Pattern Recognition Techniques Applied to Biomedical Problems

Preface
Contents
Contributors
The Classification of Independent Components for Biomedical Signal Denoising: Two Case Studies
	1 Introduction
	2 Independent Component Analysis
		2.1 FastICA
		2.2 Infomax and Ext-Infomax
		2.3 TDSep
	3 Case Study I: Denoising the Abdominal Phonogram
		3.1 The Problem Definition
		3.2 A Dataset of Single-Channel Abdominal Phonograms
		3.3 Single-Channel ICA (SCICA)
			3.3.1 Mapping a Single-Channel Signal into a Multidimensional Representation
			3.3.2 Extraction of Multiple Independent Components
			3.3.3 Automatic Classification of Fetal Independent Components
		3.4 Results
		3.5 Discussion and Conclusions
	4 Case Study II: Denoising the EEG for Recovering the Late Latency Auditory Evoked Potentials (LLAEP)
		4.1 The Problem Definition
		4.2 A Multichannel Dataset of Long Latency Auditory Evoked Potentials
		4.3 Multichannel ICA
			4.3.1 ICA Algorithms: High-Order Statistic (HOS) Based Versus Second-Order Statistic (SOS) Based
			4.3.2 Optimal ICA Parameters for the Estimation of the LLAEP Components
			4.3.3 Objective IC Selection of AEPs in Children with Cochlear Implants Using Clustering
		4.4 Results
			4.4.1 Optimal Clustering LLAEPs in Subjects with CI
			4.4.2 Maturation of the Auditory System in Children with Cochlear Implants
		4.5 Discussion and Conclusions
	5 Final Remarks
	References
Pattern Recognition Applied to the Analysis of Genomic Dataand Its Association to Diseases
	1 Relevance of Applying Pattern Recognition Techniques to the Analysis of Genomic Data
	2 Introduction to Genomic Analysis by High-Throughput Sequencing Technologies
		2.1 Genomes and Genes
		2.2 Microarrays
		2.3 DNA-RNA Sequencing
			2.3.1 Sequencing by Ligation
			2.3.2 Sequencing by Synthesis
			2.3.3 Real-Time Sequencing of a Single Molecule
			2.3.4 Challenges
	3 The Process of the Identification of Genomic Variants: Variant Calling
		3.1 Introduction
		3.2 The Importance of Genomic Variants
		3.3 Algorithms Based on Alignments
			3.3.1 Assignment of Origin Position Based on Pattern Recognition
			3.3.2 Duplicate Removal
			3.3.3 Realignment Around Insertions and Deletions
			3.3.4 Quality Score Recalibration Using Sequence Patterns
			3.3.5 Genotype Assignment
		3.4 Algorithms Not Based on Alignments
	4 Genetic Expression Profile for Diagnostic of Diseases
		4.1 Background of Genetic Profile Construction
		4.2 Methodologies for Genetic Profile Construction
		4.3 Considerations for Clustering Profiles
		4.4 Scope and Limitation
	5 Gene-Disease Association Extraction from Biomedical Literature
		5.1 Background
			5.1.1 Biomedical Natural Language Processing
			5.1.2 Information Extraction
		5.2 Automatic Extraction of Gene-Disease Associations
			5.2.1 Gene-Disease Association Databases
			5.2.2 Co-occurrence Pattern
			5.2.3 Rule-Based Approaches
			5.2.4 Machine Learning Approaches
			5.2.5 Current Challenges
	6 Conclusion
	References
Images Analysis Method for the Detection of Chagas Parasitein Blood Image
	1 Introduction
	2 Materials and Methods
		2.1 Automated System for Positioning and Image Acquisition
		2.2 Images Analysis and Processing
		2.3 Image Pre-processing
		2.4 Segmentation
		2.5 Parasite Detection
	3 Results
		3.1 Conclusions
	References
Monitoring and Evaluating Public Health Interventions
	1 Health Technology Assessment
	2 Public Health Interventions
	3 Intervention Analysis on Time Series
		3.1 Box and Tiao Approach for Intervention Analysis
	4 Case Study: Assessing an Action of the Brazilian Breast Cancer Screening Program
		4.1 Methods
			4.1.1 Pre-intervention Analysis
			4.1.2 Analysis of the Intervention\'s Effect
		4.2 Results
		4.3 Discussion
		4.4 Conclusion
	References
Recognition of Nausea Patterns by MultichannelElectrogastrography
	1 Introduction
	2 EGG Dataset
	3 Nausea Discrimination Method
		3.1 Preprocessing Step
		3.2 Independent Component Analysis (ICA)
		3.3 Feature Extraction
		3.4 Feature Selection
		3.5 Classification
		3.6 Performance Evaluation
	4 Case Study and Results
		4.1 Results for the Feature Selection Process
		4.2 Results for the Classification
	5 Conclusions
	References
Random Forest Algorithm for Prediction of HIV Drug Resistance
	1 Introduction
	2 Algorithmic Framework for Random Forest
		2.1 Bootstrap Samples
		2.2 Construction of Trees and Sampling of Variables
		2.3 Decision
	3 Error Estimation
	4 Variable Importance Measures
		4.1 Impurity Importance
		4.2 Permutation Importance
	5 Proximities
	6 Application on HIV-1 Drug Resistance Problem
		6.1 Background
	7 Conclusion
	References
Analysis of Cardiac Contraction Patterns
	1 Cardiac Contraction Dynamics
		1.1 Anatomical and Physiological Basis of Cardiac Contraction
			1.1.1 Electric Conduction System
			1.1.2 Mechanical Contraction
		1.2 Methods for Evaluation of Ventricular Dynamics
			1.2.1 Electrocardiogram
			1.2.2 Imaging Modalities
		1.3 Heart Failure and Cardiac Resynchronization Therapy
	2 Cardiac Imaging for the Assessment of Contraction Patterns
		2.1 Echocardiography
			2.1.1 Conventional M-Mode
			2.1.2 Tissue Doppler Imaging
			2.1.3 Tissue Synchronization Imaging
			2.1.4 Speckle Tracking
		2.2 Cardiac Radionuclide Imaging
			2.2.1 Equilibrium Radionuclide Angiocardiography
			2.2.2 Gated Myocardial Perfusion Single-Photon Emission Computed Tomography
		2.3 Cardiac Magnetic Resonance
	3 Analysis of Cardiac Contraction Dyssynchrony by ERNA
		3.1 Data Acquisition
		3.2 Fourier Phase Analysis
		3.3 Factor Analysis of Dynamic Structures
		3.4 Classification of Severity in Cardiac Contraction Dyssynchrony
			3.4.1 Study Population
			3.4.2 Supervised Classification
	4 Perspectives
	References
Pattern Recognition to Automate Chronic Patients Follow-Up and to Assist Outpatient Diagnostics
	1 Introduction
	2 Information and Communication Technologies in Daily Life
	3 Personal Lifetime Analyzed by Pattern Recognition
	4 Adoption of SIMIC by Patients and Health Staff
	5 Electronic Medical Records and Medical Reasoning
	6 PRAXIS Structure Follows Medical Reasoning
	7 Representing the Knowledge Base
	8 Units of Thought
	9 Interoperability
	10 Medical Records and Medical Knowledge Are Separate Entities
	11 Multiple Case Types
	12 No Diagnostics Reached and Partial Case Types
	13 Chronic Condition and Chronic Case Type
	14 Consultation Assistants
	15 Conclusion
	References
Pattern Recognition for Supporting the Replacement of Medical Equipment at Mexican Institute of Pediatrics
	1 Introduction
	2 Technological Problem
	3 Methodology
		3.1 Multi-criteria Decision Analysis
	4 Mathematical Model
		4.1 Mathematical Model
		4.2 Variables
			4.2.1 Technical Variables
			4.2.2 Clinical Variables
			4.2.3 Economic Variables
		4.3 Partial Indicators
			4.3.1 Technical Indicator (IT)
			4.3.2 Clinical Indicator (IC)
			4.3.3 Economic Indicator (IE)
		4.4 Indicator for Medical Equipment Replacement
		4.5 Qualitative Scale for the Indicator for Medical Equipment Replacement
	5 Application of the Indicator for Medical Equipment Replacement (IMER)
		5.1 Application of the Technical Indicator (IT)
		5.2 Application of the Clinical Indicator (IC)
		5.3 Application of the Economic Indicator (IE)
		5.4 Application of the Indicator for Medical Equipment Replacement (IMER)
	6 Results
		6.1 Vital Signs Monitors (VSM)
		6.2 Ventilators (V)
		6.3 Radiant Heat Cradle (RHC)
		6.4 Incubators
		6.5 Electrocardiograph (ECG)
		6.6 Ultrasound Equipment
	7 Discussion
	8 Conclusions
	References
Index