Statistical Signal Processing for Neuroscience and Neurotechnology

This is a uniquely comprehensive reference that summarizes the state of the art of signal processing theory and techniques for solving emerging problems in neuroscience, and which clearly presents new theory, algorithms, software and hardware tools that are specifically tailored to the nature of the neurobiological environment. It gives a broad overview of the basic principles, theories and methods in statistical signal processing for basic and applied neuroscience problems.Written by experts in the field, the book is an ideal reference for researchers working in the field of neural engineering, neural interface, computational neuroscience, neuroinformatics, neuropsychology and neural physiology. By giving a broad overview of the basic principles, theories and methods, it is also an ideal introduction to statistical signal processing in neuroscience. A comprehensive overview of the specific problems in neuroscience that require application of existing and development of new theory, techniques, and technology by the signal processing community Contains state-of-the-art signal processing, information theory, and machine learning algorithms and techniques for neuroscience research Presents quantitative and information-driven science that has been, or can be, applied to basic and translational neuroscience problems Statistical Signal Processing for Neuroscience and Neurotechnology, (2010) i-ii. doi:10.1016/B978-0-12-375027-3.00014-4 Statistical Signal Processing for Neuroscience and Neurotechnology 2 Copyright 3 Preface 4 About the Editor 6 About the Contributors 7 Chapter 2 7 Chapter 3 7 Chapter 4 7 Chapter 5 8 Chapter 6 9 Chapter 7 10 Chapter 8 11 Chapter 9 12 Chapter 10 13 Chapter 11 13 Introduction 15 Background 15 Motivation 17 Overview and Roadmap 19 References 26 Detection and Classification of Extracellular Action Potential Recordings 28 Introduction 28 Spike Detection 29 Known Spike 32 Unknown Spike 37 Practical Limitations 39 Transform-Based Methods 41 Theoretical Foundation 41 Performance 43 Spike Sorting 45 Pattern Recognition Approach 45 Blind Source Separation Approach 50 Fewer Sources Than Mixtures 51 More Sources Than Mixtures 55 Practical Implementation 71 Discussion and Future Directions 78 Acknowledgment 82 References 82 Information-Theoretic Analysis of Neural Data 88 Introduction 88 The Encoder 90 The Channel 96 Rate-Distortion Theory 103 Post–Shannon Information Theory 107 Discussion 112 References 113 Identification of Nonlinear Dynamics in Neural Population Activity 115 Introduction 115 Problem Statement 116 Nonlinear Model of Neural Population Dynamics 117 Model Configuration 118 Model Estimation 121 Model Selection 123 Kernel Reconstruction and Interpretation 124 Model Validation and Prediction 125 Results: Application to Hippocampal CA3-CA1 Population Activity 127 Behavioral Task 127 Data Preprocessing 128 CA3-CA1 MIMO Model 130 Discussion 133 Future Directions 137 Acknowledgments 138 References 138 Graphical Models of Functional and Effective Neuronal Connectivity 141 Introduction 141 Background and Overview 143 The Crosscorrelogram 143 Information-Theoretic Methods 145 Granger Causality 147 Generalized Linear Models 147 Graphical Models 150 Effective Connectivity 150 Graph-Based Functional Connectivity Inference 154 Results 158 Spiking Neural Model 158 Inferring Effective Connectivity 161 Deciphering Mono-Synaptic Connectivity 164 Unobserved Common Input and Scalability 165 Choosing the Markov Lag 166 Position-Specific Cell Assemblies in the Medial Prefrontal Cortex (mPFC) 166 Identifying Functional Connectivity 169 Scalability 172 Choosing the Time Scale 174 Discussion and Future Directions 176 Acknowledgments 178 References 179 State Space Modeling of Neural Spike Train and Behavioral Data 187 Introduction 187 State Space Modeling Paradigm 189 Notation 189 Recursive Form of Bayes' Rule 190 Classes of Filtering and Smoothing Problems 191 Applications of the State Space Paradigm in Neuroscience Data Analysis 192 Neural Spike Train Decoding and Point Process Filter Algorithms 192 Neural Receptive Field Plasticity and Instantaneous Steepest Descent Filtering 198 Tracking Spatial Receptive Field and Particle Filtering 202 Dynamic Analysis of Behaviorial Learning Experiments and the Expectation-Maximization Algorithm 208 Markov Chain Monte Carlo Methods and the Analysis of Cortical UP/DOWN States 214 State Space Smoothing, Dynamic Parameter Estimation, and Analysis of Population Learning 217 Discussion 223 Acknowledgments 225 References 225 Neural Decoding for Motor and Communication Prostheses 231 Introduction 231 Plan and Movement Neural Activity 234 Continuous Decoding for Motor Prostheses 237 Recursive Bayesian Decoders 239 Mixture of Trajectory Models 242 Mixture of Trajectory Models Framework 243 Mixture of Trajectory Models for Goal-Directed Movements 244 Incorporating Target Information from Plan Activity 245 Results 247 Discrete Decoding for Communication Prostheses 252 Independent Gaussian and Poisson Models 253 Factor Analysis Methods 254 Modeling Shared Variability Using Factor Analysis 257 Decoding Reach Targets Using Factor Analysis 259 A Unified Latent Space 260 Results 262 Discussion 264 Future Directions 267 Acknowledgments 269 References 270 Inner Products for Representation and Learning in the Spike Train Domain 276 Introduction 276 Functional Representations of Spike Trains 279 Synaptic Models 279 Intensity Estimation 283 Inner Products for Spike Trains 285 Defining Inner Products for Spike Trains 287 Inner Products for Spike Times 287 Functional Representations of Spike Trains 288 Properties of the Defined Inner Products 292 Distances Induced by Inner Products 294 Applications 296 Unsupervised Learning: Principal Component Analysis 296 Derivation 296 Results 300 Supervised Learning: Fisher's Linear Discriminant 305 Derivation 305 Results 308 Discussion 311 Higher-Order Spike Interactions through Nonlinearity 312 Proofs 313 Brief Introduction to RKHS Theory 315 Acknowledgments 316 References 316 Signal Processing and Machine Learning for Single-trial Analysis of Simultaneously Acquired EEG and fMRI 321 Introduction 321 Hardware Design and Setup: Challenges in EEG/fMRI Acquisition 323 EEG Cap Design 325 EEG Amplifier 325 Synchronized Sampling 327 Signal Processing and Removal of BCG Artifacts 327 The Kirchhoffian Account 329 Linking Single-Trial Variations of Task-Relevant EEG Components to the BOLD Signal 333 Identifying EEG Components Using Linear Discrimination 333 Constructing fMRI Regressors from Single-Trial Variability in EEG Components 334 Results 337 Future Directions 340 Acknowledgments 341 References 341 Statistical Pattern Recognition and Machine Learning in Brain–Computer Interfaces 345 Introduction 345 Signal Processing and Pattern Recognition in BCI Systems 347 Signal Acquisition and Major Signal Types 347 Pattern Recognition and Machine Learning 349 Preprocessing 350 Spatial Filtering: Bipolar, Laplacian, and Common Average Rereferencing 350 Spatial Filtering: Common Spatial Patterns 351 Feature Extraction 352 Classification 353 Linear Discriminant Analysis 354 Regularized Linear Discriminant Analysis 355 Quadratic Discriminant Analysis 355 Support Vector Machine 356 Multiclass Classification 357 Learning Vector Quantization and Distinction-Sensitive Learning Vector Quantization 357 Evaluation of Classification Performance 358 Applications 360 P300-Based Control of a Humanoid Robot 361 Motor Imagery—Based Control of Virtual Environments 364 Discussion 368 Acknowledgments 369 References 370 Prediction of Muscle Activity from Cortical Signals to Restore Hand Grasp in Subjects with Spinal Cord Injury 378 Introduction 378 Background 380 BMIs as a Potential Control Solution 383 BMIs for Control of Dynamics 385 Methods 386 Isometric Wrist Torque Tasks 387 Hand Grasp Tasks 388 Surgical Methods 388 Electrode Array Implantation Surgery 389 Limb Implants 389 Data Collection 390 Linear Systems Identification 390 Nerve Block Effectiveness 392 Stimulation 392 Real-Time Control 393 Results 393 Offline Signal Prediction 393 Isometric Wrist Tasks 394 Grasp Task 396 Real-Time FES Control 398 Isometric Wrist Task 399 Grasp Task 401 Discussion 403 Successful Proof of Concept 403 Limitations in the Control of Complex Motor Tasks 404 Limitations Related to the Use of FES for Control 406 Future Directions 408 Control of Higher-Dimensional Movement Using Natural Muscle Synergies 408 Adaptation 409 Acknowledgments 410 References 410 Index 416 A 416 B 416 C 416 D 416 E 416 F 417 G 417 H 417 I 418 J 418 K 418 L 418 M 418 N 419 O 419 P 419 Q 419 R 419 S 419 T 420 U 420 V 420 W 420 Z 420 This is a uniquely comprehensive reference that summarizes the state of the art of signal processing theory and techniques for solving emerging problems in neuroscience, and which clearly presents new theory, algorithms, software and hardware tools that are specifically tailored to the nature of the neurobiological environment. It gives a broad overview of the basic principles, theories and methods in statistical signal processing for basic and applied neuroscience problems.

Written by experts in the field, the book is an ideal reference for researchers working in the field of neural engineering, neural interface, computational neuroscience, neuroinformatics, neuropsychology and neural physiology. By giving a broad overview of the basic principles, theories and methods, it is also an ideal introduction to statistical signal processing in neuroscience.

• A comprehensive overview of the specific problems in neuroscience that require application of existing and development of new theory, techniques, and technology by the signal processing community
• Contains state-of-the-art signal processing, information theory, and machine learning algorithms and techniques for neuroscience research
• Presents quantitative and information-driven science that has been, or can be, applied to basic and translational neuroscience problems

Detection and classification of extracellular action potential recordings / Karim G. Oweiss and Mehdi A. Aghagolzadeh Information-theoretic analysis of neural data / Don H. Johnson Identification of nonlinear dynamics in neural population activity / Dong Song and Theodore Berger Graphical models of functional and effective neuronal connectivity / Seif M. Eldawlatly and Karim G. Oweiss State-space modeling of neural spike train and behavioral data / Zhe Chen, Riccardo Barbieri, and Emery N. Brown Neural decoding for motor and communication prostheses / Byron M. Yu ... [et al.] Inner products for representation and learning in the spike train domain / Antonio R.C. Paiva, Il Park, and Jose C. Principe Signal processing and machine learning for single-trial analysis of simultaneously acquired EEG and fMRI / Paul Sajda ... [et al.] Statistical pattern recognition and machine learning in brain-computer interfaces / Rajesh P.N. Rao and Reinhold Scherer Getting a grip on spinal cord injury : a novel application of a brain machine interface / Emily R. Oby ... [et al.]. This is a uniquely comprehensive reference that summarizes the state of the art of signal processing theory and techniques for solving emerging problems in neuroscience, and which clearly presents new theory, algorithms, software and hardware tools that are specifically tailored to the nature of the neurobiological environment. It gives a broad overview of the basic principles, theories and methods in statistical signal processing for basic and applied neuroscience problems. Written by experts in the field, the book is an ideal reference for researchers working in the field of neural engineering, neural interface, computational neuroscience, neuroinformatics, neuropsychology and neural physiology. By giving a broad overview of the basic principles, theories and methods, it is also an ideal introduction to statistical signal processing in neuroscience.