Cognitive Modeling of Human Memory and Learning

LIDIA GHOSH, PHD, is currently a post-doctoral research fellow on Brain Science and Memory Research, granted by Liverpool Hope University to Jadavpur University, India. AMIT KONAR, PHD, is currently a Professor in the dept. of Electronics and Tele-Communication Engineering (ETCE), Jadavpur University. He is an author of 15 books including a Wiley title: Emotion Recognition-A Pattern Analysis Approach. PRATYUSHA RAKSHIT, PHD, is an Assistant Professor of ETCE dept., Jadavpur University, India and is currently on lien to Basque Centre for Applied Mathematics, Bilbao, Spain.

• Preface xiAcknowledgments xviiAbout the Authors xix1 Introduction to Brain-Inspired Memory and Learning Models 11.1 Introduction 11.2 Philosophical Contributions to Memory Research 31.2.1 Atkinson and Shiffrin’s Model 41.2.2 Tveter’s Model 51.2.3 Tulving’s Model 61.2.4 The Parallel and Distributed Processing (PDP) Approach 61.2.5 Procedural and Declarative Memory 81.3 Brain-Theoretic Interpretation of Memory Formation 101.3.1 Coding for Memory 101.3.2 Memory Consolidation 121.3.3 Location of Stored Memories 141.3.4 Isolation of Information in Memory 151.4 Cognitive Maps 161.5 Neural Plasticity 171.6 Modularity 181.7 The Cellular Process Behind STM Formation 181.8 LTM Formation 201.9 Brain Signal Analysis in the Context of Memory and Learning 201.9.1 Association of EEG α and θ Band with Memory Performances 211.9.2 Oscillatory β and γ Frequency Band Activation in STM Performance 241.9.3 Change in EEG Band Power with Changing Working Memory Load 241.9.4 Effects of Electromagnetic Field on the EEG Response of Working Memory 271.9.5 EEG Analysis to Discriminate Focused Attention and WM Performance 281.9.6 EEG Power Changes in Memory Repetition Effect 291.9.7 Correlation Between LTM Retrieval and EEG Features 321.9.8 Impact of Math Anxiety on WM Response: An EEG Study 341.10 Memory Modeling by Computational Intelligence Techniques 351.11 Scope of the Book 39References 432 Working Memory Modeling Using Inverse Fuzzy Relational Approach 512.1 Introduction 522.2 Problem Formulation and Approach 542.2.1 Independent Component Analysis as a Source Localization Tool 552.2.2 Independent Component Analysis vs. Principal Component Analysis 582.2.3 Feature Extraction 582.2.4 Phase 1: WM Modeling 592.2.4.1 Step I: WM Modeling of Subject Using EEG Signals During Full Face Encoding and Recall from Specific Part of Same Face 602.2.4.2 Step II: WM Modeling of Subject Using EEG Signals During Full Face Encoding and Recall from All Parts of Same Face 622.2.5 Phase 2: WM Analysis 622.2.6 Finding Max–Min Compositional Inverse of Weight Matrix Wck 652.3 Experiments and Performance Analysis 702.3.1 Experimental Set-up 712.3.2 Source Localization Using eLORETA 732.3.3 Pre-processing 742.3.4 Selection of EEG Features 742.3.5 WM Model Consistency Across Partial Face Stimuli 772.3.6 Inter-person Variability of W 772.3.7 Variation in Imaging Attributes 772.3.8 Comparative Analysis with Existing Fuzzy Inverse Relations 842.4 Discussion 852.5 Conclusions 86References 883 Short-Term Memory Modeling in Shape-Recognition Task by Type-2 Fuzzy Deep Brain Learning 933.1 Introduction 943.2 System Overview 963.3 Brain Functional Mapping Using Type-2 Fuzzy DBLN 1013.3.1 Overview of Type-2 Fuzzy Sets 1033.3.2 Type-2 Fuzzy Mapping and Parameter Adaptation by Perceptron-Like Learning 1043.3.2.1 Construction of the Proposed Interval Type-2 Fuzzy Membership Function (IT2MF) 1043.3.2.2 Construction of IT2FS-Induced Mapping Function 1053.3.2.3 Secondary Membership Function Computation of Proposed GT2FS 1073.3.2.4 Proposed General Type-2 Fuzzy Mapping 1083.3.3 Perceptron-Like Learning for Weight Adaptation 1103.3.4 Training of the Proposed Shape-Reconstruction Architecture 1113.3.5 The Test Phase of the Memory Model 1133.4 Experiments and Results 1133.4.1 Experimental Set-up 1133.4.2 Experiment 1: Validation of the STM Model with Respect to Error Metric 𝜉 1163.4.3 Experiment 2: Similar Encoding by a Subject for Similar Input Object Shapes 1163.4.4 Experiment 3: Study of Subjects’ Learning Ability with Increasing Complexity in Object Shape 1173.4.5 Experiment 4: Convergence Time of the Weight Matrix G for Increased Complexity of the Input Shape Stimuli 1183.4.6 Experiment 5: Abnormality in G matrix for the Subjects with Brain Impairment 1193.5 Biological Implications 1203.6 Performance Analysis 1223.6.1 Performance Analysis of the Proposed T2FS Methods 1233.6.2 Computational Performance Analysis of the Proposed T2FS Methods 1233.6.3 Statistical Validation Using Wilcoxon Signed-Rank Test 1243.6.4 Optimal Parameter Selection and Robustness Study 1263.7 Conclusions 127References 1304 EEG Analysis for Subjective Assessment of Motor Learning Skill in Driving Using Type-2 Fuzzy Reasoning 1374.1 Introduction 1384.2 System Overview 1404.2.1 Rule Design to Determine the Degree of Learning 1414.2.2 Single Trial Detection of Brain Signals 1444.2.2.1 Feature Extraction 1444.2.2.2 Feature Selection 1454.2.2.3 Classification 1454.2.3 Type-2 Fuzzy Reasoning 1464.2.4 Training and Testing of the Classifiers 1464.3 Determining Type and Degree of Learning by Type-2 Fuzzy Reasoning 1474.3.1 Preliminaries on IT2FS and GT2FS 1474.3.2 Proposed Reasoning Method 1: CIT2FS-Based Reasoning 1484.3.3 Computation of Percentage Normalized Degree of Learning 1504.3.4 Optimal 𝜆 Selection in IT2FS Reasoning 1514.3.5 Proposed Reasoning Method 2: Triangular Vertical Slice (TVS)-Based CGT2FS Reasoning 1514.3.5.1 Closed General Type-2 Fuzzy Inference Generation 1514.3.5.2 Time complexity 1544.3.6 Proposed Reasoning Method 3: CGT2FS Reasoning with Gaussian Secondary MF 1544.3.6.1 Time-Complexity 1564.4 Experiments and Results 1574.4.1 The Experimental Set-up 1574.4.2 Stimulus Presentation 1574.4.3 Experiment 1: Source Localization Using eLORETA 1584.4.4 Experiment 2: Validation of the Rules 1594.4.5 Experiment 3: Pre-processing and Artifact Removal Using ICA 1594.4.6 Experiment 4: N400 Old/New Effect Observation over the Successive Trials 1634.4.7 Experiment 5: Selection of the Discriminating EEG Features Using PCA 1634.5 Performance Analysis and Statistical Validation 1644.5.1 Performance Analysis of the LSVM Classifiers 1644.5.2 Robustness Study 1654.5.3 Performance Analysis of the Proposed T2FS Reasoning Methods 1664.5.4 Computational Performance Analysis of the Proposed T2FS Reasoning Methods 1664.5.5 Statistical Validation Using Wilcoxon Signed-Rank Test 1684.6 Conclusions 169References 1695 EEG Analysis to Decode Human Memory Responses in Face Recognition Task Using Deep LSTM Network 1755.1 Introduction 1765.2 CSP Modeling 1795.2.1 The Standard CSP Algorthm 1795.2.2 The Proposed CSP Algorithm 1805.3 Proposed LSTM Classifier with Attention Mechanism 1835.3.1 Attention Mechanism in Each LSTM Unit 1845.4 Experiments and Results 1885.4.1 Experimental Set-up 1885.4.2 Experiment 1: Activated Brain Region Selection Using eLORETA 1885.4.3 Experiment 2: Detection of the ERP Signals Associated with the Familiar and Unfamiliar Face Discrimination 1905.4.4 Experiment 3: Performance Analysis of the Proposed CSP Algorithm as a Feature Extraction Technique 1915.4.5 Experiment 4: Performance Analysis of the Proposed LSTM-Based Classifier 1925.4.6 Experiment 5: Classifier Performance Analysis with Varying EEG Time-Window Length 1945.4.7 Statistical Validation of the Proposed LSTM Classifier Using McNemar’s Test 1955.5 Conclusions 196References 1976 Cognitive Load Assessment in Motor Learning Tasks by Near-Infrared Spectroscopy Using Type-2 Fuzzy Sets 2036.1 Introduction 2036.2 Principles and Methodologies 2066.2.1 Normalization of the Raw Data 2066.2.2 Pre-processing 2076.2.3 Feature Extraction 2086.2.4 Training Instance Generation for Offline Training 2086.2.5 Feature Selection Using Evolutionary Algorithm 2096.2.6 Classifier Training and Testing 2106.3 Classifier Design 2116.3.1 Preliminaries on IT2FS and GT2FS 2116.3.2 IT2FS-Induced Classifier Design 2126.3.3 GT2FS-Induced Classifier Design 2166.4 Experiments and Results 2196.4.1 Experimental Set-up 2196.4.2 Participants 2196.4.3 Stimulus Presentation for Online Classification 2216.4.4 Experiment 1: Demonstration of Decreasing Cognitive Load with Increasing Learning Epochs for Similar Stimulus 2216.4.5 Experiment 2: Automatic Extraction of Discriminating fNIRs Features 2236.4.6 Experiment 3: Optimal Parameter Setting of Feature Selection and Classifier Units 2236.5 Biological Implications 2266.6 Performance Analysis 2266.6.1 Performance Analysis of the Proposed IT2FS and GT2FS Classifier 2266.6.2 Statistical Validation of the Classifier Using McNemar’s Test 2296.7 Conclusions 232References 2327 Conclusions and Future Directions of Research on BCI-Based Memory and Learning 2397.1 Self-Review of the Works Undertaken in the Book 2397.2 Limitations of EEG BCI-Based Memory Experiments 2427.3 Further Scope of Future Research on Memory and Learning 242References 245Index 247