Advanced Methods and Tools for ECG Data Analaysis

The electrocardigram (ECG) is a recording of the electrical activity of the heart that is used to diagnose heart disorders. In recent years, new state-of-the-art approaches to ECG analysis have emerged that are now of significant interest to biomedical and electrical engineers, as well as healthcare professionals. This practical book is the first one-stop resource to offer a thorough, up-to-date treatment of the techniques and methods used in ECG data analysis, from fundamental principles to the latest tools in the field. The book places emphasis on the selection, modeling, classification, and interpretation of data based on advanced signal processing and artificial intelligence techniques. Professionals find guidance on designing, implementing, and evaluating software systems used for the analysis of ECG and related data. Moreover, this comprehensive resource offers a solid grounding in the relevant basics of physiology, data acquisition and database design, and addresses the practical issues of improving existing data analysis methods and developing new applications. Engineering in Medicine & Biology Series 1580539661 2 Chapter 1 The Physiological Basis of the Electrocardiogram 17 1.1 Cellular Processes That Underlie the ECG 17 1.2 The Physical Basis of Electrocardiography 20 1.3 Introduction to Clinical Electrocardiography: Abnormal Patterns 28 1.4 Summary 40 References 40 Selected Bibliography 41 Chapter 2 ECG Acquisition, Storage, Transmission, and Representation 43 2.1 Introduction 43 2.2 Initial Design Considerations 44 2.3 Choice of Data Libraries 51 2.4 Database Analysis--An Example Using WFDB 53 2.5 ECG Acquisition Hardware 57 2.6 Summary 66 References 66 Chapter 3 ECG Statistics, Noise, Artifacts, and Missing Data 71 3.1 Introduction 71 3.2 Spectral and Cross-Spectral Analysis of the ECG 71 3.3 Standard Clinical ECG Features 76 3.4 Nonstationarities in the ECG 80 3.5 Arrhythmia Detection 83 3.6 Noise and Artifact in the ECG 85 3.7 Heart Rate Variability 87 3.8 Dealing with Nonstationarities 99 3.9 Summary 108 References 109 Chapter 4 Models for ECG and RR Interval Processes 117 4.1 Introduction 117 4.2 RR Interval Models 118 4.3 ECG Models 131 4.4 Conclusion 142 References 143 Chapter 5 Linear Filtering Methods 151 5.1 Introduction 151 5.2 Wiener Filtering 152 5.3 Wavelet Filtering 156 5.4 Data-Determined Basis Functions 164 5.5 Summary and Conclusions 183 References 183 Chapter 6 Nonlinear Filtering Techniques 187 6.1 Introduction 187 6.2 Nonlinear Signal Processing 188 6.3 Evaluation Metrics 194 6.4 Empirical Nonlinear Filtering 195 6.5 Model-Based Filtering 202 6.6 Conclusion 209 References 210 Chapter 7 The Pathophysiology Guided Assessment of T-Wave Alternans 213 7.1 Introduction 213 7.2 Phenomenology of T-Wave Alternans 213 7.3 Pathophysiology of T-Wave Alternans 213 7.4 Measurable Indices of ECG T-Wave Alternans 215 7.5 Measurement Techniques 217 7.6 Tailoring Analysis of TWA to Its Pathophysiology 223 7.7 Conclusions 227 Acknowledgments 227 References 227 Chapter 8 ECG-Derived Respiratory Frequency Estimation 231 8.1 Introduction 231 8.2 EDR Algorithms Based on Beat Morphology 234 8.3 EDR Algorithms Based on HR Information 244 8.4 EDR Algorithms Based on Both Beat Morphology and HR 245 8.5 Estimation of the Respiratory Frequency 246 8.6 Evaluation 252 8.7 Conclusions 256 References 257 Appendix 8A Vectorcardiogram Synthesis from the 12-Lead ECG 259 Chapter 9 Introduction to Feature Extraction 261 9.1 Overview of Feature Extraction Phases 261 9.2 Preprocessing 264 9.3 Derivation of Diagnostic and Morphologic Feature Vectors 267 9.4 Shape Representation in Terms of Feature-Vector Time Series 276 References 279 Appendix 9A Description of the Karhunen-Lo`eve Transform 280 Chapter 10 ST Analysis 285 10.1 ST Segment Analysis: Perspectives and Goals 285 10.2 Overview of ST Segment Analysis Approaches 286 10.3 Detection of Transient ST Change Episodes 288 10.4 Performance Evaluation of ST Analyzers 294 References 303 Chapter 11 Probabilistic Approaches to ECG Segmentation and Feature Extraction 307 11.1 Introduction 307 11.2 The Electrocardiogram 308 11.3 Automated ECG Interval Analysis 309 11.4 The Probabilistic Modeling Approach 310 11.5 Data Collection 312 11.6 Introduction to Hidden Markov Modeling 312 11.7 Hidden Markov Models for ECG Segmentation 320 11.8 Wavelet Encoding of the ECG 327 11.9 Duration Modeling for Robust Segmentations 328 11.10 Conclusions 332 References 332 Chapter 12 Supervised Learning Methods for ECG Classification/Neural Networks and SVM Approaches 335 12.1 Introduction 335 12.2 Generation of Features 336 12.3 Supervised Neural Classifiers 340 12.4 Integration of Multiple Classifiers 346 12.5 Results of Numerical Experiments 347 12.6 Conclusions 352 Acknowledgments 352 References 352 Chapter 13 An Introduction to Unsupervised Learning for ECG Classification 355 13.1 Introduction 355 13.2 Basic Concepts and Methodologies 355 13.3 Unsupervised Learning Techniques and Their Applications in ECG Classification 357 13.4 GSOM-Based Approaches to ECG Cluster Discovery and Visualization 368 13.5 Final Remarks 375 References 378 About the Authors 383 Index 387 1580539661 1580539661......Page 2 1.1 Cellular Processes That Underlie the ECG......Page 17 1.2 The Physical Basis of Electrocardiography......Page 20 1.3 Introduction to Clinical Electrocardiography: Abnormal Patterns......Page 28 References......Page 40 Selected Bibliography......Page 41 2.1 Introduction......Page 43 2.2 Initial Design Considerations......Page 44 2.3 Choice of Data Libraries......Page 51 2.4 Database Analysis--An Example Using WFDB......Page 53 2.5 ECG Acquisition Hardware......Page 57 References......Page 66 3.2 Spectral and Cross-Spectral Analysis of the ECG......Page 71 3.3 Standard Clinical ECG Features......Page 76 3.4 Nonstationarities in the ECG......Page 80 3.5 Arrhythmia Detection......Page 83 3.6 Noise and Artifact in the ECG......Page 85 3.7 Heart Rate Variability......Page 87 3.8 Dealing with Nonstationarities......Page 99 3.9 Summary......Page 108 References......Page 109 4.1 Introduction......Page 117 4.2 RR Interval Models......Page 118 4.3 ECG Models......Page 131 4.4 Conclusion......Page 142 References......Page 143 5.1 Introduction......Page 151 5.2 Wiener Filtering......Page 152 5.3 Wavelet Filtering......Page 156 5.4 Data-Determined Basis Functions......Page 164 References......Page 183 6.1 Introduction......Page 187 6.2 Nonlinear Signal Processing......Page 188 6.3 Evaluation Metrics......Page 194 6.4 Empirical Nonlinear Filtering......Page 195 6.5 Model-Based Filtering......Page 202 6.6 Conclusion......Page 209 References......Page 210 7.3 Pathophysiology of T-Wave Alternans......Page 213 7.4 Measurable Indices of ECG T-Wave Alternans......Page 215 7.5 Measurement Techniques......Page 217 7.6 Tailoring Analysis of TWA to Its Pathophysiology......Page 223 References......Page 227 8.1 Introduction......Page 231 8.2 EDR Algorithms Based on Beat Morphology......Page 234 8.3 EDR Algorithms Based on HR Information......Page 244 8.4 EDR Algorithms Based on Both Beat Morphology and HR......Page 245 8.5 Estimation of the Respiratory Frequency......Page 246 8.6 Evaluation......Page 252 8.7 Conclusions......Page 256 References......Page 257 Appendix 8A Vectorcardiogram Synthesis from the 12-Lead ECG......Page 259 9.1 Overview of Feature Extraction Phases......Page 261 9.2 Preprocessing......Page 264 9.3 Derivation of Diagnostic and Morphologic Feature Vectors......Page 267 9.4 Shape Representation in Terms of Feature-Vector Time Series......Page 276 References......Page 279 Appendix 9A Description of the Karhunen-Lo`eve Transform......Page 280 10.1 ST Segment Analysis: Perspectives and Goals......Page 285 10.2 Overview of ST Segment Analysis Approaches......Page 286 10.3 Detection of Transient ST Change Episodes......Page 288 10.4 Performance Evaluation of ST Analyzers......Page 294 References......Page 303 11.1 Introduction......Page 307 11.2 The Electrocardiogram......Page 308 11.3 Automated ECG Interval Analysis......Page 309 11.4 The Probabilistic Modeling Approach......Page 310 11.6 Introduction to Hidden Markov Modeling......Page 312 11.7 Hidden Markov Models for ECG Segmentation......Page 320 11.8 Wavelet Encoding of the ECG......Page 327 11.9 Duration Modeling for Robust Segmentations......Page 328 References......Page 332 12.1 Introduction......Page 335 12.2 Generation of Features......Page 336 12.3 Supervised Neural Classifiers......Page 340 12.4 Integration of Multiple Classifiers......Page 346 12.5 Results of Numerical Experiments......Page 347 References......Page 352 13.2 Basic Concepts and Methodologies......Page 355 13.3 Unsupervised Learning Techniques and Their Applications in ECG Classification......Page 357 13.4 GSOM-Based Approaches to ECG Cluster Discovery and Visualization......Page 368 13.5 Final Remarks......Page 375 References......Page 378 About the Authors......Page 383 Index......Page 387 "This resource provides a practical and theoretical understanding of state-of-the-art techniques for electrocardiogram (ECG) data analysis. Placing an emphasis on the fundamentals of signal etiology, acquisition, data selection, and testing, this comprehensive volume presents guidelines to help practitioners design, implement, and evaluate algorithms used for the analysis of ECG and related data. Additionally, explanations of open source software and related databases for signal processing and system testing are included. Including over 190 illustrations, this book offers a solid grounding in the relevant basics of physiology, data acquisition, and database design, and addresses the practical issues of improving existing data analysis methods and developing new applications."--Jacket Here's a cutting-edge, practical book that offers you a thorough understanding of state-of-the-art techniques for electrocardiogram (ECG) data analysis. Placing emphasis on the selection, modeling, classification, and interpretation of data based on advanced signal processing and artificial intelligence techniques, the book helps you design, implement, and evaluate software systems used for the analysis of ECG and related data. Offers you an understanding of techniques for electrocardiogram (ECG) data analysis. Placing emphasis on the selection, modeling, classification, and interpretation of data based on advanced signal processing and artificial intelligence techniques, this book helps you design, implement, and evaluate software systems used for the analysis of ECG.