Neocortical-Medial Temporal Lobe Interactions during Visuospatial Memory Formation and Implications for Screening of Memory Impairment in Alzheimer's Disease 公开
Haque, Rafi (Spring 2020)
Published
Abstract
Alzheimer’s disease (AD) is a chronic neurodegenerative disorder characterized by pathological changes prior to onset of clinical symptoms. The earliest site of cortical pathology is the medial temporal lobe (MTL), a region critical for the formation of memories. The presence of pathology within the MTL has generated interest in developing memory assessments that involve the MTL and may identify individuals with memory impairment due to AD. To address this need, we developed a visuospatial memory paradigm that requires participants to retrieve their past memories and recognize differences between these memories and the current experience. We presented this memory paradigm while recording electrophysiological activity in visual association areas and the MTL. Successful recognition of visuospatial memories was associated with increases in 80-120 Hz power within visual association areas and the MTL and also accompanied by 8-12 Hz communication between these regions. We then developed a version of this visuospatial memory paradigm that strictly uses eye movements as an index of retrieval. Visuospatial memory performance based on eye movements was substantially reduced in participants with AD and could differentiate the two populations with high sensitivity and specificity. To improve the scalability of this assessment, we developed an iPad-based version of the memory task that utilizes iTracker, a convolutional neural network (CNN) architecture used to track eye movements on Apple devices. We found that the iPad-based implementation could also differentiate cognitively impaired participants from healthy controls with high sensitivity and specificity. This work advances our understanding of the neocortical-MTL interactions underlying visuospatial memory formation and provides a passive, sensitive, and efficient memory assessment that may be used to identify individuals at risk of future memory impairment.
Table of Contents
Table of Contents
Chapter 1: Introduction
1.1 Molecular Pathogenesis and Pathophysiology of Alzheimer’s Disease 1
1.2 Clinical Assessment of Cognitive Impairment in Alzheimer’s Disease 4
1.3 The Medial Temporal Lobe and Memory Formation 6
1.4 Neocortical and Medial Temporal Lobe Interactions during Visuospatial Memory Formation 7
1.6 Visuospatial Memory Assessments Screening for Memory Impairment 9
1.6 Thesis Aims 10
Chapter 2: Neocortical-Medial Temporal Lobe Interactions during Visuospatial Memory Formation 12
2.1 Introduction 12
2.2 Materials and Methods 14
2.2.1 Participants 14
2.2.2 Visuospatial Recognition Memory Task 15
2.2.3 Intracranial EEG (iEEG) Recordings 16
2.2.5 Spectral Power 20
2.2.6 Spectral Coherence 20
2.2.7 Generation and Characterization of Cross-Correlograms 21
2.2.8 Metrics of Reinstatement 23
2.2.9 Temporal Dynamics of Spectral Power 24
2.2.10 Statistical Analyses 27
2.3 Results 29
2.3.1 80-120 Hz power progresses down the visual hierarchy and reflects specific visual experience 33
2.3.2 80-120 Hz power increases within visual association areas and the MTL when present visual experience differs from the remembered experience 38
2.3.3 Differences in 80-120 Hz power during manipulated images progress down the visual hierarchy 45
2.4. Discussion 51
Chapter 2: Visuospatial Memory Performance during Healthy Aging, Mild Cognitive Impairment, and Alzheimer’s Disease 62
3.1 Introduction 63
3.2 Methods 64
3.2.2 Visuospatial Memory Eye-tracking Test (VisMET) 65
3.2.3 Eye movement detection 67
3.2.4 Fixation Detection 68
3.2.5 Measurement of Visual Exploration 68
3.2.6 Measurement of VisMET Performance 69
3.2.7 Logistic Regression Models 70
3.3 Results 71
3.3.1 Visual Exploration in Healthy Aging, Mild Cognitive Impairment, and Alzheimer’s Disease during Encoding 71
3.3.2 VisMET Performance in Healthy Aging 74
3.3.3 VisMET Is impaired in Mild Cognitive Impairment and Alzheimer’s Disease 77
3.3.4 VisMET as a screening tool for cognitive impairment and disease status 81
3.4 Discussion 85
Chapter 4: Deep convolutional neural networks and transfer learning for measuring cognitive impairment using eye-tracking in a distributed tablet-based environment 89
4.1 Introduction 90
4.2 Materials and Methods 91
4.2.1 Participants 91
4.2.2 Mobile Device Data Capture 92
4.2.3 Calibration Procedure 92
4.2.4 Visuospatial Memory Eye-tracking Task (VisMET) 93
4.2.5 Regression Tree Face and Eye Detection 94
4.2.7 Convolutional Neural Network and Support Vector Regression for Tablet-Based Gaze Estimation 96
4.2.8 Re-Calibration of Gaze Estimations Between Successive Images 97
4.2.9 EyeTribe-Based Gaze Estimation 98
4.2.10 Feature Extraction 98
4.2.11 Logistic Regression Models for Detection of Cognitive Impairment 99
4.3 Results 100
4.3.1 Performance Evaluation of Tablet-Based Methods for Gaze Estimation 100
4.3.2 Assessment of Visual Exploration on the Tablet and EyeTribe 100
4.3.3 Assessment of Cognitive Impairment on the Tablet and EyeTribe 103
4.3.4 Tablet Administration of VisMET as a Screening Tool for Cognitive Impairment 105
4.4 Discussion 107
Chapter 5: Future Directions and Conclusions 109
5.1 Summary 109
5.2 Future Directions 111
5.2.1 Visuospatial Memory Performance in Healthy Participants 111
5.2.2 Neurophysiological Mechanisms of Visuospatial Memory Recognition 112
5.3 Conclusions 113
About this Dissertation
| School | |
|---|---|
| Department | |
| Degree | |
| Submission | |
| Language |
|
| Research Field | |
| 关键词 | |
| Committee Chair / Thesis Advisor | |
| Committee Members |
Primary PDF
| Thumbnail | Title | Date Uploaded | Actions |
|---|---|---|---|
|
|
Neocortical-Medial Temporal Lobe Interactions during Visuospatial Memory Formation and Implications for Screening of Memory Impairment in Alzheimer's Disease () | 2020-04-02 18:33:56 -0400 |
|
Supplemental Files
| Thumbnail | Title | Date Uploaded | Actions |
|---|