The ability to read proficiently is key to social participation and an important premise for individual well-being and vocational success. Individuals with developmental dyslexia, a highly prevalent neurodevelopmental disorder affecting hundreds of millions of children and adults worldwide, face severe and persistent difficulties in attaining adequate reading levels. Despite years of extensive research efforts to elucidate the neurobiological origin of this disorder, its exact etiology remains unclear to date. In this context, most neuroimaging research on dyslexia in humans has focused on the cerebral cortex and has identified alterations in a distributed left-lateralized cortical language network. However, pioneering post-mortem human studies and animal models suggest that dyslexia might also be associated with alterations in subcortical sensory thalami and early sensory pathways. The largely cortico-centric view of dyslexia is due in part to considerable technical challenges in assessing the human sensory thalami non-invasively using conventional magnetic resonance imaging (MRI). As a result, the role that sensory thalami may play in dyslexia has been largely unaddressed. In this dissertation, I leveraged recent advances in high-field MRI to investigate the role of the human lateral geniculate nucleus (LGN) of the visual thalamus in adults with dyslexia in-vivo. In three multi-modal high-field MRI studies, I show that (i) dyslexia is associated with structural alterations in the direct V1-bypassing white matter pathway connecting the LGN with cortical motion-sensitive area V5/MT in the left hemisphere; (ii) the connectivity strength of which predicts a core symptom of the disorder, i.e., rapid naming ability. I further demonstrate that (iii) the two major functional subdivisions of the LGN can be distinguished non-invasively based on differences in tissue microstructure; and that (iv) adults with dyslexia show functional response alterations specifically in the magnocellular subdivision of the LGN. I also demonstrate that this subdivision deficit (v) is more pronounced in male than female dyslexics; and (vi) predicts rapid naming ability in male dyslexics only. The results of this doctoral thesis are the first to confirm previous post-mortem evidence of LGN alterations in dyslexia in-vivo and point to their relevance to key symptoms of the disorder. In synergy, our research findings offer new perspectives on explanatory models of dyslexia and bear potential implications also for prospective treatment strategies.:Contribution Statement i Acknowledgments iii Abstract v Table of Contents vii 1 General Introduction 1 1.1 Developmental Dyslexia 1 1.1.1 Diagnostic Criteria 1 1.1.2 Prevalence and Etiology 2 1.1.3 Cognitive and Behavioral Symptoms 3 1.1.4 Explanatory Models in Cognitive Neuroscience 4 1.2 Lateral Geniculate Nucleus 7 1.2.1 Anatomy and Function 7 1.2.2 Technical Challenges in Conventional MRI 8 1.2.3 High-Field MRI 9 1.3 Research Aim and Chapter Outline 10 2 Altered Structural Connectivity of the Left Visual Thalamus in Developmental Dyslexia 13 2.1 Summary 14 2.2 Results and Discussion 15 2.3 Conclusions 22 2.4 Materials and Methods 23 2.4.1 Subject Details 23 2.4.2 High-Resolution MRI Acquisition and Preprocessing 23 2.4.3 Lateral Geniculate Nucleus Definition 24 2.4.4 Cortical Region of Interest Definition 26 2.4.5 Probabilistic Tractography 27 2.4.6 Quantification and Statistical Analysis 29 2.5 Supplementary Information 30 3 Mapping the Human Lateral Geniculate Nucleus and its Cytoarchitectonic Subdivisions Using Quantitative MRI 33 3.1 Abstract 34 3.2 Introduction 35 3.3 Materials and Methods 37 3.3.1 In-Vivo MRI 37 3.3.2 Post-Mortem MRI and Histology 41 3.4 Results 44 3.4.1 Lateral Geniculate Nucleus Subdivisions in In-Vivo MRI 44 3.4.2 Lateral Geniculate Nucleus Subdivisions in Post-Mortem MRI 46 3.5 Discussion 50 3.6 Supplementary Information 54 3.6.1 In-Vivo MRI 54 3.6.2 Post-Mortem MRI and Histology 58 3.6.3 Data and Code Availability 60 4 Dysfunction of the Visual Sensory Thalamus in Developmental Dyslexia 61 4.1 Abstract 62 4.2 Introduction 63 4.3 Materials and Methods 66 4.3.1 Subject Details 66 4.3.2 High-Resolution MRI Experiments 66 4.3.3 High-Resolution MRI Acquisition and Preprocessing 67 4.3.4 Lateral Geniculate Nucleus Definition 68 4.3.5 Quantification and Statistical Analysis 69 4.4 Results 70 4.5 Discussion 75 4.6 Supplementary Information 77 4.6.1 Supporting Methods 77 4.6.2 Supporting Results 81 4.6.3 Data and Code Availability 82 5 General Conclusion 83 5.1 Summary of Research Findings 83 5.2 Implications for Dyslexia Models 84 5.2.1 Phonological Deficit Hypothesis 84 5.2.2 Magnocellular Theory 84 5.2.3 Model According to Ramus 85 5.2.4 Need for Revised Model 86 5.3 Implications for Remediation 87 5.4 Research Prospects 88 5.5 Brief Concluding Remarks 90 6 Bibliography 91 7 List of Tables 113 8 List of Figures 115 9 Selbstständigkeitserklärung 117

<p><a></a><a>The visual system is capable of interpreting immense sensory complexity, allowing us to quickly identify behaviorally relevant stimuli in the environment. It performs this task with a hierarchical organization that works to detect, relay, and integrate visual stimulus features into an interpretable form. To understand the complexities of this system, visual neuroscientists have benefited from the many advantages of using mice as visual models. Despite their poor visual acuity, these animals possess surprisingly complex visual systems, and have been instrumental in understanding how visual features are processed in the primary visual cortex (V1). However, a growing body of literature has shown that primary sensory areas like V1 are capable of more than basic feature detection, but can express neural activity patterns related to learning, memory, categorization, and prediction. </a></p> <p>Visual experience fundamentally changes the encoding and perception of visual stimuli at many scales, and allows us to become familiar with environmental cues. However, the neural processes that govern visual familiarity are poorly understood. By exposing awake mice to repetitively presented visual stimuli over several days, we observed the emergence of low frequency oscillations in the primary visual cortex (V1). The oscillations emerged in population level responses known as visually evoked potentials (VEPs), as well as single-unit responses, and were not observed before the perceptual experience had occurred. They were also not evoked by novel visual stimuli, suggesting that they represent a new form of visual familiarity in the form of low frequency oscillations. The oscillations also required the muscarinic acetylcholine receptors (mAChRs) for their induction and expression, highlighting the importance of the cholinergic system in this learning and memory-based phenomenon. Ongoing visually evoked oscillations were also shown to increase the VEP amplitude of incoming visual stimuli if the stimuli were presented at the high excitability phase of the oscillations, demonstrating how neural activity with unique temporal dynamics can be used to influence visual processing.</p> <p>Given the necessity of perceptual experience for the strong expression of these oscillations and their dependence on the cholinergic system, it was clear we had discovered a phenomenon grounded in visual learning or memory. To further validate this, we characterized this response in a mouse model of Fragile X syndrome (FX), the most common inherited form of autism and a condition with known visual perceptual learning deficits. Using a multifaceted experimental approach, a number of neurophysiological differences were found in the oscillations displayed in FX mice. Extracellular recordings revealed shorter durations and lower power oscillatory activity in FX mice. Furthermore, we found that the frequency of peak oscillatory activity was significantly decreased in FX mice, demonstrating a unique temporal neural impairment not previously reported in FX. In collaboration with Dr. Christopher J. Quinn at Purdue, we performed functional connectivity analysis on the extracellularly recorded spikes from WT and FX mice. This analysis revealed significant impairments in functional connections from multiple layers in FX mice after the perceptual experience; some of which were validated by another graduate student (Qiuyu Wu) using Channelrhodopsin-2 assisted circuit mapping (CRACM). Together, these results shed new light on how visual stimulus familiarity is differentially encoded in FX via persistent oscillations, and allowed us to identify impairments in cross layer connectivity that may underlie these differences. </p> <p>Finally, we asked whether these oscillations are observable in other brain areas or are intrinsic to V1. Furthermore, we sought to determine if the oscillating unit populations in V1 possess uniform firing dynamics, or contribute differentially to the population level response. By performing paired recordings, we did not find prominent oscillatory activity in two visual thalamic nuclei (dLGN and LP) or a nonvisual area (RSC) connected to V1, suggesting the oscillations may not propagate with similar dynamics via cortico-thalamic connections or retrosplenial connections, <a>but may either be uniquely distributed across the visual hierarchy or predominantly</a> restricted to V1. Using K-means clustering on a large population of oscillating units in V1, we found unique temporal profiles of visually evoked responses, demonstrating distinct contributions of different unit sub-populations to the oscillation response dynamics.</p>