on the list of leading cause of death for years in Taiwan. There is about seventeen million people pass away because of cardiovascular around the world. There is urgent need to get the early prevention tool to reduce the risk of cardiovascular disease all over the world. An effective photoplethysmography (PPG) signal processing system based on ensemble empirical mode decomposition (EEMD) method for acquiring the multiple physiological parameters is proposed in this project. The information of arterial pulse can be obtained by near-infrared. A high quality signal can be extracted through the proposed EEMD algorithm. Based on the most advanced semiconductor industry in Taiwan, the regulation of autonomic nervous system (ANS), RI and SI can be derived in real-time and monitored continuously. It makes the at-home care possible and lowers the rate of cardiovascular diseases and medical expenses through long-term monitoring. PPG signal acquired by the PPG capture circuit is sampled through the ADC at sample frequency of 200Hz after being filtered by the band pass filter. The digitized data are decomposed into IMFs with physiological meanings by the EEMD IC. The output IMFs are wirelessly sent to a computer via a Bluetooth module. Then the regulation of autonomic nervous system , RI and SI can be derived and display on the GUI. To overcome the noise and aliasing effect caused by nonstationary signals, many innovative and effective modules were developed in this thesis. The proposed HHT SoC design could be implemented in hardware with limited resources and fabricated under TSMC 90 nm CMOS technology. To assess the potential risk of cardiovascular, the IMFs with physiological meanings can be extracted from PPG. The RI, SI, LF, HF and VHF can be derived as the parameters to help the diagnosis of cardiovascular disease.

Indiana University-Purdue University Indianapolis (IUPUI)
Non-invasive surface recording of bioelectric potentials continues to be an essential tool in a variety of research and medical diagnostic procedures. However, the integrity of these recordings, and hence the reliability of subsequent analysis, diagnosis, or recommendations based on the recordings, can be significantly compromised when various types of noise are allowed to penetrate the recording circuit and contaminate the signals. In particular, for bioelectric phenomena in which the amplitude of the biosignal is relatively low, such as muscle activity (typically on the order of millivolts) or neural traffic (microvolts), external noise may substantially contaminate or even completely overwhelm the signal. In such circumstances, the tissue-electrode interface is typically the primary point of signal contamination since its impedance is relatively high compared to the rest of the recording circuit. Therefore, in the recording of low-amplitude biological signals, it is of paramount importance to minimize the impedance of the tissue-electrode interface in order to consistently obtain low-noise recordings. The aims of the current work were (1) to complete the development of a set of tools for rapid, simple, and reliable full-spectrum characterization and analytical modeling of the complex impedance of the tissue-electrode interface, and (2) to characterize the interfacial impedance and signal-to-noise ratio (SNR) at the surface of the skin across a variety of preparation methods and determine a factor or set of factors that contribute most effectively to the reduction of tissue-electrode impedance and noise contamination during recording. Specifically, we desired to test an initial hypothesis that surface abrasion is the principal determining factor in skin preparation to achieve consistently low-impedance, low-noise recordings. During the course of this master’s study, (1) a system with portable, battery-powered hardware and robust acquisition/analysis software for broadband impedance characterization has been achieved, and (2) the effects of skin preparation methods on the impedance of the tissue-electrode interface and the SNR of surface electromyographic recordings have been systematically quantified and compared in human subjects. We found our hypothesis to be strongly supported by the results: the degree of surface abrasion was the only factor that could be correlated to significant differences in either the interfacial impedance or the SNR. Given these findings, we believe that abrasion holds the key to consistently obtaining a low-impedance contact interface and high-quality recordings and should thus be considered an essential component of proper skin preparation prior to attachment of electrodes for recording of small bioelectric surface potentials.

Indiana University-Purdue University Indianapolis (IUPUI)
Carbon nanomaterials are widely produced and used in industry, medicine and scientific research. To examine the impact of exposure to nanoparticles on human health, the human airway epithelial cell line, Calu-3, was used to evaluate changes in the cellular proteome that could account for alterations in cellular function of airway epithelia after 24 h exposure to 10 μg/mL and 100 ng/mL of two common carbon nanoparticles, singleand multi-wall carbon nanotubes (SWCNT, MWCNT). After exposure to the nanoparticles, label-free quantitative mass spectrometry (LFQMS) was used to study differential protein expression. Ingenuity Pathway Analysis (IPA) was used to conduct a bioinformatics analysis of proteins identified by LFQMS. Interestingly, after exposure to a high concentration (10 μg/mL; 0.4 μg/cm2) of MWCNT or SWCNT, only 8 and 13 proteins, respectively, exhibited changes in abundance. In contrast, the abundance of hundreds of proteins was altered in response to a low concentration (100 ng/mL; 4 ng/cm2) of either CNT. Of the 281 and 282 proteins that were significantly altered in response to MWCNT or SWCNT, respectively, 231 proteins were the same. Bioinformatic analyses found that the proteins common to both kinds of nanotubes are associated with the cellular functions of cell death and survival, cell-to-cell signaling and interaction, cellular assembly and organization, cellular growth and proliferation, infectious disease, molecular transport and protein synthesis. The decrease in expression of the majority proteins suggests a general stress response to protect cells. The STRING database was used to analyze the various functional protein networks. Interestingly, some proteins like cadherin 1 (CDH1), signal transducer and activator of transcription 1 (STAT1), junction plakoglobin (JUP), and apoptosis-associated speck-like protein containing a CARD (PYCARD), appear in several functional categories and tend to be in the center of the networks. This central positioning suggests they may play important roles in multiple cellular functions and activities that are altered in response to carbon nanotube exposure. To examine the effect of nanotubes on the plasma membrane, we investigated the interaction of short purified MWCNT with model lipid membranes using a planar bilayer workstation. Bilayer lipid membranes were synthesized using neutral 1, 2-diphytanoylsn-glycero-3-phosphocholine (DPhPC) in 1 M KCl. The ion channel model protein, Gramicidin A (gA), was incorporated into the bilayers and used to measure the effect of MWCNT on ion transport. The opening and closing of ion channels, amplitude of current, and open probability and lifetime of ion channels were measured and analyzed by Clampfit. The presence of an intermediate concentration of MWCNT (2 μg/ml) could be related to a statistically significant decrease of the open probability and lifetime of gA channels. The proteomic studies revealed changes in response to CNT exposure. An analysis of the changes using multiple databases revealed alterations in pathways, which were consistent with the physiological changes that were observed in cultured cells exposed to very low concentrations of CNT. The physiological changes included the break down of the barrier function and the inhibition of the mucocillary clearance, both of which could increase the risk of CNT’s toxicity to human health. The biophysical studies indicate MWCNTs have an effect on single channel kinetics of Gramicidin A model cation channel. These changes are consistent with the inhibitory effect of nanoparticles on hormone stimulated transepithelial ion flux, but additional experiments will be necessary to substantiate this correlation.