Advances in Cardiac Signal...

Advances in Cardiac Signal Processing

123

ProcessingAdvances in Cardiac Signal

Rajendra Acharya U, Jasjit S. Suri, Jos A.E. Spaan an d S .M. Krishnan

With 268 Figures, 13 in color and 59 Tables

Library of Congress Control Number:

This work is subject to copyright. All rights are reserved, whether the whole or part of the materialis concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broad-casting, reproduction on microfilm or in any other way, and storage in data banks. Duplication ofthis publication or parts thereof is permitted only under the provisions of the German Copyright Lawof September 9, 1965, in its current version, and permission for use must always be obtained fromSpringer. Violations are liable to prosecution under the German Copyright Law.

Springer is a part of Springer Science+Business Media.

springer.com

The use of general descriptive names, registered names, trademarks, etc. in this publication does notimply, even in the absence of a specific statement, that such names are exempt from the relevant pro-tective laws and regulations and therefore free for general use.

Printed on acid-free paper 5 4 3 2 1 0

macro packageA ELT X

62/3100/SPi

2006931013

ISBN-13 978-3-540-36674-4 Springer Berlin Heidelberg New YorkISBN-10 3-540-36674-1 Springer Berlin Heidelberg New York

© Springer-Verlag Berlin Heidelberg 2007

SPi using a SpringerTypesetting by authors and

SPIN 11731283

Cover design: de’blik in Berlin

Visiting Faculty

Ngee Ann Polytechnic

Prof. Jasjit Suri

Prof. Jos A.E. Spaan

Dept. of Medical Physics

Academic Medical CenterUniversity of Amsterdam

P.O. Box 227001100 DE AmsterdamThe Netherlands

Mr. Shankar M. Krishnan8 Whittier PlaceBOSTON MA 02114

Dr. Rajendra Acharya U

E-mail: [email protected]



USA


Dept. of ECE

Singapore 599489

ID, USA and Biomedical Technologies, CO, USAIdaho State Biomedical Research Institute

Editor Chief MBEC

Preface

Various disciplines have been benefited by the advent of high-performancecomputing in achieving practical solutions to their problems and the area ofhealth care is no exception to this. Signal processing and data mining toolshave been developed to enhance the computational capabilities so as to helpclinicians in diagnosis and treatment.

The electrocardiogram (ECG) is a representative signal containing infor-mation about the condition of the heart. The shape and size of the P-QRS-Twave and the time intervals between various peaks contains useful informa-tion about the nature of disease afflicting the heart. However, the human ob-server cannot directly monitor these subtle details. Besides, since biosignalsare highly subjective, the symptoms may appear at random in the timescale.The presence of cardiac abnormalities are generally reflected in the shape ofECG waveform and heart rate. However, by the very nature of biosignals,this reflection would be random in the timescale. That is, the diseases maynot show up all the time, but would manifest at certain irregular (random)intervals during the day. Therefore the study of ECG pattern and heart ratevariability has to be carried out over extended periods of time (i.e., for 24hours). Naturally the volume of the data to be handled is enormous and itsstudy is tedious and time consuming. As a consequence, the possibility of theanalyst missing (or misreading) vital information is high. Hence, the electro-cardiogram and heart rate variability signal parameters, extracted and ana-lyzed using computers, are highly useful in diagnostics. This book deals withthe acquisition, extraction of the various morphological features, classificationand analysis of the cardiac signals. This book comprises of twenty chapters.It deals with the acquisition, extraction of the various features, classification,cardiac function during different phases, arterial pressures and analysis of fluidperfusion through soft tissues of left ventricular myocardium heart muscle.

The Chapter 1 of the book explains the recording of ECG and electro-physiologic principles underlying cardiac electrical activity. The terms usedin clinical electrocardiography are defined, and the depolarization sequence

VI Preface

of the heart is reviewed. Different modes of lead placement and the variousarrhythmia are discussed in depth.

Generally the various characteristics features of ECG are extracted andused for decision making purposes. This makes the decision making and di-agnosis process simpler and faster. Hence appropriate feature description andextraction becomes the most important component in cardiac health diagnos-tics. In Chapter 2, various techniques for feature extraction are described.

In Chapter 3, a detailed discussion is done on the prediction of heart ratesignals using Auto Regressive (Burg’s method), and Recursive Neural Network(Elman’s method). The performance of the prediction methods is evaluatedusing the Normalized Root Mean Square Error.

Visualization of ECG data is an important part of the display in life threat-ening state. The symptoms of disease may occur at random in the timescale,and the physician has to view enormously large data to diagnose the affliction.The tedium of reading large data can be considerably reduced by using thecomputer for a hierarchical visualization technique in Chapter 4.

Heart Rate Variability (HRV) is a reliable reflection of the many physio-logical factors modulating the normal rhythm of the heart. It is a powerfulmeans of observing the interplay between the sympathetic and parasympa-thetic nervous systems. Chapter 5 discusses the various applications of heartrate variability and different linear, non-linear techniques used for the analysisof the HRV.

The fusion of ECG, blood pressure, blood oxygen saturation and respi-ratory data for achieving improved clinical diagnosis of patients in cardiaccare units. Computer based analysis and display, of the heterogeneous signalsfor the detection of life threatening states is demonstrated using fuzzy logicbased data fusion. A new parameter called deterioration index is proposed toevaluate the severity of the disease and results are tabulated for various casesin Chapter 6.

In Chapter 7, a new method is developed using artificial neural network(ANN) techniques for classification of the states of patients in intensive careunit (ICU) from the electrocardiogram (ECG) obtained from the patients.The states are classified into normal, abnormal and life threatening classesusing different neural network techniques.

In Chapter 8, the AR modeling technique is proposed to classify six typesof cardiac arrhythmias. Quadratic discriminant function (QDF) based classifi-cation algorithm was performed in various stages. The AR modeling techniquecan be utilized for parameter extraction and are subsequently used for classi-fication and diagnosis in telemedicine system.

Chapter 9, presents a comparative approach for classification of eight car-diac diseases using classifiers as, neural network, fuzzy and neuro-fuzzy. Thefeatures are extracted from the raw heart rate signals using the non-linearsignal processing techniques and fed to the classifiers for classification.

Digital watermarking technique is adapted in Chapter 10, for interleavingpatient information and cardiac signals with medical images, to reduce storage

Preface VII

and transmission overheads. The text data is encrypted and the graphicalsignals like heart rate signals are compressed and subsequently interleavedwith the error correcting codes to increase the reliability during transmissionand storage of the images.

In Chapter 11, we developed a mathematical and clinical evaluation fora quantitative understanding of the cardiac function and dysfunction duringdiastolic-filling and ejection phases. The computation of clinical-diagnosticmeasures of (i) left ventricular (LV) volume-dependent passive elastance andactive elastance, (ii) LV maximal change rate of normalized wall stress withrespect to intracavitary pressure are discussed.

Chapter 12 discusses the determination of arterial pressure pulse wavepropagation velocity and arterial properties. It represents the analytical basisfor determining arterial elasticity parameters from pulse wave velocity (PWV).It is shown that the PWV can be used as an index of arterial distensibility.

Morphological filtering algorithm using modified morphological operatorsis proposed for baseline correction and noise suppression in ECG signals. Theperformance of the proposed algorithm was evaluated by using simulated sig-nals and clinically acquired ECG data from a standard set in Chapter 13.

ECG, blood pressure and respiratory signals can provide important infor-mation on the pathophysiology of the cardiovascular regulatory mechanisms.Spectral and cross-spectral analysis of these signals gives quantitative infor-mation which can be of potential interest in clinical studies. This discussedin Chapter 14 using a methodology based upon multivariate autoregressiveidentification and parametric power spectral density estimation.

Considering the heart as a nonlinear complex system and processing var-ious cardiovascular signals like ECG, HRV and ABP seems to provide veryuseful information for detection of abnormalities in the condition of the heartthan is possible by conventional means. In Chapter 15, a new multidimensionalphase space analysis of these cardiovascular signals using weighted spatial fill-ing index has been proposed for detecting cardiac dysfunction.

Chapter 16, deals with the linear, non-linear and wavelet analysis of eightkinds of cardiac abnormalities. It also, presents the ranges of linear and non-linear parameters calculated for them with a confidence level of more than90%. A unique visual pattern of scalogram and phase space plot of heart ratesignal, which may provide considerable insight into the nature and pattern ofthe disease are discussed.

Myocardial perfusion is the flow or forced passage of blood through thecoronary arteries and to the LV myocardium. A review of the biomechani-cal models focused on the analysis of fluid perfusion through soft tissues ofleft ventricular myocardium heart muscle in particular and its fundamentalconstitutive relations was included in Chapter 17.

Conventional Fourier Transform techniques are not suitable for analysis ofnonstationary signals. Wavelet analysis, on the other hand, provides a betterinsight into both the timing and intensity of transient events. Chapter 18 deals

VIII Preface

with the theory of the wavelet transforms and few applications related to thecardiac signals are dealt exhaustively in this.

Several phenomenon show the 1/f fluctuation in nature which is of interest.Recently electrical appliances have a control system included 1/f fluctuation.Heart rate in daily life also show the 1/f fluctuation. After suffering severaldiseases, the patients still show the 1/f fluctuation. In Chapter 19, the 1/ffluctuation in brain-death patients is discussed with results.

Stress measurement by heart rate variability is popular not only in dailylife but in hospital ward. In ICU, the spaghetti syndrome gives rise to stressin patients and easily estimated by heart rate monitor. After stroke, severalpeople suffer the aphasia. In Chapter 20, the stress during speech therapyusing the method of heart rate variability is investigated.

It is our humble hope that this book will assist those who seek to enrichtheir lives and those of others with the wonderful powers of cardiac signalprocessing. Electrical, Computer and Biomedical Engineering are great fields,contributing immensely to the service of humanity.

Rajendra Acharya UJasjit S. Suri

Jos A E SpaanShankar M Krishnan

Cardiac Signal Processing

Rajendra Acharya UJasjit S. SuriShankar M. KrishnanJos A. E. SpaanEditors

Rajendra Acharya U, PhD is a visiting faculty in Ngee Ann Poly-technic, Singapore. He received his Doctorate National Institute of TechnologyKarnataka, Surathkal, India in 2001. He served as Assistant Professor, in Bio-Medical Engineering department in Manipal Institute of Technology, Manipal,India till 2001. His current interests are Visualization, Bio-signal Processingand Image Processing.

Jasjit S. Suri, PhD is an innovator, scientist, and an internationally knownworld leader and has spent over 20 years in the field of biomedical engineering/sciences and its management. He received his Doctorate from Universityof Washington, Seattle and Management Sciences from Weatherhead, CaseWestern Reserve University, Cleveland. Dr. Suri was crowned with President’s

Gold medal in 1980 and the Fellow of American Institute of Medical andBiological Engineering for his outstanding contributions.

Jos A.E. Spaan is Full Professor of Medical Physics at the faculty ofMedicine, Academic Medical Center and the Faculty of Science, Universityof Amsterdam. He is Editor in Chief of Medical & Biological Engineering &Computing, the official journal of the International Federation of medical andBiological Engineering, the IFMBE.

Shankar M. Krishnan received the Ph.D. degree in Electrical Engineeringfrom the University of Rhode Island and is conducting research at the RhodeIsland Hospital. He is presently a consultant and advisor to various interna-tional projects in the biomedical and clinical engineering field.

X

Contents

1 The ElectrocardiogramJohnny Chee and Swee-Chong Seow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Anatomy of the Heart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.2 Electrical Conduction System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2.1 The Sino-Atrial (S-A) Node . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.2.2 Depolarisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.2.3 The Atrioventricular (A-V) Node . . . . . . . . . . . . . . . . . . . . . . . 91.2.4 Automaticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101.2.5 Accessory Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111.2.6 Excitable Tissue and Generation of Ionic Currents . . . . . . . . 111.2.7 ECG Signal Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131.2.8 ECG Lead Placements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141.2.9 The Limb Leads (Bipolar) – Leads I, II, III . . . . . . . . . . . . . . 151.2.10 The Augmented Limb Leads (Unipolar) – Leads aVL,

aVR, aVL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.2.11 The Precordial Leads (Unipolar) – Leads V1, V2, V3,

V4, V5, V6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161.3 Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.3.1 Sinus Node Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191.3.2 Atrial Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211.3.3 Junctional Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241.3.4 Ventricular Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261.3.5 Atrioventricular Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301.3.6 Bundle Branch and Fascicular Blocks . . . . . . . . . . . . . . . . . . . 33

1.4 Miscellaneous Electrocardiogram Changes . . . . . . . . . . . . . . . . . . . . . 351.4.1 Enlargement of the Myocardium . . . . . . . . . . . . . . . . . . . . . . . 371.4.2 Pericarditis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381.4.3 Electrolyte Imbalance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381.4.4 Drug Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411.4.5 Pulmonary Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431.4.6 Early Repolarisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

XII Contents

1.4.7 Hypothermia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451.4.8 Preexcitation Syndromes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

1.5 Myocardial Ischemia, Injury and Infarction . . . . . . . . . . . . . . . . . . . . 481.5.1 Zones of Ischemia, Injury and Infarction . . . . . . . . . . . . . . . . . 481.5.2 Myocardial Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491.5.3 Acute Myocardial Infarction (AMI) . . . . . . . . . . . . . . . . . . . . . 491.5.4 ECG Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491.5.5 Evolution of the Deep Q-Wave . . . . . . . . . . . . . . . . . . . . . . . . . 501.5.6 Silent Ischemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511.5.7 Stable Angina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511.5.8 Unstable Angina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2 Analysis of ElectrocardiogramsN. Kannathal, U. Rajendra Acharya, Paul Joseph, Lim Choo Minand Jasjit S. Suri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552.1 Steps in ECG Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552.2 Preprocessing of ECG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

2.2.1 Noise Filtering Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592.3 QRS Complex Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60

2.3.1 QRS Detection Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612.4 Detection of QRS Complex Onset and Offset . . . . . . . . . . . . . . . . . . . 652.5 ST Segment Analyzer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 662.6 Complexity Analysis of ECG Signal . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

2.6.1 Spectral Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 682.6.2 Temporal Complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

2.7 Sample Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692.7.1 Noise Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 692.7.2 QRS Complex Peak Detection . . . . . . . . . . . . . . . . . . . . . . . . . 702.7.3 Detection of QRS Complex Onset and Offset . . . . . . . . . . . . 762.7.4 ST Segment Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.7.5 T Peak Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

3 Prediction of Cardiac Signals Using Linearand Nonlinear TechniquesN. Kannathal, U. Rajendra Acharya, Lim Choo Minand Jasjit S. Suri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833.1 Data Acquisition and Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . 863.2 Modeling Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

3.2.1 Linear Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883.2.2 Nonlinear Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

3.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 943.4 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Contents XIII

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105

4 Visualization of Cardiac Health Using ElectrocardiogramsU. Rajendra Acharya, P. Subbanna Bhat, U.C. Niranjan, N. Kannathal,Lim Choo Min and Jasjit S. Suri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1094.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1094.2 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1114.3 Arrhythmia Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1114.4 Data Handling and Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1124.5 Hierarchical Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1144.6 Sector Graph Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1154.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

5 Heart Rate VariabilityRajendra Acharya U, Paul Joseph K, Kannathal N, Lim Choo Minand Jasjit Suri S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.1 Physiological Phenomenon of HRV . . . . . . . . . . . . . . . . . . . . . . . . . . . 1215.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122

5.2.1 The Autonomic Nervous System (ANS) . . . . . . . . . . . . . . . . . 1245.2.2 HRV and Blood Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1245.2.3 HRV and Myocardial Infarction . . . . . . . . . . . . . . . . . . . . . . . . 1255.2.4 HRV and Nervous System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265.2.5 HRV and Cardiac Arrhythmia . . . . . . . . . . . . . . . . . . . . . . . . . 1265.2.6 HRV in Diabetes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265.2.7 HRV and Respiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1275.2.8 HRV and Renal Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1285.2.9 HRV and Gender, Age . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1285.2.10 HRV and Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1295.2.11 HRV and Smoking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1295.2.12 HRV and Alcohol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1305.2.13 HRV and Sleep . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1305.2.14 HRV and Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130

5.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1315.3.1 Time Domain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1315.3.2 Analysis by Geometrical Method . . . . . . . . . . . . . . . . . . . . . . . 1325.3.3 Poincare Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132

5.4 Frequency Domain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1345.4.1 Limitations of Fourier Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 1365.4.2 Higher Order Spectra (HOS) . . . . . . . . . . . . . . . . . . . . . . . . . . . 1375.4.3 Short Time Fourier Transform (STFT) . . . . . . . . . . . . . . . . . . 1395.4.4 Continuous Time Wavelet Transform (CWT) Analysis . . . . 140

XIV Contents

5.5 Nonlinear Methods of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1405.5.1 Capacity Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1415.5.2 Correlation Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1415.5.3 Lyapunov Exponent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1425.5.4 Hurst Exponent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1425.5.5 Detrended Fluctuation Analysis . . . . . . . . . . . . . . . . . . . . . . . . 1435.5.6 Entropies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1455.5.7 Fractal Dimension (FD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.5.8 Recurrence Plots (RP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

5.6 Requirements for Non-Linear Analysis . . . . . . . . . . . . . . . . . . . . . . . . 1515.6.1 Surrogate Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

5.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1525.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6 Data Fusion of Multimodal Cardiovascular SignalsKannathal N, Rajendra Acharya U, E.Y.K. Ng, Lim Choo Min,Jasjit S Suri, Jos A E Spaan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1676.1 Approaches for Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1686.2 Rule Based Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1716.3 Introduction to Fuzzy Based Decision Making . . . . . . . . . . . . . . . . . . 1726.4 Fuzzy Logic Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

6.4.1 Fuzzy-Logic Decision Function Created by FuzzifyingBoolean Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

6.4.2 Fuzzy-Logic Patient Deterioration Index . . . . . . . . . . . . . . . . 1766.5 Patient States Diagnosis System Implementing Data Fusion . . . . . . 1776.6 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1796.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

7 Classification of Cardiac Patient StatesUsing Artificial Neural NetworksN. Kannathal, U. Rajendra Acharya, Lim Choo Minand Jasjit S. Suri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1877.1 Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1907.2 Backpropagation Learning Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 1917.2 Self-Organizing Maps as Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

7.2.1 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1947.2.2 Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

7.3 Radial Basis Function Networks as Classifiers . . . . . . . . . . . . . . . . . . 1967.3.1 Radial Basis Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196

7.4 Overview of Cardiac Patients States Classification System . . . . . . . 1987.4.1 Implementation using Back Propagation . . . . . . . . . . . . . . . . 1987.4.2 Implementation using Self-Organizing Maps . . . . . . . . . . . . . 2007.4.3 Implementation using Radial Basis Functions . . . . . . . . . . . . 200

Contents XV

7.5 Effect of Number of Training Data on Network Performance . . . . . 2027.6 Network Performance Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2047.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

8 The Application of Autoregressive Modelingin Cardiac Arrhythmia ClassificationDingfei Ge, Narayanan Srinivasan, S M Krishnan . . . . . . . . . . . . . . . . . . . 2098.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2098.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

8.2.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2118.2.2 AR Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2128.2.3 ECG Feature Extraction for Classification . . . . . . . . . . . . . . . 2138.2.4 QDF-based Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213

8.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2158.3.1 AR Modeling and Feature Extraction Results . . . . . . . . . . . . 2158.3.2 Classification Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

8.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2228.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224

9 Classification of Cardiac Abnormalities Using Heart RateSignals: A Comparative StudyU. Rajendra Acharya, N. Kannathal, P. Subbanna Bhat, Jasjit S. Suri,Lim Choo Min and Jos A.E. Spaan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2279.1 Neural Network Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2299.2 Inputs to the Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2309.3 Surrogate Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2319.4 Fuzzy Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2319.5 Neuro-Fuzzy Classifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2349.6 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2399.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2409.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

10 Storage and Transmission of Cardiac Datawith Medical ImagesU. Rajendra Acharya, P. Subbanna Bhat, U. C. Niranjan,Sathish Kumar, N. Kannathal, Lim Choo Min and Jasjit Suri . . . . . . . . . 24510.1 Concept of Interleaving . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24610.2 The Interleaving Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24710.3 Encryption of the Text File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24710.4 Encryption of the Graphic File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24810.5 Interleaving in DFT Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25010.6 Interleaving in DCT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25110.7 Interleaving in Wavelet Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

XVI Contents

10.8 Evaluation of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25210.9 Interleaving Error Correcting Codes . . . . . . . . . . . . . . . . . . . . . . . . . . 25510.10 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25910.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260

11 Assessment of Cardiac Function in Filling& Systolic Ejection Phases: A Mathematicaland Clinical EvaluationLiang Zhong, Dhanjoo N. Ghista, Eddie Y.K. Ng, Ru San Tan,Soo Teik Lim and Terrance S.J. Chua . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26311.1 Introduction and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26311.2 Known Characterization of Left Ventricular Function . . . . . . . . . . . 263

11.2.1 Left Ventricular Diastolic Dysfunction Characterization . . . 26411.2.2 Left Ventricular Systolic Dysfunction Assessment . . . . . . . . . 265

11.3 Justification for Clinical Indices of Cardiac Function . . . . . . . . . . . . 26611.3.1 Indices Characterizing the “Passive” Ventricle . . . . . . . . . . . 26711.3.2 Indices Characterizing Contractility . . . . . . . . . . . . . . . . . . . . 268

11.4 Passive and Active Elastances of the Left Ventricle . . . . . . . . . . . . . 27311.5 LV dσ∗/dtmax as Noninvasive Contractility Index Based on

Wall-Normalized Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27911.5.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27911.5.2 Derivation of dσ∗/dtmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27911.5.3 Clinical Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

11.6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282

12 Arterial Wave Propagation and Reflectionat a Bifurcation SiteDhanjoo N Ghista, Liang Zhong, Eddie Y.K Ng and Ru San Tan . . . . . . 28912.1 Analysis for Pulse Wave Propagation Velocity . . . . . . . . . . . . . . . . . . 29012.2 Depiction of Pulse Pressure Wave Propagation . . . . . . . . . . . . . . . . . 29612.3 Determination of Arterial Elasticity Parameters . . . . . . . . . . . . . . . . 29712.4 Determination of Arterial Impedance from PWV and Arterial

Cross-Section Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30012.4.1 Peripheral Resistance (R) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30112.4.2 Arterial Impedance (Z0) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

12.5 Reflection at Arterial Bifurcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30512.6 Conclusing Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308

13 ECG Signal Conditioning by Morphological FiltersS.M. Krishnan, Kwoh Chee Keong, Sun Yan and Chan Kap Luk . . . . . . . 31113.1 Mathematical Morphology Operators . . . . . . . . . . . . . . . . . . . . . . . . . 31313.2 Proposed MMF Algorithm for ECG Signal Conditioning . . . . . . . . 314

13.2.1 Baseline Correction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314

Contents XVII

13.2.2 Noise Suppression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31513.2.3 Filtering Performance Evaluation . . . . . . . . . . . . . . . . . . . . . . . 316

13.3 Experimental Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 31713.3.1 Algorithm Testing Using Simulated Data . . . . . . . . . . . . . . . . 31713.3.2 Algorithm Testing Using MIT-BIH

Arrhythmia Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32313.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

14 Multivariate Analysis for Cardiovascularand Respiratory SignalsNarayanan Srinivasan and S.M. Krishnan . . . . . . . . . . . . . . . . . . . . . . . . . . 32714.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32914.2 Multichannel Spectral Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33014.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331

14.3.1 ECG and ABP Signal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 33214.3.2 ECG and Respiratory Signal Analysis . . . . . . . . . . . . . . . . . . . 334


15 Phase Space Analysis for Cardiovascular SignalsS.M. Krishnan, D. Narayana Dutt, Y.W. Chanand V. Anantharaman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33915.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34115.2 Extension to Higher Dimension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34315.3 Weighted Spatial Filling Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34315.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

15.4.1 ECG Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34415.4.2 HRV and ABP Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34715.4.3 Graphical Representation in Higher Dimensions . . . . . . . . . . 349


16 Linear, Non-Linear and Wavelet Analysisof Cardiac Health Using Heart Rate SignalsU. Rajendra Acharya, N. Kannathal, Lim Choo Minand Jasjit S. Suri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35516.1 Data Used for Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35716.2 Methods Used for Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358

16.2.1 Time Domain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35816.2.2 Frequency Domain Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 35816.2.3 Nonlinear Methods of Analysis . . . . . . . . . . . . . . . . . . . . . . . . . 359

16.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35916.4 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36316.5 Surrogate Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37016.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371

XVIII Contents

17 Soft Tissue Biomechanics of the Left VentricularMyocardiumE.Y.K. Ng, Dhanjoo N. Ghista, Reginald C. Jegatheseand Jian-Jun Shu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37717.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37717.2 Previous Research on Soft Tissue Biomechanics . . . . . . . . . . . . . . . . 378

17.2.1 Model Developments for AnalyzingMyocardial Perfusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378

17.2.2 Myocardial Material Properties and Their Influenceon Perfusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380

17.2.3 Myocardial Continuum Mechanics Approachfor Analyzing Perfusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383

17.2.4 LV Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38417.2.5 Muscle Perfusion Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385

17.3 Integrative Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38717.3.1 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39217.3.2 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393


18 Wavelets and its Application in CardiologyJayachandran E S, Paul Joseph K . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40718.1 The Fourier Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40718.2 Short Time Fourier Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40818.3 Continuous Wavelet Transform (CWT) . . . . . . . . . . . . . . . . . . . . . . . . 409

18.3.1 The Inverse Continuous Wavelet Transform (ICWT) . . . . . . 41018.4 Discrete Wavelet Transform (DWT) . . . . . . . . . . . . . . . . . . . . . . . . . . 41018.5 Multi Resolution Analysis (MRA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41118.6 Some Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414

18.6.1 Decomposition of ECG signal . . . . . . . . . . . . . . . . . . . . . . . . . . 41518.6.2 Detection of Myocardial Ischemia . . . . . . . . . . . . . . . . . . . . . . 41718.6.3 De-noising ECG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41918.6.4 Classification of Arrhythmias . . . . . . . . . . . . . . . . . . . . . . . . . . 420


19 1/f Fluctuation of Heart Rate in Postoperativeand Brain-Dead PatientsNakajima K, Tamura T, Sasaki K, Maekawa T . . . . . . . . . . . . . . . . . . . . . . 42319.1 Data Acquistion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42419.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42519.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42619.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440

Contents XIX

20 Stress During Speech TherapyToshiyo Tamura, Ayako Maeda, Masaki Sekine, Yuji Higashiand Toshiro Fujimoto . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44320.1 Speech Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44320.2 Heart Rate Variability and Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44420.3 Calculating the RR Interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44420.4 Experimental Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

20.4.1 Subjects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44520.4.2 Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 446

20.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44620.6 Program with Relaxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44620.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459

List of Contributors

Johnny CheeBiomedical Engineering CentreSchool of Engineering, Ngee AnnPolytechnic, Singapore

Seow Swee ChongCardiologist, National UniversityHospital, The Heart InstituteSingapore

Kannathal N, Rajendra AcharyaU, Lim Choo MinDepartment of Electronics andComputer Engineering, Ngee AnnPolytechnic, Singapore

Jayachandran E S, PaulJoseph KDepartment of Electrical Engineer-ing, National Institute of TechnologyCalicut, India

Jasjit S. SuriIdaho’s Biomedical ResearchInstitute, ID, USA and BiomedicalTechnologies, Inc., CO, USA

and

Research ProfessorBiomedical ResearchInstitute, ISU, USA

P Subbanna BhatDepartment of Electronics & Com-munication Engineering, NationalInstitute of Technology KarnatakaSurathkal, India

Niranjan U C, Sathish KumarDepartment of Biomedical Engineer-ing, Manipal Institute of TechnologyManipal, India

E. Y. K. NgSchool of Mechanical and AerospaceEngineering, Nanyang TechnologicalUniversity, Singapore

Jos A E SpaanDepartment of Medical PhysicsAcademic Medical Center, Universityof Amsterdam, Netherlands

Liang Zhong, Eddie Y.K NgSchool of Mechanical and AerospaceEngineering, Nanyang TechnologicalUniversity, Singapore

Dhanjoo N GhistaSchool of Chemical & BiomedicalEngineering, Nanyang TechnologicalUniversity, Singapore

XXII List of Contributors

Ru San Tan, Soo Teik Lim,Terrance S. J. ChuaDepartment of Cardiology, NationalHeart Center, Singapore

S. M. Krishnan, Kwoh CheeKeong, Sun Yan, Chan KapLuk, Narayanan SrinivasanBiomedical Engineering ResearchCentre, Research TechnoPlaza, 50Nanyang Drive, XFrontiers BlockNanyang Technological UniversitySingapore, 6th Storey, Singapore637553

D Narayana DuttIndian Institute of ScienceBangalore, India

Y W Chan, V. AnantharamanSingapore General HospitalSingapore

Dingfei GeSchool of Information and Electron-ics Engineering, Zhe Jinag Universityof Science and Technology, Hangzhou310012, China

Narayanan SrinivasanIndian Institute of ScienceBangalore, India

Nakajima K, Sasaki KDivision of Bio-InformationEngineering, Faculty of EngineeringUniversity of Toyama, Japan

Tamura TDepartment of Medical SystemEngineering, Faculty of EngineeringChiba University, Japan

T MaekawaDivision of Stress and Bio-responseMedicine, School of Medicine,Yamaguchi University; AdvancedMedical Emergency and CriticalCare Center, Yamaguchi UniversityHospital, Japan

Toshiyo Tamura, Masaki SekineDepartment of Medical Systemengineering, School of EngineeringChiba University

Ayako Maeda, Yuji Higashi,Toshiro FujimotoRehabilitation Centre, Fujimoto-Hayasuzu Hospital, Japan

R.C. JegatheseCollege of Engineering, NanyangTechnological University, 50 NanyangAvenue Singapore 639798

1

The Electrocardiogram

Johnny Chee and Swee-Chong Seow

The electrocardiogram (ECG) is a graphical recording of the electrical signalsgenerated by the heart. The signals are generated when cardiac musclesdepolarise in response to electrical impulses generated by pacemaker cells.Upon depolarisation, the muscles contract and pump blood throughout thebody. The ECG reveals many things about the heart, including its rhythm,whether its electrical conduction paths are intact, whether certain chambersare enlarged, and even the approximate ischemic location in the event of aheart attack (myocardial infarction).

A typical ECG recording from a normal person (Fig. 1.1) and human heart(Fig. 1.2) are shown below.

The ECG is described by waves, segments and intervals:

• Waves are labelled using the letters P, QRS, T and U. The typical normalECG may not show a U wave.

• Segments are time durations between waves, e.g. P-R segment is theduration between the P and R waves (or P and Q waves, when Q wave ispresent).

• Intervals are time durations that include waves and segments, e.g. P-Rinterval is made up of the P-wave and the P-R segment.

In the absence of an S-wave, the junction where the R-wave joins the S-Tsegment is described as the J-point (Fig. 1.3).

The significance of the waves (Fig. 1.4) may be broadly described asfollows:

• P-wave corresponds to the depolarisation of the atrial myocardium (mus-cles of upper chambers of the heart), and indicates the start of atrialcontraction that pumps blood to the ventricles.

• The Q, R, and S waves are usually treated as a single composite waveknown as the QRS-complex. The QRS-complex reflects the depolarisa-tion of ventricular myocardium, and indicates the start of ventricularcontraction that pumps blood to the lungs and the rest of the body.

2 J. Chee and S.-C. Seow

P-wave

QRS-complex T-wave R

R-R Interval

P-R IntervalP-R Segment

q

S-T Segment

s

1 mV marker

Q-T Interval

1 mV

0.1 mV

RT

P

0.04 sec 0.2 sec

Time 0

Voltage

0

25 mm/sec

10mm

/mV

1 mm=0.04sec

1mm

=0.1m

V

U-wave

Fig. 1.1. Typical ECG of a normal person

Aorta

Left auricle

Right atrium

Rightventricle

Leftventricle

ADAM

Fig. 1.2. Human heart

J-point J-point

Fig. 1.3. In the ECG of some normal persons, the S-wave is not seen. Instead aJ-point appears in its place and is defined as the intersection of the S-T segmentwith the ending of the QRS-complex

1 The Electrocardiogram 3

Systole120

Left HeartPressure(mm Hg)

LeftVentricularVolume (ml)

Heart Sounds

Time (s)

ECG

Right HeartPressure(mm Hg)

0

120

20

0

0

0 0.2 0.4 0.6 0.8

Diastole

Aorta

L. Ventricle

L. Atrium

PulmonaryArtery

R. Ventricle

2nd1st 3rd 4th

PPR

Q S

T

Fig. 1.4. When the ventricles depolarise, the ventricular muscles contract andrapidly builds up systolic pressure. When left ventricular pressure exceeds that ofthe aorta, the aortic valve opens and blood ejects out to circulate in the body

• TheT-wavecorresponds to the repolarisationof theventricularmyocardium,which is a necessary recovery process for the myocardium to depolarise andcontract again. The end of the T-wave coincides with the end of ventricularcontraction. Atrial depolarisation (Ta-wave) is usually not visible as itnormally coincides with the QRS-complex (and is buried in the largerwaveform).

• The origin of the U-wave is uncertain, and is thought to represent repolar-isation of endocardial structures or late depolarisation of the ventricularmyocardium. U waves may be seen in a normal ECG, but are <10% ofthe height of the QRS complex. They become prominent under abnormalconditions such as electrolyte imbalance and drug toxicity.

More detailed discussion of the waves, segments and intervals in relation toarrhythmias follows in Sect. 1.3 of this chapter, after a brief study of theheart’s anatomy and its electrical conduction system.

1.1 Anatomy of the Heart

The heart is an efficient muscular organ that pumps blood throughout thewhole body. Blood brings needed nutrients and oxygen to tissue, and carries


away metabolic waste and carbon dioxide for excretion through the kidneysand the lungs, respectively.

The heart is made up of 4 chambers (Fig. 1.5). The two upper chambersare called the left and right atria or auricles, while the lower two chamb-ers are called the left and right ventricles. The atria are attached to theventricles by fibrous, non-conductive tissue that keeps the ventricles elec-trically isolated from the atria [1–4]. A thin membranous wall called theinteratrial septum separates the left atrial chamber from the right while athicker muscular wall called the interventricular septum separates the leftventricular chamber from the right.

The right atrium and the right ventricle together form a pump to the cir-culate blood to the lungs. Oxygen-poor blood is received through large veinscalled the superior and inferior vena cava and flows into the right atrium. Theright atrium contracts and forces blood into the right ventricle, stretchingthe ventricle and maximising its pumping (contraction) efficiency. The rightventricle then pumps the blood to the lungs where the blood is oxygenated.Similarly, the left atrium and the left ventricle together form a pump to

InferiorVena Cava

SuperiorVena Cava

RightAtrium

LeftAtrium

MitralValve

PulmonaryVein

PulmonaryArtery

Aorta

TricuspidValve

PulmonaryValve

RightVentricle

LeftVentricle

AorticValve

Fig. 1.5. A diagrammatic structure of the human heart


circulate oxygen-enriched blood received from the lungs (via the pulmonaryveins) to the rest of the body.

Pumping is performed with synchronised motion. The right and left atriacontract in unison to force-fill the ventricles, and then the right and leftventricles contract in unison to forcefully pump blood to the lungs and otherparts of the body, respectively. The time duration during which the ventri-cles contract is known as “systole” while the time duration during which theventricles relax to receive blood is called “diastole”.

The left ventricle typically has a muscular wall about three times as thickas that of the right ventricle (Fig. 1.6) because of the heavier workload to cir-culate blood to the rest of the body, as compared to that required to circulateblood to the lungs.

The muscle wall of the heart is made up of three layers. The inner layer,called the endocardium, lines the chambers of the heart. The centre layer isthe myocardium, which forms the bulk and provides the contractile force forpumping. This layer of myocardium is further divided into the subendocardialarea which is the inner half of the myocardium, and the subepicardial area, theouter half. The outermost layer of the heart wall overlying the myocardiumis called the epicardium. The entire heart is encased in a thin membranecalled the serous pericardium, which is made up of 2 layers viz the visceral(inner) and parietal (outer) pericardium. Pericardial fluid between these 2layers minimises friction against heart movements as the heart beats.

Superiorvena cava

Aortic valve

Right pulmonaryveins

Right atrium

Tricuspid valve

Inferior vena cava

Right ventricle

Mitral valve

Left pulmonaryveins

Left atrium

Pulmonary artery

Aorta

Left ventricle

Fig. 1.6. Cutaway section of the human heart showing the myocardium


The pericardium is the thin sacenclosing the heart

Pericardium

Fig. 1.7. The pericardium anchors the heart within the body

External to the serous pericardium is an outer tough, fibrous sac calledthe fibrous pericardium which holds the heart in place within the body(Fig. 1.7). Ligaments attach the fibrous pericardium inferiorly to the centre ofthe diaphragm, anteriorly to the sternum, and posteriorly to the oesophagus,trachea, and main bronchi.

Valves regulate the flow of blood by ensuring that blood flows only in thedesired direction, and prevent backflow. If the valves are incompetent (i.e.“leaky”) or stenotic (i.e. narrowed), the heart has to work harder. This addi-tional workload can, over time, cause the heart to compensate by developinga thicker myocardium (hypertrophy); and in severe cases eventually lead toheart failure.

1.2 Electrical Conduction System

Pumping is only efficient when the heart contracts in a coordinated manner.Blood must first fill the atria, and then be pumped into the ventricles beforebeing forcefully ejected. This coordination is achieved by an elaborate elec-trical conduction system that controls the precise timing for depolarising thesubstantial mass of electrically excitable myocardium. This delicate controlstarts with an intrinsic self-excitable cardiac pacemaker which sets the rateat which the heart beats. The pacemaker spontaneously generates regularelectrical impulses, which then spread through the conduction system of theheart and initiate contraction of the myocardium. This pacemaker is calledthe Sino-atrial (S-A) node.

1.2.1 The Sino-Atrial (S-A) Node

The S-A node lies in the upper wall of the right atrium, near the entrance ofthe superior vena cava (Fig. 1.8). It is normally the initial source of electrical


Intrinsic conduction system of the heart

Sinoatrial node

Right atrium

Right ventricle

Left ventricle

Left atrium

Atrioventricularnode

Fig. 1.8. The intrinsic conduction system

excitation. The S-A node is a network of pacemaker cells – excitable tissuethat exhibits automaticity. Automaticity is a property of the cell to periodic-ally generate an electrical impulse even without the presence of an externalstimulus. Automaticity of the S-A node may be modulated by the balanceof sympathetic and parasympathetic inputs, or by drugs. Sympathetic stim-ulation (from nerves connected to the brain) speeds up the rate of impulsegeneration while parasympathetic stimulation slows the rate. This variabilityin rate allows the heart to respond to demands for higher or lower cardiacoutput (i.e. faster or slower heart rate).

The electrical impulse from the S-A node spreads through the myocardiumof the right atrium, stimulating its contraction (Fig. 1.9). At the same time,the interatrial conduction tract (Bachman’s Bundle) between the S-A nodeand the left atrium carries the impulse quickly to the left atrium, spreading itthrough the left atrial myocardium so that the contraction of the left atriumoccurs almost simultaneously with that of the right. In addition, three intern-odal conduction tracts carry the impulse from the S-A node to Atrioventricular(A-V) node, the gateway to the ventricular conduction system.

1.2.2 Depolarisation

Propagation of an electrical impulse through excitable tissue is achievedthrough a process called depolarisation. At rest, a potential difference (volt-age) exists across the cell membrane, between the fluid external to the celland its internal cellular fluid as the cell membrane separates the ions of bothfluids. When a trigger occurs (e.g. an electrical impulse exceeding a threshold


NORMAL HEARTBEAT

A normal heartbeat is triggered byan electrical impulse which startsin the Sinoatrial (SA) Node.

The impulse then travels acrossthe Atrioventricular (AV) Nodeand triggers the ventriclesto contract.

Fig. 1.9. Sino-atrial node in the right atrium connects electrically via conductiontracts to the atrioventricular node

voltage) the cell membrane suddenly becomes permeable. The ions that werepreviously held at bay across the membrane cross the membrane generatingan ionic current flow.

The depolarisation of a myocardial cell will lead to a physical contractionwithin milliseconds. The contraction would then last for tens of milliseconds.At the end of depolarisation, the cell membrane once again becomes imper-meable. Repolarisation begins as the ion channels within the cell membranepump out unwanted ions and brings the ionic balance of the cell back to itsresting (equilibrium) state.

Myocardial cells that make up the heart muscles are interconnected toadjacent myocardial cells through specialised cellular membranes known asintercalated disks (Fig. 1.10). These disks contain areas of low electricalresistance called gap junctions, which permit rapid conduction of electricalimpulses from one cell to another [5].

With its large muscle mass, depolarisation of the heart muscles collec-tively generates a strong ionic current. This current flows through the resistivebody tissue generating a voltage drop. The magnitude of the voltage drop issufficiently large to be detected by electrodes attached to the skin (typically3 mV-peak at the chest). ECGs are thus recordings of voltage drops across theskin caused by ionic current flow generated from myocardial depolarisations.Although nerve depolarisations also generate ionic currents, the magnitudesof such currents are too small to be detected by skin electrodes.


Intercalateddisk

Fig. 1.10. Myocardial cells are interconnected to one another electrically andphysically by the intercalated disks

Atrial depolarisation results in the spreading of the electrical impulsethrough the atrial myocardium and appears as the P-wave. This wave is nor-mally less than 120 ms wide and corresponds with the start of atrial muscularcontraction. The P-R interval, which is measured from the onset of the P-waveto the onset of the QRS-complex, is normally within 120–200 ms. (Note thatif the Q-wave is present, the P-R interval should terminate on the onset ofQ-wave although it would still be labelled as P-R interval.) Atrial contractiontypically lasts longer than the P-R interval.

Similarly, ventricular depolarisation results in the spreading of the elec-trical impulse throughout the ventricular myocardium. Depolarisation istriggered when the pacemaker impulse from the S-A node comes through theatrioventricular node [1–4] and spreads through the ventricular conductionsystem to the myocardium.

1.2.3 The Atrioventricular (A-V) Node

The A-V node lies partly in the right side of the interatrial septum and partlyin the interventricular septum. Since the ventricles are separated from theatria by a fibrous layer of non-conducting tissue, no electrical impulse fromthe S-A node will reach the ventricles except through the A-V node (Fig. 1.10).This allows the A-V node to control the impulse received from the S-A node.Each impulse, received through the atrial conduction tracts, is propagated ata slow rate (∼0.05 m/s) through the A-V node, generating a time delay. Thistime delay is to allow blood from the atria to fully fill the ventricles beforethe latter contract.


Left atriumSinoatrial node

Right atrium

Right ventricle

Atrioventricularnode

Left ventricle

Fig. 1.11. The ventricular conduction system

The conduction system of the ventricles is more elaborate than that of theatria. The A-V node connects to the Bundle of His which in turn connectsto two bundle branches – the Right Bundle Branch and the Left BundleBranch (Fig. 1.11). The right bundle branch travels down the right side ofthe interventricular septum to subdivide again and again before connectingto the Purkinje Network of the right ventricle. The Purkinje network of fineconducting fibres is embedded in subendocardium of the ventricles. In likemanner, the left bundle branch divides into the left posterior fascicle and leftanterior fascicle which travel down the left side of the interventricular septumto further subdivide repeatedly before connecting with the Purkinje Networkof the left ventricle. In addition, the left anterior fascicle also branches to formseptal fibres which innervate the interventricular septum.

Thus from the A-V node, the impulse propagates through the bundle ofHis, spreads through the bundle branches and sub-branches until it reachesthe Purkinje network. The spread of impulses to depolarise the ventricularmyocardium is nearly simultaneous despite the large mass of myocardiumbecause of the high speed of propagation through the conduction fibres.

1.2.4 Automaticity

Cells that form different parts of the conduction system exhibit different ratesof automaticity. The S-A node has the fastest inherent rate of 60 ∼ 100pulses per minute. It becomes the dominant pacemaker in the heart, settingthe rhythm at which the heart contracts by overriding (resetting) potentialpacemakers running at slower rates. The A-V node and the Bundle of His arethe next fastest with an inherent rate of 40 ∼ 60 pulses per minute. Finally theBundle Branches and Purkinje Network exhibit the slowest rates, at 30 ∼ 40pulses per minute.

Date post:	26-Sep-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Advances in Cardiac Signal...

Documents