Contents

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

Preface xvii
Acknowledgments xix

Part I: Introduction 1

1 The Purpose of This Book, Who Should Read It, and How to Use It 3

1.1 What Is Cognitive Electrophysiology? 3

1.2 What Is the Purpose of This Book? 4
1.3 Why Shouldn’t You Use <Insert Name of M/EEG Software Analysis Package>? 6
1.4 Why Program Analyses, and Why in Matlab? 7
1.5 How Best to Learn from and Use This Book 8
1.6 Sample Data and Online Code 9
1.7 Terminology Used in This Book 10
1.8 Exercises 11
1.9 Is Everything There Is to Know about EEG Analyses in This Book? 11
1.10 Who Should Read This Book? 12
1.11 Is This Book Difficult? 13
1.12 Questions? 14

2 Advantages and Limitations of Time- and Time-Frequency-Domain Analyses 15

2.1 Why EEG? 15

2.2 Why Not EEG? 16
2.3 Interpreting Voltage Values from the EEG Signal 17
Copyright @ 2014. The MIT Press.

2.4 Advantages of Event-Related Potentials 18

2.5 Limitations of ERPs 19
2.6 Advantages of Time-Frequency-Based Approaches 21
2.7 Limitations of Time-Frequency-Based Approaches 23
2.8 Temporal Resolution, Precision, and Accuracy of EEG 24
2.9 Spatial Resolution, Precision, and Accuracy of EEG 26

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 vi Contents

2.10 Topographical Localization versus Brain Localization 27

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

2.11 EEG or MEG? 28

2.12 Costs of EEG Research 29

3 Interpreting and Asking Questions about Time-Frequency Results 31

3.1 EEG Time-Frequency: The Basics 31

3.2 Ways to View Time-Frequency Results 35
3.3 tfviewerx and erpviewerx 37
3.4 How to View and Interpret Time-Frequency Results 38
3.5 Things to Be Suspicious of When Viewing Time-Frequency Results 40
3.6 Do Results in Time-Frequency Plots Mean That There Were Neural Oscillations? 41

4 Introduction to Matlab Programming 43

4.1 Write Clean and Efficient Code 43

4.2 Use Meaningful File and Variable Names 45
4.3 Make Regular Backups of Your Code and Keep Original Copies of Modified Code 45
4.4 Initialize Variables 46
4.5 Help! 47
4.6 Be Patient and Embrace the Learning Experience 49
4.7 Exercises 50

5 Introduction to the Physiological Bases of EEG 51

5.1 Biophysical Events That Are Measurable with EEG 51

5.2 Neurobiological Mechanisms of Oscillations 54
5.3 Phase-Locked, Time-Locked, Task-Related 55
5.4 Neurophysiological Mechanisms of ERPs 56
5.5 Are Electrical Fields Causally Involved in Cognition? 57
5.6 What if Electrical Fields Are Not Causally Involved in Cognition? 59

6 Practicalities of EEG Measurement and Experiment Design 61

6.1 Designing Experiments: Discuss, Pilot, Discuss, Pilot 61

6.2 Event Markers 62
6.3 Intra- and Intertrial Timing 63
6.4 How Many Trials You Will Need 65
6.5 How Many Electrodes You Will Need 65
6.6 Which Sampling Rate to Use When Recording Data 66
Copyright @ 2014. The MIT Press.

6.7 Other Optional Equipment to Consider 68

Part II: Preprocessing and Time-Domain Analyses 71

7 Preprocessing Steps Necessary and Useful for Advanced Data Analysis 73

7.1 What Is Preprocessing? 73

7.2 The Balance between Signal and Noise 73

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 Contents vii

7.3 Creating Epochs 75

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

7.4 Matching Trial Count across Conditions 78

7.5 Filtering 80
7.6 Trial Rejection 80
7.7 Spatial Filtering 81
7.8 Referencing 82
7.9 Interpolating Bad Electrodes 83
7.10 Start with Clean Data 85

8 EEG Artifacts: Their Detection, Influence, and Removal 87

8.1 Removing Data Based on Independent Components Analysis 87

8.2 Removing Trials because of Blinks 89
8.3 Removing Trials because of Oculomotor Activity 90
8.4 Removing Trials Based on EMG in EEG Channels 92
8.5 Removing Trials Based on Task Performance 93
8.6 Removing Trials Based on Response Hand EMG 94
8.7 Train Subjects to Minimize Artifacts 95
8.8 Minimize Artifacts during Data Collection 96

9 Overview of Time-Domain EEG Analyses 97

9.1 Event-Related Potentials 97

9.2 Filtering ERPs 97
9.3 Butterfly Plots and Global Field Power/Topographical Variance Plots 100
9.4 The Flicker Effect 101
9.5 Topographical Maps 102
9.6 Microstates 104
9.7 ERP Images 105
9.8 Exercises 106

Part III: Frequency and Time-Frequency Domains Analyses 109

10 The Dot Product and Convolution 111

10.1 Dot Product 111

10.2 Convolution 113
10.3 How Does Convolution Work? 113
Copyright @ 2014. The MIT Press.

10.4 Convolution versus Cross-Covariance 118

10.5 The Purpose of Convolution for EEG Data Analyses 118
10.6 Exercises 119

11 The Discrete Time Fourier Transform, the FFT, and the Convolution
Theorem 121

11.1 Making Waves 121

11.2 Finding Waves in EEG Data with the Fourier Transform 123

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 viii Contents

11.3 The Discrete Time Fourier Transform 124

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

11.4 Visualizing the Results of a Fourier Transform 126

11.5 Complex Results and Negative Frequencies 127
11.6 Inverse Fourier Transform 129
11.7 The Fast Fourier Transform 130
11.8 Stationarity and the Fourier Transform 132
11.9 Extracting More or Fewer Frequencies than Data Points 134
11.10 The Convolution Theorem 135
11.11 Tips for Performing FFT-Based Convolution in Matlab 138
11.12 Exercises 139

12 Morlet Wavelets and Wavelet Convolution 141

12.1 Why Wavelets? 141

12.2 How to Make Wavelets 144
12.3 Wavelet Convolution as a Bandpass Filter 146
12.4 Limitations of Wavelet Convolution as Discussed Thus Far 147
12.5 Exercises 150

13 Complex Morlet Wavelets and Extracting Power and Phase 151

13.1 The Wavelet Complex 151

13.2 Imagining the Imaginary 153
13.3 Rectangular and Polar Notation and the Complex Plane 153
13.4 Euler’s Formula 155
13.5 Euler’s Formula and the Result of Complex Wavelet Convolution 158
13.6 From Time Point to Time Series 162
13.7 Parameters of Wavelets and Recommended Settings 162
13.8 Determining the Frequency Smoothing of Wavelets 172
13.9 Tips for Writing Efficient Convolution Code in Matlab 173
13.10 Describing This Analysis in Your Methods Section 173
13.11 Exercises 173

14 Bandpass Filtering and the Hilbert Transform 175

14.1 Hilbert Transform 175

14.2 Filtering Data before Applying the Hilbert Transform 178
14.3 Finite versus Infinite Impulse Response Filters 179
14.4 Bandpass, Band-Stop, High-Pass, Low-Pass 179
Copyright @ 2014. The MIT Press.

14.5 Constructing a Filter 180

14.6 Check Your Filters 186
14.7 Applying the Filter to Data 188
14.8 Butterworth (IIR) Filter 189
14.9 Filtering Each Trial versus Filtering Concatenated Trials 190
14.10 Multiple Frequencies 190

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 Contents ix

14.11 A World of Filters 192

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

14.12 Describing This Analysis in Your Methods Section 193

14.13 Exercises 193

15 Short-Time FFT 195

15.1 How the Short-Time FFT Works 195

15.2 Taper the Time Series 195
15.3 Time Segment Lengths and Overlap 199
15.4 Power and Phase 200
15.5 Describing This Analysis in Your Methods Section 201
15.6 Exercises 201

16 Multitapers 203

16.1 How the Multitaper Method Works 203

16.2 The Tapers 205
16.3 When You Should and Should Not Use Multitapers 207
16.4 The Multitaper Framework and Advanced Topics 208
16.5 Describing This Analysis in Your Methods Section 208
16.6 Exercises 208

17 Less Commonly Used Time-Frequency Decomposition Methods 211

17.1 Autoregressive Modeling 211

17.2 Hilbert-Huang (Empirical Mode Decomposition) 211
17.3 Matching Pursuit 212
17.4 P-Episode 213
17.5 S-Transform 214

18 Time-Frequency Power and Baseline Normalizations 217

18.1 1/f Power Scaling 217

18.2 The Solution to 1/f Power in Task Designs 220
18.3 Decibel Conversion 220
18.4 Percentage Change and Baseline Division 223
18.5 Z-Transform 224
18.6 Not All Transforms Are Equal 224
18.7 Other Transforms 227
18.8 Mean versus Median 227
Copyright @ 2014. The MIT Press.

18.9 Single-Trial Baseline Normalization 230

18.10 The Choice of Baseline Time Window 230
18.11 Disadvantages of Baseline-Normalized Power 234
18.12 Signal-to-Noise Estimates 234
18.13 Number of Trials and Power Estimates 237
18.14 Downsampling Results after Analyses 239

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 x Contents

18.15 Describing This Analysis in Your Methods Section 239

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

18.16 Exercises 239

19 Intertrial Phase Clustering 241

19.1 Why Phase Values Cannot Be Averaged 241

19.2 Intertrial Phase Clustering 242
19.3 Strength in Numbers 246
19.4 Using ITPC When There Are Few Trials or Condition Differences in Trial Count 248
19.5 Effects of Temporal Jitter on ITPC and Power 249
19.6 ITPC and Power 251
19.7 Weighted ITPC 253
19.8 Multimodal Phase Distributions 257
19.9 Spike-Field Coherence 257
19.10 Describing This Analysis in Your Methods Section 257
19.11 Exercises 258

20 Differences among Total, Phase-Locked, and Non-Phase-Locked Power and Intertrial

Phase Consistency 259

20.1 Total Power 259

20.2 Non-Phase-Locked Power 259
20.3 Phase-Locked Power 260
20.4 ERP Time-Frequency Power 262
20.5 Intertrial Phase Clustering 262
20.6 When to Use What Approach 263
20.7 Exercise 264

21 Interpretations and Limitations of Time-Frequency Power and ITPC Analyses 265

21.1 Terminology 265

21.2 When to Use What Time-Frequency Decomposition Method 266
21.3 Interpreting Time-Frequency Power 267
21.4 Interpreting Time-Frequency Intertrial Phase Clustering 268
21.5 Limitations of Time-Frequency Power and Intertrial Phase Clustering 269
21.6 Do Time-Frequency Analyses Reveal Neural Oscillations? 270

Part IV: Spatial Filters 273

Copyright @ 2014. The MIT Press.

22 Surface Laplacian 275

22.1 What Is the Surface Laplacian? 275

22.2 Algorithms for Computing the Surface Laplacian for EEG Data 278
22.3 Surface Laplacian for Topographical Localization 283
22.4 Surface Laplacian for Connectivity Analyses 286

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 Contents xi

22.5 Surface Laplacian for Cleaning Topographical Noise 288

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

22.6 Describing This Analysis in Your Methods Section 289

22.7 Exercises 289

23 Principal Components Analysis 291

23.1 Purpose and Interpretations of Principal Components Analysis 291

23.2 How PCA Is Computed 293
23.3 Distinguishing Significant from Nonsignificant Components 297
23.4 Rotating PCA Solutions 300
23.5 Time-Resolved PCA 300
23.6 PCA with Time-Frequency Information 301
23.7 PCA across Conditions 303
23.8 Independent Components Analysis 304
23.9 Describing This Method in Your Methods Section 305
23.10 Exercises 305

24 Basics of Single-Dipole and Distributed-Source Imaging 307

24.1 The Forward Solution 307

24.2 The Inverse Problem 310
24.3 Dipole Fitting 311
24.4 Nonadaptive Distributed-Source Imaging Methods 312
24.5 Adaptive Distributed-Source Imaging 313
24.6 Theoretical and Practical Limits of Spatial Precision and Resolution 314

Part V: Connectivity 317

25 Introduction to the Various Connectivity Analyses 319

25.1 Why Only Two Sites (Bivariate Connectivity)? 319

25.2 Important Concepts Related to Bivariate Connectivity 320
25.3 Which Measure of Connectivity Should Be Used? 322
25.4 Phase-Based Connectivity 323
25.5 Power-Based Connectivity 324
25.6 Granger Prediction 324
25.7 Mutual Information 325
25.8 Cross-Frequency Coupling 325
Copyright @ 2014. The MIT Press.

25.9 Graph Theory 326

25.10 Potential Confound of Volume Conduction 326

26 Phase-Based Connectivity 333

26.1 Terminology 333

26.2 ISPC over Time 333

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 xii Contents

26.3 ISPC-Trials 337

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

26.4 ISPC and the Number of Trials 340

26.5 Relation between ISPC and Power 341
26.6 Weighted ISPC-Trials 342
26.7 Spectral Coherence (Magnitude-Squared Coherence) 342
26.8 Phase Lag-Based Measures 346
26.9 Which Measure of Phase Connectivity Should You Use? 351
26.10 Testing the Mean Phase Angle 352
26.11 Describing These Analyses in Your Methods Section 355
26.12 Exercises 355

27 Power-Based Connectivity 357

27.1 Spearman versus Pearson Coefficient for Power Correlations 357

27.2 Power Correlations over Time 361
27.3 Power Correlations over Trials 363
27.4 Partial Correlations 366
27.5 Matlab Programming Tips 367
27.6 Describing This Analysis in Your Methods Section 370
27.7 Exercises 370

28 Granger Prediction 371

28.1 Univariate Autoregression 372

28.2 Bivariate Autoregression 373
28.3 Autoregression Errors and Error Variances 375
28.4 Granger Prediction over Time 376
28.5 Model Order 377
28.6 Frequency Domain Granger Prediction 380
28.7 Time Series Covariance Stationarity 383
28.8 Baseline Normalization of Granger Prediction Results 384
28.9 Statistics 385
28.10 Additional Applications of Granger Prediction 387
28.11 Exercises 388

29 Mutual Information 389

29.1 Entropy 389

29.2 How Many Histogram Bins to Use 391
Copyright @ 2014. The MIT Press.

29.3 Enjoy the Entropy 395

29.4 Joint Entropy 396
29.5 Mutual Information 396
29.6 Mutual Information and Amount of Data 398
29.7 Mutual Information with Noisy Data 400
29.8 Mutual Information over Time or over Trials 402

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 Contents xiii

29.9 Mutual Information on Real Data 402

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

29.10 Mutual Information on Frequency-Band-Specific Data 402

29.11 Lagged Mutual Information 404
29.12 Statistics 405
29.13 More Information 406
29.14 Describing This Analysis in Your Methods Section 407
29.15 Exercises 407

30 Cross-Frequency Coupling 409

30.1 Visual Inspection of Cross-Frequency Coupling 409

30.2 Power-Power Correlations 410
30.3 A Priori Phase-Amplitude Coupling 410
30.4 Separating Task-Related Phase and Power Coactivations from Phase-Amplitude Coupling 418
30.5 Mixed A Priori/Exploratory Phase-Amplitude Coupling 419
30.6 Exploratory Phase-Amplitude Coupling 421
30.7 Notes about Phase-Amplitude Coupling 422
30.8 Phase-Phase Coupling 425
30.9 Other Methods for Quantifying Cross-Frequency Coupling 426
30.10 Cross-Frequency Coupling over Time or over Trials 426
30.11 Describing This Analysis in Your Methods Section 426
30.12 Exercises 427

31 Graph Theory 429

31.1 Networks as Matrices and Graphs 429

31.2 Thresholding Connectivity Matrices 432
31.3 Connectivity Degree 433
31.3 Clustering Coefficient 436
31.4 Path Length 440
31.5 Small-World Networks 441
31.6 Statistics 444
31.7 How to Describe These Analyses in Your Paper 446
31.8 Exercises 446

Part VI: Statistical Analyses 447

32 Advantages and Limitations of Different Statistical Procedures 449

Copyright @ 2014. The MIT Press.

32.1 Are Statistics Necessary? 449

32.2 At What Level Should Statistics Be Performed? 450
32.3 What p-Value Should Be Used, and Should Multiple-Comparisons Corrections Be Applied? 452
32.4 Are p-Values the Only Statistical Metric? 453
32.5 Statistical Significance versus Practical Significance 454
32.6 Type I and Type II Errors 455

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 xiv Contents

32.7 What Kinds of Statistics Should Be Applied? 456

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

32.8 How to Combine Data across Subjects 457

33 Nonparametric Permutation Testing 459

33.1 Advantages of Nonparametric Permutation Testing 459

33.2 Creating a Null-Hypothesis Distribution 461
33.3 How Many Iterations Are Necessary for the Null-Hypothesis Distribution? 462
33.4 Determining Statistical Significance 463
33.5 Multiple Comparisons and Their Corrections 468
33.6 Correction for Multiple Comparisons Using Pixel-Based Statistics 470
33.7 Corrections for Multiple Comparisons Using Cluster-Based Statistics 471
33.8 False Discovery Rate for Multiple-Comparisons Correction 475
33.9 What Should Be Permuted? 475
33.10 Nonparametric Permutation Testing beyond Simple Bivariate Cases 477
33.11 Describing This Analysis in Your Methods Section 477

34 Within-Subject Statistical Analyses 479

34.1 Changes in Task-Related Power Compared to Baseline 479

34.2 Discrete Condition Differences in Power 481
34.3 Continuous Relationship with Power: Single-Trial Correlations 482
34.4 Continuous Relationships with Power: Single-Trial Multiple Regression 485
34.5 Determining Statistical Significance of Phase-Based Data 487
34.6 Testing Preferred Phase Angle across Conditions 490
34.7 Testing the Statistical Significance of Correlation Coefficients 490

35 Group-Level Analyses 493

35.1 Avoid Circular Inferences 493

35.2 Group-Level Analysis Strategy 1: Test Each Pixel and Apply a Mapwise Threshold 494
35.3 Group-Level Analysis Strategy 2a: Time-Frequency Windows for Hypothesis-Driven Analyses 495
35.4 Group-Level Analysis Strategy 2b: Subject-Specific Time-Frequency Windows for
Hypothesis-Driven Analyses 499
35.5 Determining How Many Subjects You Need for Group-Level Analyses 501

36 Recommendations for Reporting Results in Figures, Tables, and Text 503

36.1 Recommendation 1: One Figure, One Idea 504

36.2 Recommendation 2: Show Data 504
Copyright @ 2014. The MIT Press.

36.3 Recommendation 3: Highlight Significant Effects Instead of Removing Nonsignificant Effects 506
36.4 Recommendation 4: Show Specificity (or Lack Thereof) in Frequency, Time, and Space 508
36.5 Recommendation 5: Use Color 509
36.6 Recommendation 6: Use Informative Figure Labels and Captions 511
36.7 Recommendation 7: Avoid Showing “Representative” Data 512
36.8 A Checklist for Making Figures 514

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
 Contents xv

36.9 Tables 514

All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.

36.10 Reporting Results in the Results Section 515

Part VII: Conclusions and Future Directions 519

37 Recurring Themes in This Book and Some Personal Advice 521

37.1 Theme: Myriad Possible Analyses 521

37.2 Advice: Avoid the Paralysis of Analysis 522
37.3 Theme: You Don’t Have to Program Your Own Analyses, but You Should Know How Analyses
Work 522
37.4 Advice: If It Feels Wrong, It Probably Is 523
37.5 Advice: When in Doubt, Plot It Out 523
37.6 Advice: Know These Three Formulas like the Back of Your Hand 524
37.7 Theme: Connectivity over Trials or over Time 524
37.8 Theme: Most Analysis Parameters Introduce Bias 525
37.9 Theme: Write a Clear Methods Section so Others Can Replicate Your Analyses 525
37.10 Theme: Use Descriptive and Appropriate Analysis Terms 526
37.11 Advice: Interpret Null Results Cautiously 527
37.12 Advice: Try Simulations but Also Trust Real Data 528
37.13 Advice: Trust Replications 529
37.14 Theme: Analyses Are Not Right or Wrong; They Are Appropriate or Inappropriate 529
37.15 Advice: Hypothesis Testing Is Good/Bad, and So Is Data-Driven Exploration 530
37.16 Advice: Find Something That Drives You and Study It 531
37.17 Cognitive Electrophysiology: The Art of Finding Anthills on Mountains 532

38 The Future of Cognitive Electrophysiology 535

38.1 Developments in Analysis Methods 535

38.2 Developments in Understanding the Neurophysiology of EEG 537
38.3 Developments in Experiment Design 538
38.4 Developments in Measurement Technology 539
38.5 The Role of the Body in Brain Function 540
38.6 Determining Causality 541
38.7 Inferring Cognitive States from EEG Signatures: Inverse Inference 542
38.8 Tables of Activation 543
38.9 Disease Diagnosis and Predicting Treatment Course and Success 543
Copyright @ 2014. The MIT Press.

38.10 Clinical Relevance Is Not Necessary for the Advancement of Science 544
38.11 Replications 545
38.12 Double-Blind Review for Scientific Publications 547
38.13 ? 548

References 549
Index 573

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost
All rights reserved. May not be reproduced in any form without permission from the publisher, except fair uses permitted under U.S. or applicable copyright law.
Copyright @ 2014. The MIT Press.

EBSCO : eBook Collection (EBSCOhost) - printed on 10/8/2018 5:17 AM via UNIVERSITY OF AUCKLAND
AN: 689432 ; Cohen, Mike X..; Analyzing Neural Time Series Data : Theory and Practice
Account: s2802549.main.ehost