Biomedical Signal Analysis

Citation

Rangayyan, R. M., & Krishnan, S. (2024). Biomedical Signal Analysis (3rd ed.). Wiley-IEEE Press.

Chapter Summary

The third edition has ten chapters and about 720 pages, with forewords by Willis Tompkins and Alan Oppenheim, a list of symbols and abbreviations, and a companion website. The publisher describes the approach shared by all editions: each chapter begins with a statement of a biomedical signal problem, presents real-life case studies with the signals concerned, introduces techniques from simple "textbook" methods to more sophisticated research-informed ones, and ends with practical applications. Compared with the second edition, which had nine chapters, the third adds essential background and advanced topics, and a second author, Sridhar Krishnan. Chapters end with study questions and laboratory exercises or projects, and an online repository supplies solutions and data files.

Chapter 1: Introduction to Biomedical Signals

Sets out the nature of biomedical signals and surveys seventeen examples: the action potentials of cardiac myocytes and neurons, the electroneurogram, electromyogram, electrocardiogram and electroencephalogram, event-related potentials, the electrogastrogram, the phonocardiogram, the carotid pulse, the photoplethysmogram, catheter-tip signals, the speech signal, vibroarthrogram and vibromyogram signals, otoacoustic emissions and bioacoustic signals. It then states the objectives and challenges of signal analysis and the case for computer-aided monitoring and diagnosis.

Chapter 2: Analysis of Concurrent, Coupled, and Correlated Processes

Case studies of signals recorded together, such as the ECG with the PCG, the carotid pulse and the atrial electrogram, cardiorespiratory interaction, heart-rate variability and knee-joint and muscle-vibration signals. Applications are the segmentation of the PCG and the diagnosis and monitoring of sleep apnea, including polysomnography, home monitoring and multivariate, multiorgan analysis.

Chapter 3: Filtering for Removal of Artifacts

The longest chapter on technique. It separates random, structured and physiological noise and stationary, nonstationary and cyclostationary processes, reviews the fundamentals of filtering (convolution, transform-domain analysis, pole–zero plots, the Fourier and discrete Fourier transforms), and then covers synchronized averaging, time-domain filters, Butterworth and notch and comb filters, order-statistic filters, the Wiener filter and adaptive filters. Applications include artifacts in event-related potentials and the ECG, maternal–fetal ECG and muscle-contraction interference.

Chapter 4: Detection of Events

Covers detecting P, QRS and T waves and heart sounds, correlation analysis of EEG rhythms, cross-spectral techniques, the matched filter and homomorphic filtering. Applications are ECG rhythm analysis, identifying heart sounds and detecting the aortic component of the second heart sound.

Chapter 5: Analysis of Waveshape and Waveform Complexity

Moves from finding events to describing their shape: analysis of event-related potentials, morphology of ECG waves, envelope extraction and analysis of activity. Applications include normal and ectopic beats, exercise ECG, EMG in relation to force, respiration, statistical analysis of VAG signals and fractal analysis of the EMG.

Chapter 6: Frequency-domain Characterization of Signals and Systems

Treats estimation of the power spectral density and measures derived from it, with applications to prosthetic heart valves, fractal analysis of VAG signals and spectral analysis of EEG.

Chapter 7: Modeling of Biomedical Signal-generating Processes and Systems

Covers point processes, parametric system modeling, autoregressive and pole–zero modeling, electromechanical models of signal generation and electrophysiological models of the heart. Applications are heart-rate variability, spectral modeling of PCG signals and coronary artery disease.

Chapter 8: Adaptive Analysis of Nonstationary Signals

Time-variant systems, fixed and adaptive segmentation, adaptive filters for segmentation, the Kalman filter, wavelet analysis and bilinear time–frequency distributions. The applications are broad: EEG and PCG segmentation, heart-rate variability, infant cry sounds, PPG denoising, CPR studies, ventricular fibrillation, epileptic seizures and neural decoding for prosthesis control.

Chapter 9: Signal Analysis via Adaptive Decomposition

Matching pursuit, empirical mode decomposition, dictionary learning, decomposition-based adaptive time–frequency distributions and separation of mixtures of signals. Applications include seizure detection, VAG signals, T-wave alternans, extraction of the fetal ECG from a single channel and EEG analysis for brain–computer interfaces.

Chapter 10: Computer-aided Diagnosis and Healthcare

Pattern classification (supervised and unsupervised), probabilistic models and statistical decision, logistic regression, neural networks, diagnostic accuracy and cost, and the reliability of features and classifiers. Applications are ectopic ECG beats, knee-joint cartilage pathology, sleep apnea and Parkinson's disease monitoring, and the chapter ends with the strengths and limitations of computer-aided diagnosis.

Key Concepts

Chapters 1–3

• Biomedical signal characteristics: the origin and typical form of signals such as the ECG, EMG and EEG.
• Concurrent, coupled and correlated processes: signals from interacting physiological systems analysed together.
• Random, structured and physiological noise: the categories of interference that filtering must address.
• Synchronized averaging, Wiener and adaptive filtering: the main filtering approaches for artifacts.

Chapters 4–7

• Event detection and the matched filter: finding waves and sounds in a recording.
• Waveshape and complexity: envelope, activity and fractal measures.
• Power spectral density estimation: spectral description of signals.
• Autoregressive and pole–zero models: parametric models of signal generation.

Chapters 8–10

• Adaptive segmentation and wavelet analysis: methods for signals whose properties change.
• Matching pursuit, empirical mode decomposition and dictionary learning: decomposition-based analysis of multicomponent signals.
• Pattern classification and diagnostic accuracy: moving from features to computer-aided decisions.

Critical Analysis

Strengths

• Problem-first structure. Each chapter's pattern of problem statement, case studies, methods and applications ties techniques to concrete signals, and the section lists show a consistent layout across the ten chapters.
• Methods from simple to advanced. The publisher describes presenting basic textbook methods before research-informed ones, and the chapter contents bear this out: Chapter 3 moves from moving-average filters to adaptive filters, and Chapters 8 and 9 reach wavelets, empirical mode decomposition and dictionary learning.
• Wide range of signals and applications. The applications span cardiac, neural, muscular, respiratory and joint signals, from ECG rhythm analysis to brain–computer interfaces.
• Practice material. Every chapter ends with study questions and laboratory exercises, and the online repository supplies data files and solutions.

Limitations

• Mathematical and signal-processing preparation. The book states that introductory sections on signals, systems and transforms make it accessible beyond electrical engineering, but Chapters 3, 7 and 8 are demanding for readers without that background.
• Case studies drawn from the authors' research. The publisher notes that the case studies come from the authors' own work, and some recur across chapters: vibroarthrogram (VAG) signals, for example, appear in Chapters 1, 3, 5, 6 and 9.
• Edition differences. The third edition has ten chapters against nine in the second, so chapter numbers, exercises and page references differ from older course materials.

Where It Fits in the Course

Rangayyan is the case-study-driven, graduate-level text in this course. Semmlow's Biosignal and Medical Image Processing is written specifically for biomedical engineers and covers filtering and spectral and time-frequency analysis alongside image processing. Tompkins's Biomedical Digital Signal Processing (1993) is a laboratory-oriented introduction centred on the ECG. Circuits, Signals, and Systems for Bioengineers teaches the electrical engineering foundations with MATLAB, and Smith's The Scientist and Engineer's Guide to Digital Signal Processing is a plain-language introduction to DSP.

Read one of the last two first if filtering, convolution and the Fourier transform are new. Rangayyan then supplies the biomedical depth: the signals themselves, the artifacts that affect them and the route from analysis to diagnosis.

How to Use This Book

The book is for senior undergraduate and graduate students in biomedical and electrical engineering, and for researchers who need methods for physiological signals. Familiarity with signals, systems and transforms is the main prerequisite, and the opening sections give a refresher.

A practical plan:

• Core: Chapters 1 to 4 for the signals, noise and detection.
• Next: Chapters 5 to 7 for waveshape, spectral analysis and modeling.
• Advanced: Chapters 8 and 9 on nonstationary signals and decomposition, and Chapter 10 for classification and diagnosis; Chapter 10 can be read earlier if the goal is computer-aided diagnosis.
• Practice: work the study questions and use the laboratory exercises and online data files to reproduce results on real recordings.