University of Southern California Dissertations and Theses

Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

(USC Thesis Other)

Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content University of Southern California Department of Biomedical Engineering A DISSERTATION FOR THE DEGREE OF DOCTORATE OF PHILOSOPHY Characterization of Visual Cortex Function in Late-Blind Individuals with Retinitis Pigmentosa and Argus II Patients by Samantha Irene Cunningham December 2014 Dissertation Committee: Professor James D. Weiland, Ph.D. (Chair and Advisor) Associate Professor Bosco S. Tjan, Ph.D. (Advisor) Professor Krishna S. Nayak, Ph.D. (Outside Member) Copyright 2014 Samantha I. Cunningham For my Mom and Dad, Grandmommy and Granddaddy, Aunt Monie, and Fiancé, Whom I love dearly. Therefore, since we are surrounded by such a great cloud of witnesses, …let us run with endurance the race marked out for us, fixing our eyes on Jesus, the pioneer and perfecter of faith. Hebrews 12:1-2 iii | P a g e ABSTRACT Imaging studies show that vision deprivation causes the visual cortex to become responsive to tactile stimulation. The following work determined if this cross-modal activation in the primary visual cortex (V1) is correlated with vision loss in sighted individuals and late-blind patients with retinitis pigmentosa (RP)—an inherited degenerative photoreceptor disease that progressively diminishes vision—and further determined if these tactile-evoked responses were correlated with changes in resting-state and structural connectivity of the cortical visual streams. The effect of sight restoration on these tactile-evoked activities was determined by comparing V1 responses in Argus II patients with sighted and RP subjects. Based on our findings, we determined if tactile- evoked responses can account for visual performance variability among Argus II patients and gauged the feasibility of using cross-modal activation as a predictive measure for assessing outcomes in patients following a sight restoration treatment. RP patients, sighted control subjects, and recipients of the Argus II retinal prosthesis completed a series of tactile tasks in the scanner. We measured tactile-evoked blood oxygenation level dependent (BOLD) responses using functional magnetic resonance imaging (fMRI) and quantified the extent and strength of activation in V1; these quantities were evaluated as a function of visual acuity and size of the preserved visual field in V1. Subjects also completed whole-brain resting-state and diffusion tensor imaging (DTI) scans. Regions of interest (ROIs) were selected along the dorsal and ventral visual streams and both resting-state correlations and the probability of fibers connecting two regions were calculated for pairs of ROIs. iv | P a g e A highly significant effect of vision loss was found on tactile-evoked activation observed in V1: as visual acuity and visual field loss worsened, V1 became more responsive to tactile stimulation. These cross-modal responses in V1 were found to decrease after consistent use of a retinal prosthesis in one, but not the other, Argus II patient. A significant resting-state pathway was also found between primary somatosensory region S1 and V1 that is mediated by the inferior parietal lobe in RP patients. However, the overall functional or structural connectivity of the visual streams was independent of vision loss and did not change as vision worsened in our late-blind subjects. Tactile-evoked responses may be used as a cortically localized biomarker to explain individual differences in visual performance following sight recovery treatments. This would require that a relationship be established between a patient’s visual performance following treatment and the level of tactile-evoked responses in V1. If cross-modal responses are found to have a detrimental effect on adaptation to a retinal prosthesis, our results show that these responses may decrease with prolonged use of the device, a testament to the plastic nature of the adult visual cortex. v | P a g e ACKNOWLEDGMENTS The following material is based on work supported by the National Science Foundation under Grant No. EEC-0310723, the National Eye Institute under Grant No. R01 EY017707, Research to Prevent Blindness (RPB), the USC Dana and David Dornsife Cognitive Neuroscience Imaging Center, and the NSF Graduate Research Fellowship Program (GRFP). I am forever thankful to the numerous individuals whose support, kindness, knowledge, and advice made it possible for me to complete my thesis research. First to my co-advisors Dr. James Weiland and Dr. Bosco Tjan—thank you for your constant patience, wisdom, and guidance from the first year of my research until now. You not only immersed me in the field by sharing your expertise, but also instilled in me a new level of confidence, perseverance, and positivity when approaching the challenges, triumphs, and setbacks inherent in scientific research. The guidance I received from you has shaped my thesis into a body of work that I am proud of. I will take the skills and wisdom you have passed on to me (including bringing doughnuts to every major future experiment!) and use them to continue my pursuit of biomedical research with joy, and I look forward to the opportunity to pass these same lessons on to someone else. I am also grateful for the instruction given to me by my qualifying and thesis committee members Dr. Krishna Nayak, Dr. Mark Humayun, and Dr. Norberto Grzywacz, as well as the faculty members of USC’s Biomedical Engineering Department who have allowed me to pick their brains over the years on projects both related and unrelated to my thesis research. I have grown from your honest critique and the passion you have for your field. And a special thank you vi | P a g e to Mischalgrace Diasanta—you have been a constant kind and supportive face in the department. I will miss our talks and your flower pens. Thank you also to the USC Marshall School of Business Technology Commercialization Certificate Program and specifically to Dr. Kathleen Allen. Your courses introduced me to the business aspect of scientific research, while the numerous opportunities you gave me outside the classroom gave me the practical knowledge and skills I will use to one day pursue my own biomedical startup company. This came in addition to my many experiences from participating in the USC Biomimetic MicroElectronic Systems Engineering Research System (BMES ERC), from which I learned the importance of collaboration between government, industry, and academia. Thank you to Lindy Yow, Doris Lee, Diana Sabogal, Dr. Jack Whalen, and Dr. Joe Coccozza for sharing your experiences with our lab: from river rafting to navigating the NSF ERC Annual Meetings. Of utmost importance to my research, thank you to Second Sight Medical Products, Inc. for providing a device that we could modify and the resources we needed to test the system in an MRI environment. A huge thank you to Arup Roy, Punita Christopher, Richard Castro, Walter Little, Scott Loftin, and Thomas Lauritzen for lending your time and expertise to my research. Thank you to all of my current and past labmates in both Dr. Weiland’s and Dr. Tjan’s research labs. I have learned so much from all of you over the years. Your comradery during tough times and quick willingness to celebrate with each other during times of accomplishment has provided the support system I needed to grow as a researcher and individual. I will miss you all dearly, including our many side projects, runs to Sam’s lunch truck, random conversations in the kitchen, poking fun at each other’s research and bringing light-heartedness to challenging situations, and the fun we had outside of lab together. And a special thank you to Mort Arditti for your help in developing the external coil, as well as the many life lessons you shared with me over the years. I have enjoyed our many talks and am thankful for the practical electrical engineering skills I have taken away from our work together. And last but most important of all, thank you to my mom, dad, grandparents, fiancé, and the rest of my family—your sacrifice, unconditional love, support, and encouragement has made all the difference since I was a child and has shaped me into the person I am today. Perseverance through my PhD work has been possible because you were always rooting for my success. And I will always thank Jesus Christ, in whom I have placed my faith and who has given me a life filled with amazing adventures and stories to tell. I constantly strive to glorify you. vii | P a g e CONTENTS ABSTRACT ................................................................................................................................................ iii ACKNOWLEDGMENTS .......................................................................................................................... v LIST OF ABBREVIATIONS ................................................................................................................... xi LIST OF TABLES .................................................................................................................................... xiv LIST OF FIGURES .................................................................................................................................. xv PART I: INTRODUCTION ....................................................................................................................... 1 Chapter 1: Overview ................................................................................................................................... 2 1.1 Motivation ........................................................................................................................................... 2 1.2 Problem Statement .............................................................................................................................. 3 1.3 Hypotheses .......................................................................................................................................... 3 1.4 Dissertation Aims and Organization ................................................................................................... 4 Chapter 2: Background .............................................................................................................................. 6 2.1 Review of Retinal Prosthesis Technology .......................................................................................... 6 2.1.1 Overview of Retinal Degenerative Diseases ................................................................................ 6 2.1.2 Argus II Epiretinal Prosthesis System........................................................................................ 12 2.1.3 Other Visual Prostheses ............................................................................................................. 15 2.2 The Healthy Human Visual Cortex ................................................................................................... 20 2.2.1 Overview of General Anatomy .................................................................................................. 20 2.2.2. Imaging the Human Visual Cortex ........................................................................................... 25 2.3 The Human Visual Cortex and Blindness ......................................................................................... 28 2.3.1. Cross-Modal Plasticity in Blind Individuals ............................................................................. 29 2.3.2. Compensation Following Sensory Deprivation ........................................................................ 31 2.3.3. Theories on the Neural Basis of Cross-Modal Activity following Blindness ........................... 33 2.4 Implications of Cross-Modal Plasticity on Visual Prostheses .......................................................... 34 2.4.1. Insights from Vision Recovery Studies..................................................................................... 34 2.4.2. Insights from the Cochlear Implant........................................................................................... 35 viii | P a g e PART II: THE EFFECT OF BLINDNESS ON THE VISUAL CORTEX ......................................... 37 Chapter 3: Characterization of tactile-evoked V1 BOLD responses in late-blind individuals with retinitis pigmentosa ................................................................................................................................... 38 3.1 Overview ........................................................................................................................................... 38 3.2 Study Design ..................................................................................................................................... 40 3.2.1. Participants ................................................................................................................................ 40 3.2.2. Experimental Stimuli and Tasks ............................................................................................... 42 3.2.3. Image Acquisition ..................................................................................................................... 44 3.2.4. Defining Vision Loss ................................................................................................................ 44 3.2.5. fMRI Data Analysis .................................................................................................................. 45 3.2.6. Statistical Modeling .................................................................................................................. 46 3.3 Results ............................................................................................................................................... 48 3.3.1. Extent and strength of V1 BOLD responses to tactile stimulation in RP and sighted subjects 48 3.3.2. Relationship between cross-modal V1 BOLD responses and degree of preserved visual functions .............................................................................................................................................. 51 3.3.3. Comparison of tactile-evoked BOLD responses in V1 and S1 ................................................. 53 3.4 Discussion ......................................................................................................................................... 60 3.5 Conclusions ....................................................................................................................................... 62 Chapter 4: A case study of tactile-evoked V1 BOLD responses in patients with RP, AMD, and optic nerve hypoplasia........................................................................................................................................ 63 4.1 Overview ........................................................................................................................................... 63 4.2 Study Design ..................................................................................................................................... 64 4.2.1. Participants ................................................................................................................................ 64 4.2.2. Image Acquisition ..................................................................................................................... 65 4.2.3. Experimental Stimuli and Tasks ............................................................................................... 66 4.2.4. fMRI Data Analysis .................................................................................................................. 67 4.3 Results ............................................................................................................................................... 68 4.3.1. Visual Comparison of Visually- and Tactile-Evoked Responses in V1 ................................... 68 4.3.2. Comparison of Tactile-Evoked Responses in V1 to Retinotopic Maps .................................... 71 4.4 Discussion ......................................................................................................................................... 72 Chapter 5: The effect of vision loss on functional connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams ......................................... 74 5.1 Overview ........................................................................................................................................... 74 5.2 Study Design ..................................................................................................................................... 75 ix | P a g e 5.2.1. Participants ................................................................................................................................ 75 5.2.2. Image Acquisition ..................................................................................................................... 76 5.2.3. fMRI Data Analysis .................................................................................................................. 76 5.2.4. Statistical Modeling .................................................................................................................. 80 5.3 Results ............................................................................................................................................... 80 5.3.1. Task-state Correlation Between Tactile-Evoked Responses in V1 and SI ............................... 80 5.3.1. Mapping Resting-State Connectivity between V1, SI, and along the Visual Stream ............... 81 5.4 Discussion ......................................................................................................................................... 86 Chapter 6: The effect of vision loss on structural connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams ......................................... 87 6.1 Overview ........................................................................................................................................... 87 6.2 Study Design ..................................................................................................................................... 89 6.2.1. Participants ................................................................................................................................ 89 6.2.2. Image Acquisition ..................................................................................................................... 89 6.2.3. MRI Data Analysis.................................................................................................................... 90 6.3 Results ............................................................................................................................................... 92 6.3.1. Cortical Thickness of Regions along the Dorsal and Ventral Visual Streams .......................... 92 6.3.2. Analysis of White Matter Integrity using FA and MD Values ................................................. 94 6.3.3. Probabilistic Connectivity of the Visual Streams ..................................................................... 96 6.3.4. Tract-Specific Analysis of Visual Stream Connectivity ........................................................... 97 6.3.5. Comparison of Structural and Functional Connectivity along the Visual Streams ................... 99 6.4 Discussion ....................................................................................................................................... 100 PART III: THE EFFECT OF RETINAL PROSTHESIS USE ON THE VISUAL CORTEX........ 102 Chapter 7: Effect of extended Argus II prosthesis use on tactile-evoked BOLD responses in V1 .. 103 7.1 Overview ......................................................................................................................................... 103 7.2 Study Design ................................................................................................................................... 104 7.2.1. Participants .............................................................................................................................. 104 7.2.2. Image Acquisition ................................................................................................................... 105 7.2.3. Experimental Stimuli and Tasks ............................................................................................. 105 7.2.4. Experimental Procedure .......................................................................................................... 106 7.2.5. fMRI Data Analysis ................................................................................................................ 109 7.3 Results ............................................................................................................................................. 110 x | P a g e 7.3.1. Comparison of Visually and Tactile-Evoked Responses with and without Device Use ......... 110 7.3.2. Comparison of Argus II Subject Responses to RP and Sighted Groups ................................. 115 7.4 Discussion ....................................................................................................................................... 117 Chapter 8: Development and validation of a protocol for suprathreshold stimulation of the Argus II device in a 3T MRI Machine .................................................................................................................. 120 8.1 Overview ......................................................................................................................................... 120 8.2 Study Design ................................................................................................................................... 121 8.2.1. Preliminary Phantom and Canine Experiments ...................................................................... 121 8.2.2. Phantom Experiments with a Sparse Scanning Protocol ........................................................ 122 8.2.3. Validation of Sparse Scanning Protocol in a Canine Model ................................................... 124 8.3 Results ............................................................................................................................................. 124 8.3.1. Preliminary Phantom and Canine Experiments ...................................................................... 124 8.3.2. Phantom Experiments with Sparse Scanning Protocol ........................................................... 125 8.3.3. Validation of Sparse Scanning Protocol in a Canine Model ................................................... 126 8.3.4. Testing RF link between our Modified External Coil and an Argus II Patient ....................... 127 PART IV: CONCLUSION ..................................................................................................................... 128 Chapter 9: Context of results and overall conclusions ........................................................................ 129 9.1 Summary of Work and Contribution to the Field ........................................................................... 129 9.2 Suggested Future Work ................................................................................................................... 130 REFERENCES ........................................................................................................................................ 132 xi | P a g e LIST OF ABBREVIATIONS 3T 3 tesla ACPC anterior commissure and posterior commissure AD axial diffusivity ADC apparent diffusion coefficient AMD age-related macular degeneration ASR artificial silicon retina BMA blind mobility aid BOLD blood oxygenation level dependent CAS calcarine sulcus CC corpus callosum CI cochlear implant DTI diffusion tensor imaging EPI echo-planar imaging FA fractional anisotropy FDR false discovery rate FFA fundus fluorescein angiography fMRI functional magnetic resonance imaging FWER familywise error rate GLM general linear model GM gray matter ICA independent component analysis IFO inferior fronto-occipital tracts ILF inferior longitudinal tracts xii | P a g e IP inferior parietal cortex IT inferior temporal cortex LGN lateral geniculate nucleus LOC lateral occipital cortex LPZ lesion projection zone M magnocellular MD mean diffusion MPDA micro-photodiode arrays MR magnetic resonance MRI magnetic resonance imaging MT medial temporal cortex MST superior temporal areas OCT optical coherence tomography ONH optic nerve hypoplasia P parvocellular PACE prospective acquisition correction PBS phosphate buffered saline RD mean radial diffusivity RF radio frequency ROI region of interest RP retinitis pigmentosa RPE retinal pigment epithelium rs-fMRI resting-state functional magnetic resonance imaging S1 primary somatosensory cortex SC superior colliculus xiii | P a g e SLF superior longitudinal tracts SSMP Second Sight Medical Products, Inc. ST superior temporal cortex T1 longitudinal relaxation time T2 transverse relaxation time TR repetition time TE echo time UEA Utah electrode array V1 primary visual cortex V2 prestriate cortex V3, V4, and V5 extrastriate cortex VPU video processing unit WM white matter xiv | P a g e LIST OF TABLES Table 1. Subject demographics for tactile stimulation fMRI study .............................................. 41 Table 2. The effect of visual function on V1 BOLD responses after controlling for S1 BOLD responses ................................................................................................................................ 59 Table 3. Subject demographics for V1 BOLD response localization study ................................. 65 Table 4. Subject demographics for resting-state fMRI study ....................................................... 76 Table 5. List of resting-state and DTI ROIs ................................................................................. 78 Table 6. Resting-state correlations between V1 and S1 for each subject ..................................... 81 Table 7. Subject demographics for DTI study ............................................................................. 89 Table 8. Correlation of FA and MD values with visual function for select ROI pairs ................. 96 Table 9. Resting-state correlations along 4 white-matter tracts .................................................... 99 Table 10. Argus II subject demographics. .................................................................................. 104 Table 11. Summary of postoperative scanning procedure for Argus II subjects ........................ 108 xv | P a g e LIST OF FIGURES Chapter 2: Background Figure 1. Normal structure of the eye and organization of the retina ............................................. 7 Figure 2. Visual pathway from the retina to the visual cortex ........................................................ 8 Figure 3. High resolution, stained images of a normal versus degenerated retina ....................... 10 Figure 4. FFA and OCT images comparing a normal retina to one affected by wet AMD ......... 11 Figure 5. The Argus II retinal prosthesis system .......................................................................... 13 Figure 6. Subjects’ percept responses to short versus long distance electrode stimulation .......... 15 Figure 7. MPDA subretinal prosthesis .......................................................................................... 16 Figure 8. Comparison of visual prosthesis stimulation sites in the eye ........................................ 18 Figure 9. Schematic of an optic nerve visual prosthesis ............................................................... 19 Figure 10. Illustration of intracortical visual prosthesis ............................................................... 19 Figure 11. Subcortical visual pathways ........................................................................................ 22 Figure 12. Organization of the primary visual cortex ................................................................... 24 Figure 13. Illustration of dorsal and ventral visual pathways. ...................................................... 25 Figure 14. Illustration of how BOLD response maps are generated from the presentation of a stimulus .................................................................................................................................. 26 Figure 15. Illustration of reconstructed fiber tracts between regions of activation in the occipitotemporal cortex ......................................................................................................... 27 Figure 16. ICA results for low-frequency resting-state patterns in 10 subjects............................ 28 Figure 17. The four primary theories on cross-modal plasticity ................................................... 34 Chapter 3: Characterization of tactile-evoked V1 BOLD responses in RP patients Figure 18. Example of the three tactile tasks ................................................................................ 42 Figure 19. V1 BOLD responses for a healthy sighted subject to the shapes tactile task under an “eyes-closed” versus “eyes-open” condition ......................................................................... 43 Figure 20. V1 BOLD responses to the three tactile tasks in four representative RP subjects and two sighted control subjects ......................................................................................................... 49 Figure 21. Extent and strength of tactile-evoked responses in V1 ............................................... 50 xvi | P a g e Figure 22. Predictive margins from a linear mixed effects model relating visual function to tactile- evoked responses in V1 ......................................................................................................... 52 Figure 23. S1 BOLD responses to the three tactile tasks in four representative RP subjects and two sighted control subjects ......................................................................................................... 54 Figure 24. Extent and strength of tactile-evoked responses in S1 ................................................ 56 Figure 25. Predictive margins from a linear mixed effects model relating visual function to tactile- evoked responses in S1 .......................................................................................................... 57 Figure 26. Graphical illustration of pseudo-partial correlations between visual function and tactile- evoked responses in V1 and S1 across all subjects ............................................................... 60 Chapter 4: Case study of tactile-evoked V1 BOLD responses in patients with RP, AMD, and optic nerve hypoplasia Figure 27. Illustration of the checkerboard stimulus .................................................................... 66 Figure 28. Comparison of V1 responses to the sandpaper task in one RP patient and one AMD patient .................................................................................................................................... 68 Figure 29. Comparison between V1 BOLD responses to a visual stimulus and responses to the three tactile tasks in one RP subject ...................................................................................... 70 Figure 30. V1 BOLD responses to a tactile task in 1 subject with optic nerve hypoplasia, 1 RP subject, and 1 sighted control subject .................................................................................... 72 Chapter 5: Effect of vision loss on functional connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams Figure 31. Color maps depicting strength of resting state correlations between pairs of ROIs for individual subjects ................................................................................................................. 83 Figure 32. Resting-state network including all 11 ROIs among RP and sighted subjects ............ 84 Figure 33. Graphical illustration of pairwise correlations between the residual time courses of areas SI, IP, MT, and V1 ................................................................................................................ 85 Chapter 6: Effect of vision loss on structural connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams Figure 34. Comparison of cortical thickness in 2 representative RP subjects and 1 representative sighted subject ....................................................................................................................... 93 Figure 35. FA maps for 4 representative RP subjects and 2 representative sighted subjects ....... 95 xvii | P a g e Figure 36. Probabilistic connectivity of ROI pair IP—MT for 4 representative RP subjects and 2 sighted subjects ...................................................................................................................... 97 Figure 37. Tract-specific regression between FA values and preserved visual field across subjects ............................................................................................................................................... 98 Figure 38. Correlation between average resting-state connectivity and average probabilistic connectivity across ROI pairs .............................................................................................. 100 Chapter 7: Effect of Argus II prosthesis use on tactile-evoked BOLD responses in V1 Figure 39. Illustration of the cross-modal responsivity differences between sighted and low vision individuals ........................................................................................................................... 102 Figure 40. Illustration of the checkerboard stimulus presented to Argus II subjects ................. 106 Figure 41. V1 BOLD responses to the three tactile tasks for subject A1 following 4 days of consistent device use and 7 days of no device use .............................................................. 111 Figure 42. V1 BOLD responses to the three tactile tasks for subject A2 following 7 days of consistent device use and 4 days of no device use .............................................................. 112 Figure 43. S1 BOLD responses to the three tactile tasks for subject A1 following 4 days of consistent device use and 7 days of no device use .............................................................. 113 Figure 44. S1 BOLD responses to the three tactile tasks for subject A2 following 4 days of consistent device use and 4 days of no device use .............................................................. 114 Figure 45. Comparison of extent and strength of tactile-evoked responses in V1 for Argus II subject A1 to RP and sighted groups ............................................................................................... 116 Figure 46. Comparison of extent and strength of tactile-evoked responses in V1 for Argus II subject A2 to RP and sighted groups ............................................................................................... 117 Chapter 8: Development and validation of a protocol for suprathreshold stimulation of the Argus II device in a 3T MRI Machine Figure 47. Illustration of a simulated Argus II implant and MRI experimental setup ................ 122 Figure 48. Sparse temporal scanning protocol sequence of events ............................................ 123 Figure 49. Phantom experiment anatomical images of simulated implant ................................. 125 Figure 50. Structural and functional MR images of a canine implanted with the Argus II ........ 125 Figure 51. MR image results of phantom experiment with sparse scanning protocol ................ 126 Figure 52. Light-evoked fMRI BOLD signal acquired from a canine implanted with the Argus II ............................................................................................................................................. 127 1 | P a g e PART I: INTRODUCTION The story of sight restoration is one that continues to be written. It is bursting with challenges, triumphs, and set-backs as we seek to mold prosthetic devices to our understanding of the human visual system. As our knowledge of this system and its response to external sensory devices increases, vision restoration treatments will inevitably improve. The Argus II epiretinal prosthesis is one such system that provides blind patients with a low level of vision restoration. However, clinical trial outcomes have shown that visual performance is variable among recipients of the device. Much headway has been made in understanding how vision prostheses interact with degenerate retinas, but little is known about the effect of artificial vision on the cortex. In order to account for the cortical contribution to this variability, it is necessary to first gain a solid understanding of how certain types of blindness come about, their effect on the brain, and how these effects might interact with vision prostheses. 2 | P a g e Chapter 1: Overview 1.1 Motivation Performance results from the ongoing Argus II clinical trial have shown that the majority of recipients are able to better perceive light when using the device. However, a minority of patients do not show the same improvement in visual performance as their peers—while some patients improve from minimal light perception to perceiving hand motions, others progress further to counting individual fingers. There are several possible reasons for this variability in performance, including: inconsistencies in how the device interfaces with the retina, differences in retinal degeneration across retinitis pigmentosa (RP) patients, and difficulties in completing the visual tasks during clinical trial testing. The following project sought to address variability in Argus II patient performance from a cortical point-of-view. How does the structure and function of the visual cortex change as vision loss progresses in RP patients? How do these changes impact visual performance following a sight restoration treatment? Are these changes reversible with the level of vision restoration provided by a retinal prosthesis? Does cortical plasticity in blind individuals improve or hinder their visual performance with the device? How does the visual cortex respond to stimulation from the device? Answers to these questions may help provide an additional explanation for why some individuals adapt to the device better than others. Sight restoration procedures are often financially and time-consuming endeavors that carry some degree of medical risk to the patient. While the benefits typically far outweigh the costs 3 | P a g e (Vaidya et al., 2014), there is a need to better identify which patients have a greater likelihood of increased visual function following treatment. If a relationship is found between the cortical changes associated with vision loss and an individual’s visual performance with the Argus II device, these functional and structural changes may be used as a measure to help predict whether a patient will be able to sufficiently respond to the device’s electrical signal and adapt to the treatment. Such a measure may also be translate into predicting patient outcomes following other forms of sight restoration treatment. 1.2 Problem Statement The novelty of retinal prosthesis technology has produced several new questions: while studies have demonstrated the ability of these devices to restore a low level of vision to blind patients, little is known about the effect of vision restoration on the visual cortex. Here, we sought to determine how cross-modal responses in the primary visual cortex and connectivity along the visual streams are affected by vision loss in adulthood and by vision restoration. This will provide a first step towards addressing performance variability among retinal prosthesis recipients and predicting outcomes in sight restoration treatment candidates. 1.3 Hypotheses Previous literature has demonstrated that several functional and structural changes occur in late-blind individuals, including an increase in cross-modal responses in the visual cortex and decreased integrity of the optic radiation. We expect our findings to support these results by finding a consistent correlation between vision loss and visual cortex function: as vision loss becomes more severe, RP patients will exhibit increased tactile-evoked responses in primary visual cortex (V1) and little-to-no decrease in white matter connectivity along the visual streams. Prolonged use of an Argus II device will reverse these cross-modal responses, so that Argus II users will have decreased tactile-evoked responses in V1 when compared to their late-blind RP counterparts. As a predictive measure, we expect to find that individuals with more extensive cross-modal responses in V1 prior to treatment will have better visual task performance results with the Argus II than those who had less cross-modal activity. This latter hypothesis is based on findings from cochlear implant (CI) studies showing that postlingual deaf patients with greater cross-modal (audio-visual) 4 | P a g e activity in the visual cortex were more successful at rehabilitation with a CI—higher performers utilized this enhanced cross-modal activity to compensate for an imperfect implant signal. 1.4 Dissertation Aims and Organization To address these hypotheses, we investigated a series of aims that have been organized in this dissertation as follows: Chapter 2 provides an overview of retinal prosthesis technology and the human visual system from the retina to the cortex. A review of current literature is presented describing how blindness affects the visual areas of the brain and several neuroimaging techniques used to evaluate these changes. Chapter 3 describes our method of evoking visual cortex responses in blind individuals. We used fMRI to characterize how tactile-evoked V1 blood oxygen level dependent (BOLD) responses are modulated by vision loss in late-blind patients with RP and sighted control subjects. Chapter 4 presents a case study exploring a possible relationship between the location of tactile- evoked activity in V1 and the location of visual field loss in patients with RP, age-related macular degeneration (AMD), and optic nerve hypoplasia. Chapter 5 begins an investigation of the source of V1 cross-modal activity in late-blind RP subjects using resting-state fMRI. We determined the effect of blindness on functional connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams. Chapter 6 further determines the effect of blindness on the structural connectivity of the visual streams in RP and sighted subjects using DTI. Functional connectivity was compared with the integrity of white matter connections between these regions. Chapter 7 addresses how tactile-evoked responses in V1 change once vision is restored in a case study of two Argus II patients. Chapter 8 describes our development of an Argus II external coil that is functional in a 3T MRI 5 | P a g e machine, as well as a novel method that allows for suprathreshold stimulation with the device during scanning without compromising device functionality or image quality in phantom and canine models. Chapter 9 summarizes our findings, contributions to the field, and suggested future work. 6 | P a g e Chapter 2: Background 2.1 Review of Retinal Prosthesis Technology There are currently more than 285 million visually impaired individuals worldwide, including 39 million blind and 246 million low vision patients (Congdon et al., 2011). In the United States alone, there are 3.4 million individuals over 40 years old whose visual impairment cannot be corrected with glasses or contact lenses (Chou et al., 2013). Visual impairment can arise from a range of pathologies including cataracts, glaucoma, macular degeneration, and diabetic retinopathy—the foremost causes of blindness and low vision in the U.S. (WHO 2012, NIH 2008). Recent therapeutic research has targeted retinal degenerative diseases for the development of neuroprotection strategies, gene therapy, and retinal prostheses. Disorders like age-related macular degeneration and retinitis pigmentosa follow a degenerative process that preferentially affects the retinal layers of the eye, allowing for targeted intervention within the retina. Other diseases allow for treatments targeting the optic nerve or visual cortex. 2.1.1 Overview of Retinal Degenerative Diseases Normal Retina Morphology The human retina is composed of five major classes of neurons organized into layers that line the eyeball wall (Figure 1, Left). These neurons serve to convert light into electrical signals that are sent to the visual cortex. Light coming into the eye initially stimulates photoreceptor neurons, which are contained within an outer nuclear layer at the posterior end of the retina. The 7 | P a g e following inner nuclear layer is comprised of retinal interneurons (horizontal, bipolar, and amacrine cells), while a subsequent ganglion cell layer transmits the final output of the retina to the cortex via the optic nerve. Photoreceptors, horizontal cells, and bipolar cells all form synaptic connections with one another in the outer plexiform layer, while bipolar, amacrine, and ganglion cells connect within the inner plexiform layer. In back of the retina lies the retinal pigment epithelium (RPE), a single layer of pigmented cells that acts as a blood/retina barrier and helps maintain visual function by: absorbing light energy focused on the retina, delivering nutrients from nearby blood vessels to the photoreceptors, maintaining the photoreceptor outer segments, and secreting immunosuppressive factors into the eye. The RPE sits atop a vascular choroidal layer that is located in front of the white, fibrous sclera that forms the outermost eye wall. Figure 1. Left: Normal structure of the eye and organization of the retina 1 . Light enters the eye through the lens and travels to the outer layer of the retina, where it is absorbed by rod and cone photoreceptors. A phototransduction cascade ensues which transduces light energy into neural activity, allowing interneurons to transmit and integrate the photoreceptor signal. The final output of the retina is conveyed by the ganglion cell layer to the cortex via the optic nerve. Right: Structure of healthy rod and cone photoreceptors 2 . Light enters through the inner segment of the photoreceptor and is funneled to the outer segment where it is absorbed by photopigment rhodopsin. Light is focused by the cornea and lens onto the retina and detected by photoreceptor cells in the outer nuclear layer. These cells are of two types: cones mediate day and color vision and are primarily located in the macula (foveal) region of the retina, while the more numerous and sensitive rods are responsible for night vision and are distributed peripherally. Normal rod and cone photoreceptor cells contain an outer segment of photosensitive discs that absorb light energy (using 1 webvision.med.utah.edu 2 https://www.stanford.edu/group/vista/cgi-bin/FOV/ 8 | P a g e photopigment rhodopsin for rod photoreceptors and photopsins for cones) and transduce it into neural activity in the inner segment (Figure 1, Right). Exposure of these discs to light triggers a biochemical phototransduction cascade that hyperpolarizes the photoreceptor and results in neural impulses that are transmitted throughout the other layers of the retina. Horizontal, biopolar, and amacrine interneurons normally integrate photoreceptor signals to convey temporal and spatial information to the ganglion cell layer, which in turn conveys information about the visual image to the cortex. This process is supported by cytoskeletal proteins, regular phagocytosis of the discs by retinal pigment epithelium (RPE), and trafficking of newly synthesized proteins from the rod inner segment to the outer segment. Figure 2. Left: Visual pathway relating how images are translated from the visual field to the retina and visual cortex 3 . 3 <http://www.studyblue.com/notes/note/n/movesci-quiz-2/deck/3676042> 9 | P a g e Light from the visual field is arranged on the retina in a highly organized manner. The retinal field is divided into nasal (medial to the fovea) and temporal (lateral to the fovea) hemifields, which are further divided into dorsal and ventral quadrants (Figure 2). These regions allow for organized processing of the binocular and monocular regions of the visual field. Light reaching both eyes from the central visual field is projected onto the retina based on these four regions: the left visual field projects onto the nasal hemiretina of the left eye and temporal hemiretina of the right eye, while the right visual field projects onto the nasal hemiretina of the right eye and temporal hemiretina of the left eye. Light originating from the monocular regions of the visual field is projected onto the ipsilateral side of the retina in the eye on the same side (e.g. light from the right monocular zone reaches the right eye). Information from each retinal quadrant exits the eye through the optic disc via ganglion cell axons and is projected to subcortical regions of the brain. Retinitis Pigmentosa Retinitis pigmentosa (RP) is an inherited pigmentary retinopathy affecting 1 in 4000 individuals worldwide that results in progressive peripheral vision loss and eventual blindness over the course of decades (Hamel et al., 2006). The typical form of RP (also known as rod-cone dystrophy) is characterized by initial degeneration of photoreceptor rods followed by loss of cone cells—a sequence which causes night blindness to develop before patients experience visual impairment in daylight. Early stage night blindness gives way to photophobia, retinal pigment deposits, decreased visual acuity, and mild peripheral scotomas during mid-stage RP. These symptoms are exacerbated in the final stages, with patients suffering from severe peripheral vision loss, intense photophobia, and eventual loss of foveal vision. Most RP patients maintain minimal light perception following loss of central and peripheral vision. An individual may be diagnosed with early-onset RP if symptoms reach mid-stage by the age of two; this is often indistinguishable from Leber’s congenital amaurosis (LCA), a disease that shares several genetic mutations with RP (including mutation of RPE65, CRB1, CRX and TULP1 genes; Hamel et al., 2006). In contrast, late-onset RP is characterized by early-stage symptoms that develop at or after mid-life. Degeneration in RP is caused by mutations in genes that are expressed in the photoreceptors and surrounding tissues. These include genes that encode proteins for rod visual pigment and rod visual transduction (Rhodopsin), retinal pigment epithelium (RPE), photoreceptor differentiation (NRL, NR2E3, CRX), and other genes that support cell cytoskeleton, trafficking, 10 | P a g e and metabolic processes. Mutations in these proteins destabilize the photoreceptor outer segment and prevent visual transduction. The result is a degenerated photoreceptor cell layer that can no longer absorb light and transmit visual information to other layers of the retina (Figure 3). Initial loss of rod cells forms the basis for the gradual tunnel vision characteristic of RP. Figure 3. High resolution, stained images of a normal versus degenerated retina (rods are shown in green). Left: Layers of a normal retina are organized and differentiable, with the outer and inner segments of rods and cones intact. Right: RP causes deterioration of rods cones and disorganization of the photoreceptor (outer nuclear) layer, including invasion of photoreceptors into other layers of the retina. 4 No treatments currently exist for humans that prevent, stop, or restore degeneration of the photoreceptor layer. Instead, treatments for RP patients focus on slowing down vision loss through the use of light protection (sunglasses) and Vitamin A and E therapy. Other treatments for preserving visual function are being explored in animal models and include gene therapy, pharmacological treatments, and neuroprotection using growth factors, rod-derived cone viability factor, and antiapoptotic factors (Hamel et al., 2006). Sight restoration techniques are also being developed that are specific to RP pathology, including cell or tissue transplantation and retinal prosthetic devices. Age-Related Macular Degeneration Age-related macular degeneration (AMD) is a highly prevalent retinal degenerative disease that is the leading cause of blindness among the aging population, affecting 1.75 million adults in the United States alone (Freidman et al., 2004). AMD manifests itself in wet or dry forms: symptoms of dry AMD include gradual loss of central vision, while the onset of wet AMD is 4 < http://www.nei.nih.gov/eyeonnei/snapshot/archive/0909.asp> 11 | P a g e relatively more abrupt and includes development of a central scotoma, loss of central vision, and object distortions. The majority of late-stage AMD patients experience complete loss of foveal vision while retaining some peripheral vision. Onset of AMD begins with the development of hard, soft, and Reticular drusen that forms following the accumulation of extracellular matrix that is normally phagocytosed by RPE cells (Khandhadia et al., 2012). Dry AMD may also be accompanied by choriocapillaris atrophy and deterioration of the RPE. The resulting geographic atrophy can lead to gradual loss of vision and complete blindness in its most advanced form. The sudden decrease in vision that is characteristic of wet AMD manifests from growth of abnormal blood vessels within the retina or from the choroidal layer under the retina. Sub-retinal, sub-RPE, or intra-retinal hemorrhaging can then occur, leading to fibrosis and macular scarring (Figure 4). Thickening of the macula and retinal distortion further contribute to vision deterioration in wet AMD patients. Figure 4. Fundus fluorescein angiograms (FFA) and optical coherence tomography (OCT) images comparing normal (top) and wet AMD (bottom) images (Khandhadia et al., 2012). A: Normal FFA of macula. B: Normal OCT of macula. C: FFA of macula in a wet AMD patient, with hemorrhaging from choroidal layer shown in center. D: OCT of retina in a wet AMD patient, showing intraretinal and sub- retinal fluid-filled cysts. Similar to RP, there is currently no treatment available to prevent AMD. However, injections of vascular endothelial growth factor (VEGF) inhibitors into the vitreous of the eye have recently been shown to stabilize and often restore sight in wet AMD patients by slowing the growth of abnormal blood vessels. Removal of blood from the retina and RPE can also help prevent cell 12 | P a g e death and permanent macular scarring. The Age-Related Eye Disease Study has found that dry AMD patients can slow the progression of vision loss by taking high doses of zinc and vitamins A, C, and E, and by decreasing their risk for cardiovascular disease (by reducing blood pressure, smoking, obesity, and cholesterol levels; Age-Related Eye Disease Study Research Group et al., 2001). 2.1.2 Argus II Epiretinal Prosthesis System It has long been shown that electrical stimulation of the eyeball surface and retina evokes visual phosphenes in sighted individuals and RP patients (Brindley et al., 1955; Brindley et al., 1964; Humayun et al., 1996; Potts et al., 1968). Visual prosthesis technologies are currently being developed that attempt to restore vision to visually impaired patients by stimulating some component of the visual system. Among these are retinal prostheses created for patients suffering from retinal degenerative diseases like retinitis pigmentosa. While RP leads to the degeneration of photoreceptor cells, inner retinal cells remain functional and may be stimulated to evoke an electrical response. The Argus II retinal prosthesis system is one such technology developed by Second Sight Medical Products, Inc. that electrically stimulates these remaining retinal cells to produce phosphenes and partially restore vision in RP patients. Description of the Device The Argus II epiretinal prosthesis (Humayun et al., 2012) is a device consisting of a small glasses-mounted video camera that captures and transmits video to a video processing unit (VPU) worn on a belt or shoulder strap (Figure 5, Top). The VPU then converts the image into electrical signals that are conveyed to a transmitter coil fixed to the glasses. This external radiofrequency (RF) coil wirelessly transmits image data and power to a receiving coil that, along with an electronics case, is fixed to the sclera around the eye (Figure 5, Bottom Left). The electronics case is connected to a 6x10 electrode array (covering a 20 degree visual field) via a polymer ribbon cable that is inserted through the sclera, allowing the electrode array to be fixed to the macula region of the retina (contacting the ganglion cells and their axons) using a retinal tack (Figure 5, Bottom Right). The number of enabled electrodes typically varies from 46-60 and depends on how well each electrode meets conservative electrical stimulation criteria. Image data (in the form of electrical signals) is transmitted through the ribbon cable to the microelectrode array, which in turn stimulates retinal cells by emitting spatially and temporally controlled electrical pulses based on 13 | P a g e the image recorded by the camera. If the stimulations contain enough electrical energy to excite the retinal cells, the signal is then processed by the retina and sent through the optic nerve to the cortex, where the signal is interpreted as a phosphene. Figure 5. The Argus II Retinal Prosthesis System. Top: Photo showing the camera-mounted glasses with external RF coil and video processing unit (VPU) 5 . Bottom Left: Implantable component of the device including electronics case, internal RF coil, and 6x10 electrode array. Bottom Right: Fundus image of electrode array tacked to macular region of retina (Humayun et al., 2012). Argus II Clinical Trial Outcomes A feasibility study conducted by Second Sight tested the visual acuity and navigation abilities of 30 RP patients following implantation (Humayun et al., 2012). While all subjects reported seeing percepts in response to electrical stimulation with the device, their performance during the tasks varied. During a square localization task in which subjects had to identify a white square on a black background, 27 of 28 subjects (96%) performed better with the system on versus off, and no subjects performed significantly better with the system off. When describing the path of a white line moving across a black background, 16 of 28 subjects (57%) performed the test 5 <www.2-sight.eu/en> 14 | P a g e better with the system on versus off, while one subject performed the task significantly better with the system off. A grading visual acuity test was conducted by asking subjects to differentiate the orientation of black and white bars with a range of widths—7 subjects improved their visual acuity from worse than 2.9 logMAR to between 2.9 and 1.6 logMAR when the device was on. 22 subjects further completed a letter identification task, where individuals were asked to identify letters in the full alphabet or in a limited letter set. 6 subjects were able to identify any letter of the alphabet after 100 seconds at a 63.5% success rate (vs. 9.5% with the device off). All 22 subjects were able to correctly identify a small set of 8 letters after 44 seconds at a 72.5% success rate (vs. 16.8% with the device off). These performance scores demonstrate that letter identification abilities can be restored with a retinal implant if the optimal degree of "image magnification" is used (Fornos et al., 2011). Argus II patients demonstrated a similar improvement in performance during mobility tasks over a 2 year period: during two tasks (finding a door across a room and following a white line), subjects performed significantly better with the device on. The variability in success rates among patients may have been due in part to changes in task difficulty part way through the trial. However, it remains uncertain as to why patients differed in their overall visual task performance (e.g. some subjects improved from light perception to perceiving hand motions, while others improved to counting fingers). Visual percepts differ depending on whether retinal ganglion cell bodies are stimulated versus their axons (subjects reported seeing spots of light in the first case and elongated phosphenes in the second). A study by Nanduri et al., 2012 demonstrated that the size, shape, and brightness of these phosphenes are dependent upon stimulation frequency and amplitude, so that an optimal pulse train amplitude and frequency may allow fine-tuning of an individual’s percepts. A study by Lauritzen et al., 2011 was conducted to determine if Argus II patients see one or two phosphenes when two electrodes are stimulated together. Electrode pairs of different spacing were stimulated either individually or together, after which subjects were asked to draw the perceived phosphene on a touch screen. 25 electrode pairs with inter-electrode separation from 525μm (adjacent electrodes) to 5405μm (longest diagonal of array) as well as the corresponding single electrodes were stimulated for each subject. Results demonstrated that small inter-electrode distances yielded a single phophene percept response, while longer inter-electrode distances yielded two distinct phosphenes (Figure 6). Single electrode stimulation tended to yield single phosphene percepts. It was further found that basic visual stimulus discriminability correlated well 15 | P a g e with visual performance in Argus II subjects, including grating visual acuity (subjects’ ability to detect the orientation of a grating displayed for 5 seconds) and ability to identify the direction of motion of a bar moving across the screen. Figure 6. Subjects percept responses to short (left) versus long (right) distance electrode stimulation. (Lauritzen et al., 2011) Current challenges An epiretinal design contains several limitations: in addition to difficulties in controlling the types of phosphenes generated during stimulation, use of an extraocular camera requires that individuals move their head to change the camera position instead of their eye, leading to possible disorientation. The interface between the electrode array and retina similarly poses some challenges. Pressure from the array could cause harmful side effects, including dislodgment of the array during rapid eye movement, erosion of the retina, dislodged retinal tacks, inflammation, and retinal tears and detachment (Humayun 2012). 2.1.3 Other Visual Prostheses Several other approaches to visual prosthetic devices are being investigated and are summarized as follows. These methods attempt to create artificial vision by placing a stimulating electrode array between the RPE and inner nuclear layers (subretinal), between the sclera and choroid (suprachoroidal), on the optic nerve, or in the cortex (intracortical). Subretinal Prostheses While an epiretinal approach to phosphene generation has been widely successful, others have developed a subretinal prosthesis concept that replaces the photoreceptor layer by processing 16 | P a g e light instead of camera images. This method requires lower stimulating currents since the implant is surgically placed in the subretinal space. This design targets individuals with retinal degenerative diseases and is being accomplished in several different ways. The Tübingen Alpha IMS project utilizes a 3.0 x 3.1 mm micro-photodiode array (MPDA) composed of 1500 silicon-based micro-photodiodes, each with individual stimulating electrodes and amplifiers that generate a pattern of 38 x 40 pixels (Schwahn et al., 2001; Zrenner et al., 2011; Zrenner et al., 2012). The MPDA is implanted into the subretinal space—the area separating the RPE and photoreceptor layer—such that the electrodes contact the degenerated retina (Figure 7). Figure 7. MPDA Subretinal Prosthesis. “(a) The cable from the implanted chip in the eye leads under the temporal muscle to the exit behind the ear, and connects with a wirelessly operated power control unit. (b) Position of the implant under the transparent retina. (c) MPDA photodiodes, amplifiers and electrodes in relation to retinal neurons and pigment epithelium.” (Zrenner et al., 2011). Instead of bypassing damaged photoreceptors by collecting image data from a camera, the photodiodes collect light that reaches the back of the retina and transforms it into localized electrical currents that are proportional to the light source. This signal is amplified using an additional energy source (such as an external power supply) to ensure that these electrical currents stimulate bipolar and/or amacrine cells, and propagate to the ganglion cell layer which carries the signal to the visual cortex. Preclinical studies in the Yucatan micropig and rabbit demonstrated that subretinal electrical stimulation leads to evoked potentials in the cortex that were comparable to visual evoked potentials (Schwahn et al., 2001). A clinical study in 3 RP and choroideraemia patients implanted with the MPDA device was conducted to determine their visual function. Two subjects were able to identify grating patterns while all 3 subjects were able to locate bright objects on a dark table. One of these subjects correctly identified objects on the table (including forks and 17 | P a g e geometric patterns) with 15% contrast, read larger letters, and localized other people in a room (Zrenner et al., 2011). An artificial silicon retina (ASR) has also been developed (Chow et al., 2010) that is similar to the Alpha IMS device: it is composed of photovoltaic, silicon chip micro-photodiodes which each contain their own capacitive electrode and serves to replace degenerate photoreceptors. The device produces stimulating electrical currents that either induces hyperpolarization of photoreceptor outer segments when exposed to light or induces depolarization in darkness. This allows patients to experience images composed of both light and darkness information. Unlike the Alpha IMS device, the ASR is powered by incident light only and does not require use of an additional power source. Clinical studies in 6 implanted RP patients demonstrated that the device improved and/or slowed vision loss in all subjects, including improvements in visual acuity and the perceptions of color, contrast, and darkness (Chow et al., 2010). Suprachoroidal Prostheses A suprachroidal-transretinal stimulation system has been developed for indirectly stimulating retinal cells, an approach that eliminates risks due to direct attachment of electrodes to the delicate retina. An example of this is a suprachoroidal device designed by the Bionics Institute in Melbourne, Australia (Cicione et al., 2012; Shivdasani et al., 2012). The device consists of a 7 x 12 array of 400µm diameter electroplated platinum electrodes that is inserted into the scleral pocket (Figure 8). Pre-clinical studies in 6 normally sighted anesthetized cats revealed that suprachoroidal stimulation successfully evoked responses in primary visual cortex. A pilot study of the suprachoroidal implant in RP and choroideremia patients—led by the Center for Eye Research Australia—is currently underway. 18 | P a g e Figure 8. Comparison of visual prosthesis stimulation sites in the eye (Zrenner et al., 2012). Optic Nerve Prosthesis The optic nerve serves as the only outlet for transmitting information from the retina to the brain. It is composed primarily of axons from retinal ganglion cells that terminate in the lateral geniculate nucleus (LGN), from which information is transmitted to the visual cortex. Since the optic nerve is unaffected by the majority of degenerative eye diseases, it serves as a promising point of intervention for sight restoration. Electrical stimulation of the optic nerve has been shown to produce phosphenes in one RP patient (Veraart et al., 2004). A prosthesis has thus been created that uses a penetrating multi-electrode array wrapped around the optic nerve for direct stimulation of ganglion axons (Figure 9). Images are captured by an intraocular micro-camera placed in the lens of the blind eye. The resulting image data is processed using a feature-extraction algorithm and is transformed by a retinal encoder into a spatiotemporal pattern of stimulating electrical pulses. The penetrating electrode array delivers the pulses to the optic nerve, and the signal is propagated to the visual cortex. Electrophysiological experiments in rabbits have confirmed that the prosthesis can generate visual cortex responses via direct electrical stimulation (Chai et al., 2008). 19 | P a g e Figure 9. Schematic of an optic nerve visual prosthesis (Chai et al., 2008). Intracortical Prosthesis In addition to stimulating retinal cells and ganglion axons, studies have shown that direct electrical stimulation of the visual cortex (specifically neurons in V1) can produce visual percepts (Brindley et al., 1968; Dobelle et al., 1974; Schmidt et al., 1996). Several devices are currently in development that utilize intracortical stimulation as an alternative solution to sight restoration (Figure 10). Among these are Utah Electrode Arrays (UEAs) developed by Torab et al., 2011, which consist of 100 1-mm long penetrating microelectrodes spaced 400 μm apart on a 10 × 10 array. These UAEs are implanted directly into the cortical layers of V1. A stimulator battery is used to deliver constant-current pulses to V1 of non-human primates using stimulation parameters similar to those found to evoke phosphenes in human and macaque monkey subjects. Experiments demonstrated that inducing phosphenes using penetrating microlectrodes depended on the location of the electrode tips in V1 and the distance of the tips from the neuronal somas (Torab et al., 2011). Figure 10. Left: Illustration of intracortical visual prosthesis 6 ; Center: Top view of implant; Right: Side view of implant. 6 <http://neural.iit.edu/research/visual_prosthesis/> 20 | P a g e The ultimate goal of intracortical prostheses, however, is to generate percepts in real time as a means of providing artificial visual input, a feat which requires use of an image processing system that can convert real-time images from a camera into cortical stimulation patterns that may be interpreted by the implantee. One such system is being developed by Srivastava et al., 2005 which consists of three parts: an image detection system, image processing system, and an interface between the image processor that will generate stimulation patterns and deliver them to the implanted microelectrode arrays using a transcutaneous telemetry link. An intracortical prosthesis design by Mohammadi et al., 2012 goes further to incorporate a blind mobility aid (BMA) system that estimates the distance of objects within 7.5m of the device and alerts the blind person to the presence of the closest object. An image processing system similar to that of Srivastava et al., 2005 is used to extract salient information from the camera’s images and provide the appropriate stimulation commands—this serves to supplement the otherwise limited number of phosphenes typically evoked by the prosthesis. 2.2 The Healthy Human Visual Cortex The seeming simplicity of seeing is based on a series of highly complex processes that begin in the retina and are developed in the visual cortex and through its interactions with other regions of the brain. 2.2.1 Overview of General Anatomy Subcortical Visual Pathways Visual field information leaving the retina through the bilateral optic nerves is immediately projected to three primary subcortical regions in the brain: the superior colliculi (SC), pretectum, and LGN (Figure 2, Left). Both the SC and pretectum are necessary for imperceptible residual vision (also known as “blindsight”), while visual perception is largely mediated by the LGN. The optic nerves from each eye converge at the optic chiasm, where signal from the retina is distributed into one of two optic tracts that cross to the opposite side of the brain (Figure 2). Retinal ganglion cell axons project directly onto the SC, a layered structure located in the midbrain that also receives inputs from the cortex. The superficial layers gather sensory information from the retina and visual cortex and project it to the cerebral cortex, while deeper layers receive more multisensory input (including somatosensory and auditory information) from 21 | P a g e other regions of the brain. In this way, the superior colliculus provides a pathway for signals to travel from the retina to the cortex. The SC further contains a topographic map of the space around an individual, including a map of the visual field and maps based on other sensory modalities. SC neurons in the deeper layers control saccadic eye movements that help to shift an individual’s gaze to a region of the visual field, based on cortical inputs into the intermediate layers. This helps to direct behavioral responses towards places of interest around an individual. Information from the retina is also sent to the pretectum of the midbrain, which projects the signal to preganglionic parasympathetic neurons. These cells, located next to oculomotor neurons, help to control pupillary responses to incoming light. Visual information traveling to the visual cortex primarily converges on the lateral geniculate nucleus, the termination point for 90% of retinal ganglion cell axons. These axons terminate in a manner that maintains a retinotopic representation in the LGN: adjacent regions on the retina project to adjacent regions in the LGN. Since the highest concentration of ganglion cells is located in the fovea of the retina, the LGN more largely represents the foveal region of the retina than the periphery. This difference in central and peripheral representation in the retina (i.e. visual field) versus cortex is described by a cortical magnification factor that accounts for changes in topographic representation of visual input as it travels throughout the cortex. The LGN is highly organized into two ventral (i.e. magnocellular) and four dorsal (i.e. parvocellular) layers (Figure 11): M ganglion cells terminate ventrally, while P ganglion cells terminate in dorsal layers. Information from the right hemiretinal field is projected to layers of the right LGN, while input from the left hemiretinal field is projected to the left LGN (in general, each hemiretina projects to two ventral layers and one dorsal layer). Together, these inputs form a complete representation of the visual field and information from each layer is projected to the primary visual cortex. Both M and P pathways remain segregated in the visual cortex and contribute to different aspects of vision depending on how they integrate input from rod and cone photoreceptors: P cells mediate color, high spatial and low temporal vision. M cells facilitate low spatial and high temporal visual stimuli. 22 | P a g e Figure 11. Parallel pathways from the retina to the cortex via the LGN (Nassi et al., 2009) While the LGN is the main terminus for retinal input, the LGN itself receives feedback connections from many other regions of the brain, including the brain stem and cortex. These additional inputs may help to modulate information traveling from the retina to the visual cortex. The Visual Cortex Visual information travels from the LGN to the primary visual cortex (V1) via the optic radiation fiber tract in both hemispheres (Figure 2). Similar to the LGN, each hemisphere of V1 processes information from the contralateral visual field. Also known as the striate cortex, area V1 is a 2mm thick region of the visual cortex containing around 200 million neurons that are arranged into six distinct anatomical layers (Figure 11). Inputs from the LGN primarily terminate in layer 4, characterized by the stria of Gennari (for which the striate cortex derives its name)—a massive stripe of afferent fibers from the LGN. M and P cell axons terminate in separate sublayers, thus maintaining M and P pathway segregation in V1. Axons from intralaminar areas of the LGN terminate in striate layers 2 and 3. Input from the LGN is initially distributed to the cortex by non- pyramidal cells, a type of local cortical neuron characterized by their small somas and spiny or smooth dendrites that project to other regions of V1. Information is then integrated across the striate layers (including layers 5 and 6) by pyramidal cells, which are characterized by their large somas and long dendrites that project to other brain regions. In addition, V1 layers output to extrastriate cortex (i.e. the higher visual areas including V2, V3, V4, and area MT) and subcortical areas, including back to the LGN and SC. 23 | P a g e The layers of striate cortex are topographically organized according to how V1 neurons are tuned to different features (such as stimulus size, orientation, motion, color, location, and eye that the stimulus is presented to). V1 is organized into a retinotopic map that preserves the relative spatial relationship of the retina (where each hemisphere is organized into a retinotopic map of half of the contralateral visual field). Similar to the LGN, this retinotopy is distorted according to a cortical magnification factor such that the foveal region of the retina is represented on a greater area of V1, while the periphery has a smaller representation. V1 is further organized into functional modules that consist of a set of individual columns, each of which processes specific features in the visual field (Figure 12); the columns are 30- 100µm wide and 2mm deep and span the region between the pial surface and white matter of the brain. These include alternating ocular dominance columns, where neurons that receive input from a particular eye are arranged together in the same column. Each column thus responds best to stimuli presented to either the left or right eye and is crucial in processing binocular interactions. On a finer scale, V1 is organized into hypercolumns that traverse all six layers. Each hypercolumn responds to input from one eye and from a similar region of the visual field and is composed of several orientation columns—columns that contain neurons tuned to one particular stimulus orientation. Interspersed among these columns are groups of cells in layers 2-3 called blobs and interblobs which process color and form, respectively. One functional module is thus able to process all information about a visual stimulus. In general, neurons in different columns that respond to similar stimuli (such as a similar axis of orientation) are linked by horizontal connections within the same layer. 24 | P a g e Figure 12. Organization of the primary visual cortex 7 Following area V1, information is projected to more specialized regions of the visual cortex that are categorized as the extrastriate cortex (also known as V2-V5 or visual association cortex). V1 primarily projects to area V2, which is organized into thin, thick, and pale stripes: thin stripes received input about color from blobs in V1 layers 2-3, interblobs provide input about form to the pale stripes, and V1 layer 4B provides motion information to the thick stripes. While cross- connections exist between the stripes, both M and P pathways are partially preserved through V2. Area V3 also receives input from both M and P pathways and has projections to both the middle temporal area (MT) and V4, and as such has been implicated in integrating and transforming visual signals (specifically processing color and complex motion information). Together, areas V2/V3 are important for the perception of first-order motion (motion defined by a change in luminance). Area V4 is located anterior to V2 and has traditionally been known as a visual object recognition processing center driven by bottom-up specification of object features or by top-down attention- based feedback. MT and V5 and are generally considered the area of extrastriate cortex that processes second-order motion (motion defined by a change in contrast). Recent results have shown that V2/V3 and V5/MT may be directly connected such that both may contribute to the processing of first and second order motions. The visual system can be summarized as being composed of ventral (magnocellular) and dorsal (parvocellular) streams (Figure 13). The ventral stream, also known as the “What Pathway”, mediates processing of color, form, and object identification and projects to ventral 7 <http://www.biopsychology.com/6e/step1002.html> 25 | P a g e regions of the higher visual areas and temporal lobe. Beginning in the occipital lobe, this includes areas V1, V2, V4, and inferior temporal areas (IT). The dorsal stream is often referred to as the “Where Pathway” and goes on to link dorsal extrastriate and parietal lobe areas that process spatial properties like stimulus motion and location. The pathway includes V1, V2, V3, MT, medial superior temporal areas (MST), inferior parietal cortex, and superior temporal sulcus. Figure 13. Illustration of dorsal and ventral visual pathways 8 . 2.2.2. Imaging the Human Visual Cortex Functional Magnetic Resonance Imaging Functional magnetic resonance imaging (fMRI) is a critical tool in the study of neuroscience and behavior, and has become an invaluable method for studying the human visual cortex. Functional MRI allows us to understand how the brain responds to outside stimuli by detecting fluctuations in blood oxygenation (also known as the hemodynamic response) that occur following changes in neuronal activity. Increased neuronal activity results in an increased demand for oxygen—this demand is met by locally increasing blood flow in the active regions. The oxygenated blood that is delivered is diamagnetic (i.e. weakly interacts with a magnetic field), while deoxygenated blood is paramagnetic (i.e. strongly attracted to a magnetic field). These differences in hemoglobin magnetization alter the MR signal that is created by the MRI machine’s static and dynamic magnetic, and RF fields. Since different levels of neuronal activity yield 8 <http://philosophy.hku.hk/courses/cogsci/ncc.php> 26 | P a g e different levels of oxygenated blood flow, blood-oxygen-level dependent (BOLD) contrast allows us to turn the resulting MR signal distortions into an image that indicates levels of activity across the brain (Figure 14). Figure 14. Illustration of how BOLD response maps are generated from the presentation of a stimulus. From left to right: a visual stimulus is presented with a hand movement and cognitive task, resulting in an increased neuronal responses in the corresponding cortical regions. This increase in synaptic activity increases blood flow to those areas. BOLD contrast is then used to generate functional maps of the responses on the cortex. For example, presentation of a visual stimulus leads to a hemodynamic response which can be measured by comparing the time-course of activity in the visual cortex with the time course predicted based on when the stimulus was presented. Activity that correlates with the prediction is concluded to be a result of the visual task. Activation maps can then be created describing the brain’s response. Diffusion Tensor Imaging A subset of MR imaging, diffusion tensor imaging (DTI) utilizes diffusion-weighted scanning techniques to develop maps of relative water diffusivity within the brain. This technique has been used extensively to determine tissue architecture and structural connectivity in the cortex. Diffusivity is measured by determining the fractional anisotropy (FA) of a particular voxel (valued from 0 to1): voxels that are highly anisotropic (i.e. in which water is diffusing along the same preferred axis) are assumed to be part of a white matter tract where water molecules are diffusing 27 | P a g e along the length of the fiber. Changes in the integrity of axonal fiber tracts can thus be determined by changes in FA values within a particular region. Magnetic gradients are applied in several different directions and attenuate the MRI signal in a manner that is dependent upon the diffusion characteristics of the voxel. The result is a diffusion tensor which mathematically describes the strength and direction of diffusion within each voxel. Based on the diffusion tensors, axonal bundles maps can be reconstructed using a technique called probabilistic tractography which describes the probability that fiber tracts exist between two regions in the brain based on the area’s diffusion characteristics. Fiber reconstruction may also be combined with maps of fMRI BOLD responses to show if and how activated regions are connected by white matter tracts (Figure 15). Figure 15. Illustration of reconstructed fiber tracts between regions of activation in the occipitotemporal cortex (Kim et al., 2006). Resting-State fMRI An understanding of the visual cortex is made more complete by investigating its functional connections with other regions of the brain. This is being accomplished through resting-state fMRI (rs-fMRI), which seeks to characterize the baseline activity of the brain in the absence of stimulation and hence determine which regions belong to functionally distinct networks (also known as “default networks”). Resting state networks are identified by temporally or spatially correlating spontaneous fluctuations in BOLD signals across the brain—regions that are highly correlated with one another are considered to be part of the same functional resting-state network. 28 | P a g e Beckmann et al., 2005 used independent component analysis (ICA) to identify 6 primary functional networks in the cortex, including 4 regions associated with vision in healthy individuals (Figure 16). The medial visual cortical areas include low-frequency activation ranging from the primary visual cortex located around the calcarine sulcus (Figure 16a) to the LGN of the thalamus— this functional region is thus implicated in relaying visual input from the retina to V1. Lateral visual cortical areas (Figure 16b) are functionally defined as the higher visual areas and extend from the occipital pole toward the temporal lobe boundary. As such, they are implicated in visuo-spatial attention, orientation, and navigation. A more pronounced visuo-spatial system is evident as a region extending from the parietal-occipital junction to the frontal pole (Figure 16e) and is associated with helping an individual to orient to salient visuo-spatial cues. The dorsal visual streams were identified as two spatially overlapping networks (i.e. networks that are functionally integrated) that include the occipital lobe and parietal cortex (Figure 16g, h). Figure 16. ICA results for low-frequency resting-state patterns in 10 subjects. Maps including the visual areas are seen in a, b, e, g, and h. (Beckmann et al., 2005) 2.3 The Human Visual Cortex and Blindness The human visual cortex is remarkably dynamic. Vision loss does not result in inactivation of the visual areas, but instead initiates a series of ongoing neuroplastic changes that allow the region to continuously adapt to changes in sensory inputs. Studies have shown that the visual 29 | P a g e cortex of blind individuals exhibits a high level of metabolism that is associated with neural activity (Volder et al., 1997), demonstrating that the region still remains biochemically functional without visual input. 2.3.1. Cross-Modal Plasticity in Blind Individuals Imaging studies have shown that the visual cortex adapts to vision loss in significant ways. Even in adults, cross-modal responses develop in partially-sighted individuals, where their visual cortex becomes active during non-visual tasks. Congenital and early blindness yields the most drastic cortical and behavioral changes. Lack of vision during the earliest stages of development— when the brain is most plastic—results in a functional reorganization of the region to accommodate other non-visual sensory inputs. Sadato et al., 2002 measured changes in cerebral blood flow to V1 during tactile discrimination tasks in 15 early and late-blind individuals to determine that the first 16 years of life marked a critical period for this shift in function. Subjects who went blind during the first 16 years of life exhibited increased activation to the tasks in V1, while late-blind subjects showed decreased activation during the same tasks. This is consistent with a PET and TMS study showing strong activation to a Braille reading task in congenital and early blind subjects (those who went blind before the age of 14), but not in late-blind subjects (Cohen et al., 1999). While V1 responses to tactile tasks were shown to be age-dependent, the visual association cortex showed higher activation in both early and late-blind groups. Subsequent studies similarly concluded that tactile-evoked responses in late-blind individuals is limited to extrastriate cortex, visual association areas, sensorimotor, and fusiform cortex (Goyal et al., 2006; Sadato et al, 2004; Sathian et al., 2005). This difference in responses between the two groups was attributed to blindness causing the visual association cortex to be recruited for tactile processing in early and late blind groups, with this processing being supplemented by V1 recruitment in early blind individuals (Sadato et al., 2002). The notion of visual association areas being recruited for tactile processing was promoted by Amedi et al., 2010 following his study showing dorsal and ventral visual stream recruitment for tactile processing in congenitally blind. Tactile-evoked V1 responses found in three late-blind patients was hypothesized to be the result of visual imagery in subjects with early visual experience (Buchel et al., 1998). While the occipital cortex of sighted individuals has also been shown to respond to tactile input (Zangaladze et al., 1999; Kauffman et al., 2002), this activity is intensified with blindness. 30 | P a g e In both congenital and early-blind groups, vision loss results in functional connectivity changes that enhance multimodal processing, resulting in a cortical network that is substantially different from sighted individuals. In most cases, tactile processing in visual areas is attributed to changes in functional connectivity between visual and somatosensory cortices. Delivery of rTMS to primary somatosensory cortex (S1) resulted in activation of early visual areas in early blind but not sighted groups, suggesting that functional connections between somatosensory and visual areas are enhanced with early vision deprivation (Wittenberg et al., 2004). Similar results were reported in early blind monkeys who showed extrastriate responses to somatosensory stimulation (Carlson et al., 1987; Hyvärinen et al., 1981a). Evidence for the functional relevance of these connections was shown when rTMS of the occipital cortex resulted in decreased Braille reading abilities in early blind, but not late-blind subjects (Cohen et al., 1999). An effective connectivity study by Stilla et al., 2008 used Granger causality analysis to investigate how cortical regions interact with one another in early and late- blind patients. In both groups, the visual areas were found to interact primarily with themselves and with the somatosensory cortices; these regions exhibited heightened spatial selectivity in early- blind when compared to late-blind subjects. These results support the notion that blindness causes tactile processing to be expanded from somatosensory areas into primary visual and association areas, and that the visual cortex is functionally relevant to processing tactile information. However, a few conflicting results have been observed: ischemic damage to the occipital lobe in one early- blind woman caused her to lose her ability to read Braille but did not interfere with her tactile performance (Hamilton et al., 2000). Whole brain connectivity studies in early blind by Liu et al., 2007 determined that early blind individuals exhibited decreased resting-state connectivity in the visual cortex as well as between the visual cortex and the parietal somatosensory, frontal motor, and temporal multisensory cortices when compared to their sighted counterparts. Instead, early- blind subjects showed increased resting-state connections between the visual cortex and frontal language cortices. These differences in tactile responses in visual cortex may be due to Braille stimuli being processed differently from other purely tactile input. While Braille reading has been shown to activate the primary visual cortex of blind individuals (Sadato et al., 1996), application of peripheral electrical stimuli to the Braille reading hand of twelve congenitally and early blind individuals did not elicit responses in V1 (Gizewski et al., 2003). This finding led to the conclusion 31 | P a g e that occipital activation during Braille and other tactile discrimination tasks in early and congenital blind subjects is not the result of plasticity of pure sensory function. Instead, visual cortex activation is due to changes in higher cortical areas that enable the brain to learn the difference between “finger touching” and “finger reading”. Contrary to this theory, another group suggests that visual area responses during Braille tasks in early and late-blind subjects may be the product of enhanced linguistic processing mechanisms already present in the normal visual cortex (Burton et al., 2002). Anatomically, congenital and early blindness have a significant effect on the structure of the visual cortex. When compared to normally-sighted subjects, congenitally blind show significantly reduced gray and white matter volumes, and decreased total brain volume (Ptito et al., 2008, Shimony et al., 2006). These include the optic nerves, optic radiations, and optic tracts, as well as decreases in both occipital lobes (specifically in striate and extrastriate cortices). However, segregation of the visual cortex into both dorsal and ventral visual streams is preserved in congenitally blind (Ptito et al., 2012). This leads to several theories on the nature and location of multimodal pathways in early blind individuals. Besides the notion that visual association areas (including dorsal and ventral visual streams) are implicated in tactile processing (Ptito et al., 2012), some suggest that tactile input is processed by a multimodal network extending from the somatosensory cortex to the dorsal posterior occipital cortex via the posterior parietal cortex (Sadato et al., 1996). Shu et al., 2009 used DTI to determine that early-onset blindness results in increased connections between motor and somatosensory cortices and other regions of the brain, hinting at a form of experience-dependent compensatory plasticity. In contrast to congenital and early-blind groups, late-blind individuals exhibited slight or no significant changes in cortical thickness, and generally had a greater connectivity density and higher global efficiency similar to sighted control subjects (Lepore et al., 2010; Li et al., 2012; Park et al., 2009). In addition, no significant changes have been found in fiber tract density of the optic radiation after decades of acquired blindness, suggesting that transmission of electric signals from retinal implants to the visual cortex is still possible (Schoth et al., 2006). 2.3.2. Compensation Following Sensory Deprivation Loss of a sense is often accompanied by an improvement in another sensory function. Several studies have demonstrated that vision deprivation results in both enhanced tactile acuity 32 | P a g e and hearing. Van Boven et al., 2000 and Goldreich et al., 2003 used tactile stimulation tests to show that blind individuals had greater tactile acuity when compared to their sighted counterparts. Even in sighted individuals, Kaufmann et al., 2002 found enhanced tactile acuity in subjects after being blindfolded for five days, while Facchini et al., 2009 found improved tactile acuity after only 90 minutes of blindfolding. Goldreich concluded that improved tactile ability in the blind was a result of the cross-modal plasticity observed with vision deprivation, though additional evidence is needed to support this theory. Blind subjects also show an amazing ability to maintain a spatial map of their environment even without any residual vision. Monaural and binaural listening tasks demonstrate that lack of vision causes early-blind individuals to create auditory maps of their surroundings that are as or more accurate than their sighted counterparts (Doucet et al., 2005; Lessard et al., 1998). This auditory compensation seems to vary with the extent of vision loss, where subjects with central vision loss (such as AMD patients) localize sounds less accurately than those with both central and foveal vision loss. Electrophysiological recordings in blind revealed that these enhanced sound localization abilities were correlated with differences in scalp distribution of electrical activity when compared to sighted controls (Roder et al., 1999). Similar to conclusions by Goldriech et al., 2003, these findings suggest that enhanced hearing in blind may be the result of reorganization in early auditory areas. While enhanced hearing is evident in both congenital and late-blind individuals, the neural mechanism differs between the two groups: congenitally blind are characterized by highly tuned early attentional filtering, while late-blind patients show enhanced sound target discrimination and recognition (Fieger et al., 2006). Fine et al., 2007 offered the hypothesis that sensory compensation is due to either “compensatory hypertrophy” (where improvements in a sense are due to modifications of that sense’s primary region of the brain), or cross-modal plasticity within the deprived sense’s cortex. For example, blind individuals with heightened tactile abilities may be experiencing compensatory hypertrophy within the somatosensory cortex or cross-modal plasticity in the visual cortex. While the majority of recent literature supports the notion of cross-modal plasticity, some evidence exists in favor of compensatory hypertrophy. Electrophysiological studies have found that somatosensory cortex representation of the Braille reading finger is expanded in blind subjects (Giriyappa et al., 2009; Pascual-Leone et al., 1993). Similarly, the topographical arrangement of 33 | P a g e fingers along the postcentral gyrus was disordered in Braille-reading blind individuals when compared to sighted controls (Sterr et al., 1998). 2.3.3. Theories on the Neural Basis of Cross-Modal Activity following Blindness While cross-modal activity has been well documented in cases of vision loss, the many conflicting results have made it difficult to encapsulate these findings under a unifying framework. However, several theories exist that attempt to explain the neural basis of cross-modal activation following vision deprivation (Figure 17). A recent study by Merabet et al., 2008 observed tactile-evoked visual cortex activation in 32 sighted subjects after being blindfolded for five days, suggesting that preexisting multisensory pathways remain suppressed in sighted individuals and are “unmasked” with vision deprivation. These pathways may then be re-masked if vision is restored—tactile activity disappeared in the sighted group within 24 hours of removing their blindfolds. Some endorse the notion that vision deprivation results in the creation of new neural networks and sensory associations that support cross-modal responses (Burton et al., 2003), while others suggest that the occipital cortex may act as an operator of a function based on the best-suited input available (Pascual-Leone et al., 2001). This operator hypothesis views the visual cortex as a metamodal structure that only responds to vision because the striate cortex is computationally best suited to process visual input. Once vision is no longer available, the visual cortex “unmasks” its tactile and auditory inputs to begin computing other available sensory information. However, several recent studies on humans and monkeys have questioned these functional claims and instead suggest that the observed activations may be non-functional unmasked feedback signals driven by task demands that are otherwise suppressed in the presence of visual inputs (Smirnakis et al., 2007; Wandell et al., 2009). Studies in AMD patients found that cortical activity was expanded into the lesion projection zone (LPZ) when performing a judgment-related task, compared to activity being absent from the LPZ when passively viewing the same stimulus. It was concluded that activity in LPZ was not the result of visual pathway reorganization, but instead was caused by task-related feedback signals (Masuda et al., 2010). 34 | P a g e Figure 17. Table illustrating four key theories on cross-modal plasticity. Theories are arranged in terms of whether they support the creation of new neural connections and whether they agree that tactile-evoked activity plays a functional role. 2.4 Implications of Cross-Modal Plasticity on Visual Prostheses As sight restoration treatments become an increasing reality, the question remains of how cross-modal plasticity in the brain will affect the ability of patients to use their restored vision. 2.4.1. Insights from Vision Recovery Studies Several studies have documented the effects of vision recovery following several types of sight restoration treatments. Vision restoration training in patients with six months to eight years of cerebral blindness (i.e. vision loss resulting from postchiasmatic damage to the visual pathways) resulted in improved visual fields. Functional MRI further revealed a change in size and location of receptive fields that accounted for local (but not large-scale) increases in visual field (Raemaekers 2010). Blindfold studies in sighted individuals (Merabet et al., 2008) revealed that a much shorter-term vision loss resulted in cross-modal processing in the visual cortex, while vision restoration yielded an immediate return to normal visual input processing. 35 | P a g e The most extensive study of the cortical effects of sight restoration following long-term vision deprivation was documented by Fine et al. 2003. After losing his vision at the age of three due to chemical and thermal damage to his cornea, subject MM only experienced minimal light perception (with no contrast or form vision) until the age of 43 when he received a corneal and limbal stem-cell transplant in his right eye. Psychophysical and neuroimaging techniques were used to assess the effects of the long-term deprivation on his cortical functions and visual perception. While MM’s electroretinogram responses were normal following surgery—indicating a healthy retina—fMRI responses revealed that the spatial resolution of his early visual cortex was degraded compared to normally-sighted subjects. Further tests revealed that while MM’s simple form, color, and motion processing remained normal, his complex form, object, and face recognition were severely impaired. The neural networks in those cortical regions responsible for complex form recognition were inactive, resulting in his inability to recognize faces and suggesting that MM would maintain this impairment unless a novel network was created to process complex stimuli. Fine et al., 2003 hypothesized that since novel complex forms (objects and faces) are encountered throughout life, complex form processing may remain plastic following the early stages of development and therefore be more vulnerable to vision deprivation. Consequently, MM’s ability to recognize and interpret complex forms experienced long-lasting deficits. Simple motion processing is typically established during infancy, and so remained unimpaired even after 40 years of vision loss. The results from these studies agree with subsequent studies that have found the visual cortex to contain both plastic and stable elements (studies reviewed in Wandell et al., 2009). The extent of plasticity within distinct systems (e.g. within the primary visual cortex versus higher visual areas) and the effects of specific injuries and rehabilitation efforts continues to be investigated. However, these findings provide evidence that the visual cortex of blind patients can once again process (to some degree) visual input following vision restoration. 2.4.2. Insights from the Cochlear Implant Functional MRI has been used to investigate the cortical effects of other electrical stimulation systems, including cochlear implants (CIs). In order to ensure that patients are not excluded from using clinical MRI for diagnostic purposes, CIs have been developed that either do not have electronics in the implanted element or allow removal of ferrous components. This avoids 36 | P a g e incompatibility issues due to image artifacts induced by ferrous electronic parts, unwanted implant stimulation during scanning, and any implant damage, heating, or movement that may occur within the strong magnetic fields. Systems like the Ineraid Cochlear Implant consist of intracochlear electrodes that are connected to a percutaneous plug and activated by an external current generator; electrodes are disconnected from the stimulation hardware during scanning. More recent systems (such as the HiRes 90K Implant produced by Advanced Bionics) may enter a 1.5-Tesla/64-MHz environment after the individual’s external sound processor, headpiece, and implant’s internal magnet have been removed. In a similar manner to individuals suffering from vision deprivation, deaf patients experience a cross-modal reorganization in which their visual cortex becomes responsive to auditory stimuli. Several imaging studies have demonstrated that the degree of cortical reorganization can be used to predict the effectiveness of rehabilitation post-implantation. In a study completed by Green et al. 2008, CI recipients underwent PET scans while listening to a pre- recorded story. Subjects with higher speech perception scores exhibited greater activity in the primary and association auditory cortices when compared to their poorer performing counterparts. A study by Doucet et al. 2005 showed that postlingual deaf patients with a greater extent of audiovisual association were more successful at rehabilitation with a CI than individuals whose visual cortex was less responsive to speech sounds. They further concluded that higher performers utilized this enhanced visual cortex activity to compensate for a CI’s imperfect auditory signal. This so-called “mutual reinforcement of audio and visual modalities” (leading to stronger auditory- induced visual cortex activation) was similarly recorded by Giraud et al. 2001 as being accompanied by an improvement in speech and lip-reading comprehension after receiving a CI. Cochlear implant studies thus provide some insight into the effect of sensory restoration on cross-modal activity. They suggest that the brain is plastic enough to process signal from electrical stimulation systems, and that subject performance may be predicted by the extent of activity in the association regions of the brain. 37 | P a g e PART II: THE EFFECT OF BLINDNESS ON THE VISUAL CORTEX The first step in determining the cortical causes of Argus II performance variability must begin with a basic understanding of how vision loss affects the visual cortex. The brain has proven to be both a plastic and stable entity that responds to different conditions in different ways: sensory deprivation may or may not result in functional or structural changes depending on the age of the individual, the duration of the deprivation, or the mechanisms that caused loss of the sense. The extent of this plasticity can have a significant impact on an individual’s response to sensory treatments. Our investigation began with an attempt to answer this basic question: what types of cortical changes result from gradual vision deprivation in late-blind RP patients? Honing-in on this specific patient population will allow us to more accurately predict patient outcomes in RP patients who receive the Argus II device. The results may also be generalizable to late-blind patients who suffer from other forms of blindness and will contribute to a general understanding of plasticity (or a lack thereof) in the adult human visual cortex. 38 | P a g e Chapter 3: Characterization of tactile- evoked V1 BOLD responses in late-blind individuals with retinitis pigmentosa 3.1 Overview Characterizing cross-modal responses in partially-sighted and blind adults can provide us with a cortically localized biomarker for vision loss, which may be useful for assessing sight restoration treatments. While some studies suggest that cross-modal plasticity will have a negative impact on an individual’s ability to adapt to vision restoration (Lee et al., 2001), others believe these cross-modal responses increase performance by compensating for the sensory device’s imperfect signal (Doucet 2005). In either case, understanding the relationship between vision loss and cross-modal responses may help us predict how an individual will respond to treatment. Cross-modal activations have been described in early and late-blind groups (Amedi et al., 2003, 2010; Buchel et al., 1998; Burton 2003; Cheung et al., 2009; Merabet et al., 2006; Ptito et al., 2005; Sadato et al. 1996, 2004; Sathian 2005). However, variability among individuals and correlations between severity of vision loss and extent of the responses is poorly understood. Tactile discrimination studies of blind have shown increased activation of occipital areas when compared to their sighted counterparts (Amedi et al., 2010; Cheung et al., 2009). Sadato et al. 39 | P a g e (1996) described activation of the primary and secondary visual cortical areas in early-blind subjects during a Braille reading task, and deactivation of those regions in a sighted control group. A similar pattern of tactile-evoked visual cortex activation was observed in sighted subjects after being blindfolded for five days and was reversed within 24 hours of removing the blindfold (Merabet et al., 2008). The presence of these occipital tactile-evoked activations is well- established, but the cause is in dispute (Burton 2003; Merabet et al., 2008; Pascual-Leone et al., 2001; Smirnakis et al., 2007; Wandell and Smirnakis, 2009). Nevertheless, tactile-evoked response in the visual cortex may still be valuable as a cortically localizable physiological marker for assessing the effect of vision loss on V1. For such applications, cross-modal responses should correlate with vision loss, and the effect size should be larger than the individual variation of cross- modal responses in the normally-sighted population. The state of V1 and other visual areas after vision loss can affect the efficacy of sight recovery treatment. The outcome of sight restoration procedures (such as implantation with a retinal prosthesis) often varies for reasons that are not fully understood (Humayun et al., 2012). Ocular imaging can measure the position of a prosthesis in the eye and psychophysics can record patient perceptions during stimulation, but no direct measures of brain activity are available. Functional imaging of cross-modal responses may provide part of the critical data that could account for the individual differences to treatment. For example, a decrease in tactile-evoked responses in the part of V1 that corresponds retinotopically to the implant would suggest that V1 is effectively driven by the signals evoked by the retinal stimulation. Conversely, undiminished tactile-evoked responses in V1 may indicate less effective stimulation. Similar analysis may be applied to other visual areas, provided these areas can be sufficiently localized in a blind subject (Benson et al., 2012; Henriksson et al., 2012). In this study, we took the first step towards these goals. We sought to determine if a basic relationship exists between severity of vision loss and the extent and strength of tactile-evoked V1 responses in late-blind individuals with retinitis pigmentosa (Hamel 2006). We used three simple tactile experiments to elicit a strong cross-modal response in primary visual cortex and found that the cross-modal response is significantly linked to degrees of vision loss across individuals with RP. 40 | P a g e 3.2 Study Design 3.2.1. Participants Eighteen subjects participated in the study (9 normally-sighted individuals and 9 individuals diagnosed with retinitis pigmentosa) having a mean ± SD age of 45.11 ± 13.78 years (range: 21-67 years); sighted control subjects were gender-matched and had a similar age range (24-66 years) to the RP subjects (21-67 years, Table 1). Five additional individuals with RP were also enrolled in the study but were excluded from our data analysis for the following reasons: two subjects were unable to perform the specified tasks correctly; a third subject was removed from the scanner after he exhibited claustrophobic symptoms; a fourth subject had posture restrictions that forced the use of a larger single-channel circular-polarization coil in place of our typical multi- channel coil, resulting in poor data quality; and a fifth subject was excluded from all analyses after we were unable to obtain consistent visual field information. Individuals were recruited from the community and received monetary compensation for their participation. The study received approval from the University of Southern California’s University Park Campus Institutional Review Board and all subjects provided written informed consent after explanation of the nature and possible consequences of the study. MRI experiments were conducted at the USC David and Dana Dornsife Cognitive Neuroscience Imaging Center, while additional visual acuity and Goldmann visual field measurements were obtained by ophthalmologists and study staff at the USC Doheny Vision Research Center. This research followed the tenets of the Declaration of Helsinki. 41 | P a g e Subject ID Age, Gender Visual Acuity Years since onset of symptoms Can Subject Read Braille? Diagnosis Additional Description of Vision RP1 41, F L.P. 22 Yes RP Light perception only RP2 52, M 20/60 46 No RP Partial tunnel vision; cataract removal from both eyes RP3 67, M L.P. 24 No RP Light perception only RP4 43, M 20/800 38 Yes RP Partial tunnel vision in right eye RP5 58, M 3/200 23 No RP Cataract removal from both eyes RP6 57, F 20/50 32 Yes RP Cataract removal from both eyes RP7 24, F 20/25 11 No RP Partial tunnel vision RP8 51, F 20/40 44 No RP Loss of peripheral vision, night blindness, blurred vision in right eye RP9 21, M 20/20 1 No RP Beginning loss of peripheral vision and some night blindness S1 M, 53 20/20 --- No --- Sighted S2 43, M 20/20 --- No --- Sighted S3 51, F 20/20 --- No --- Sighted S4 24, M 20/20 --- No --- Sighted S5 30, M 20/20 --- No --- Sighted S6 34, F 20/20 --- No --- Sighted S7 F, 40 20/20 --- No --- Sighted S8 F, 57 20/20 --- No --- Sighted S9 M, 66 20/20 --- No --- Sighted Table 1. Subject Demographics. Normally sighted subjects S1-S9 served as sighted, gender-matched control subjects for RP subjects RP1-RP9. L.P. = light perception only. All RP subjects were diagnosed with a typical form of retinitis pigmentosa. Diagnosis of RP was confirmed by each individual’s primary ophthalmologist and additional information regarding the individuals’ vision was obtained from their most recent medical records after receiving HIPAA authorization from the subjects. Subject RP8 (female, age 51) also had untreated cataract in her left eye, while subjects RP2 (male, age 52), RP5 (male, age 58) and RP6 (female, age 57) underwent successful cataract removal surgery prior to participating in the study. Following informed consent, the visual field of all RP subjects were measured during a Goldmann visual field examination. Four RP subjects completed visual acuity exams on-site, while visual acuity information for the other five RP subjects was obtained from their recent medical records. 42 | P a g e RP subjects with minimal light perception only (whose visual acuity could not be measured) were assigned a visual acuity of 20/20000 (LogMAR = 3) for analysis purposes. The RP group was divided into “Low Vision” (n = 5) and “Blind” (n = 4) subgroups. Those in the “Blind” subgroup were subjects whose visual acuity was worse than 20/200 (LogMAR = 1), a definition of legally blind. All remaining RP subjects were placed in the “Low Vision” subgroup. 3.2.2. Experimental Stimuli and Tasks Both sighted and RP groups completed the same three tactile tasks in the following order: 1) a shapes task requiring subjects to determine if any of a series of raised-line shapes was bilaterally symmetric, 2) a Braille-dot counting task in which subjects counted the number of dots in a series of random Braille letters (subjects were not asked to read the letters), and 3) a sandpaper task requiring individuals to determine the relative roughness between a strip of sandpaper and the sandpaper disc surrounding it (Figure 18). Each subject was given a sheet composed of 4 columns and 5 rows of tactile elements spaced approximately 25 mm apart, for a total of 20 tactile elements per sheet. Each sheet was attached to a plastic clipboard and handed to the subject by the experimenter before each functional scan, after which the subject placed his/her dominant hand in a “ready position” on the bottom left-hand corner of the sheet until the task began. Subjects completed two sheets for each task, where the second sheet consisted of the same tactile elements as the first in a rearranged order. Design of the shapes and Braille elements were based on a tactile stimuli setup used by Cheung et al. (2009) during tactile experiments in one blind subject. Figure 18. Example of the three tactile tasks, with shapes in red, Braille in green, and sandpaper in blue. The tasks were performed in a block design paradigm, in which individuals scanned a column during active blocks and rested their fingers in the empty space between columns during 43 | P a g e rest blocks. Each run was composed of four 20s active blocks (one active block per column) and five 20s rest blocks. These blocks were interleaved, with the run starting and ending on a rest block. Two runs were completed for each task, with each run lasting a total of 180s. Subjects were given 4s per tactile element (for a total of 20s per active block/column) for determining symmetry, number of dots, or relative roughness and were instructed to either explore the tactile elements or rest between columns. We chose to use these simple tasks as we are mindful of the challenges associated with testing patient populations. Subjects wore headphones and auditory instructions were given under computer control using a text-to-speech function. These instructions also indicated when subjects should move from one tactile element to another in a column; the auditory instructions were presented during both rest and active blocks. Participants did not report their answers during scanning. All subjects were asked to use their dominant hand and keep their eyes open while wearing a light excluding eye mask (made of black molded cell foam and nylon interlock fabric with a contoured rim) throughout the task. This “eyes-open-in-darkness” condition was found to mimic an interoceptive mental state that minimizes visual cortex activity due to imagination and multisensory activity (Marx et al., 2004, Figure 19). Both the scanner and scanner room lights were turned off. All completed a training session prior to entering the scanner and completed a verbal survey about their performance following the scans to ensure that the task was completed properly. Figure 19. V1 BOLD responses for a healthy sighted subject to the shapes tactile task under an “eyes- closed” versus “eyes-open” condition. Significant responses (FDR < 0.05) are color-coded, with warm colors denoting increases in BOLD responses relative to rest. The response patterns were projected onto an inflated representation of the occipital lobe, where the center white line represents the calcarine sulcus that runs through the primary visual cortex. Left: The subject was asked to keep their eyes closed while completing a tactile shapes task (Figure 18 Left). Right: The subject was asked to keep their eyes open while blindfolded throughout the task. 44 | P a g e 3.2.3. Image Acquisition MR images were acquired in a 3 Tesla Siemens MAGNETOM TIM Trio scanner using a 12-channel Matrix head array coil. Anatomical images were obtained using a T1-weighted sequence (MPRAGE) with TR/TE/flip angle/slice thickness = 1.95s/2.26ms/9°/1.2mm for sighted subjects, and TR/TE/flip angle/slice thickness = 2.3s/2.98ms/9°/1.0mm for RP subjects. Functional images with blood oxygenation level-dependent (BOLD) contrast were acquired using an echo-planar imaging (EPI) sequence with TR/TE/flip angle = 2s/25ms/60° and Prospective Acquisition Correction (PACE). 36 slices with isotropic voxels of 3x3x3 mm 3 were axially oriented and covered the entire cerebral cortex except for the tip of the temporal lobe for some subjects. Subjects laid head first and supine in the scanner. Foam padding was placed around the head to minimize movement during scanning, while earplugs and sound-attenuating headphones were provided to dampen scanner noise. 3.2.4. Defining Vision Loss Two measures were used to define each subject’s level of vision loss: visual acuity and preserved visual field in V1. Visual acuities were measured using a Snellan eye chart. Subjects were asked to stand 20 feet away from the chart and read each line with both eyes open and without any corrective lenses. The smallest best read line was considered to be their OU visual acuity. This fractional Snellan value was then converted to logMAR for analysis purposes. Preserved visual field was quantified in terms of the fractional areal size of a subject’s preserved visual field as projected to V1 cortex and is referred to as the “preserved visual field in V1”. Since we were unable to perform functional retinotopic mapping with our RP subjects, the amount of preserved visual field in V1 was estimated by mapping subjects’ visual field to a commonly used model of V1 based on macaque monkey data (Daniel and Whitteridge, 1961; Schira et al., 2010). The interspecies difference is largely irrelevant for our purpose. Goldmann visual field maps were determined based on subject responses to a 15 dB, 64 mm light stimulus. The maps were transferred to ImageJ, where each image pixel within the sighted regions of the visual field was isolated (see Figure 20 for illustrations of Goldmann visual field results). We then found the eccentricity value for each pixel and its corresponding areal cortical magnification factor, according to the following equation described in Motter (2009): 45 | P a g e 𝑀 𝑎 = 100(0.8 + 𝑤 ) −2.0 Equation 1. where w is the eccentricity in degrees and magnification is in square millimeters of cortex per square degree of visual field. Application of Ma according to Equation 2 yielded each pixel’s V1 cortical area: 𝐴 𝐶 = 𝑀 𝑎 ∗ 𝐴 𝑃 Equation 2. where AC is the cortical area for a single pixel and AP is the area of a single pixel as determined in ImageJ. The cortical area of all pixels within a sighted region were then summed to give the total V1 cortical area corresponding to the sighted region of each subject’s visual field (one areal value was obtained for each eye). To minimize interspecies difference, the total area derived from the macaque model was divided by the maximum of 2171.3 mm 2 such that a subject with normal visual field corresponds to a fractional V1 area of 1.0 (the maximum area was derived from the typical spatial extent of a nominal human visual field, as described in Walker et al., 1990). The total fractional V1 area of preserved visual field reported here are averages of values for both eyes for each subject. 3.2.5. fMRI Data Analysis Image data was analyzed using BrainVoyager QX (Goebel et al., 2006) in subjects’ native space (as opposed to normalizing to a standard space). Anatomical data underwent inhomogeneity correction and were reoriented via rigid-body rotation and translation to place the origin at the Anterior Commissure and the Posterior Commissure on the y-axis. All functional data was preprocessed with 3D motion correction (PACE and post hoc), slice timing correction, and temporal filtering. In cases of excessive head movement, which occurred in 3 RP subjects and 1 sighted subject, volumes in which a subject exhibited movement greater than 0.6 mm/degree of motion (based on online PACE estimation) and the corresponding entries in the design matrix were excluded from the analysis. Spatial smoothing was not applied to the functional data. Whole-brain voxel-wise BOLD modulation was obtained by estimating the signal level during the active blocks with respect to that during the resting blocks using a general linear model (GLM), with head-motion parameters as covariates. For each subject, individual functional data sets of each run were concatenated after normalization (z-transform). Significant voxel-wise activations were identified at false discovery rates (FDR) less than 0.05 with a cluster threshold of 46 | P a g e 25mm 2 . The activation maps displayed below for each subject were constructed by projecting the GLM contrast (t-statistics) obtained from voxels on the cortex onto the reconstructed and inflated cortical surface meshes of the subject. Putative primary visual cortex (V1) was identified anatomically for each subject, consisting of both banks of the Calcarine fissure, the parietal-occipital fissure, and the posterior end of the calcarine sulcus (Hinds et al., 2009). We calculated, for each subject, two complimentary measures (extent and strength) of the unsigned cross-modal response. The areal extent of cross-modal activation in the primary visual cortex was defined as the percentage of significantly modulated voxels on the cortex within the V1 ROI, while the strength of the response was calculated as the mean absolute parameter estimate (beta value) of the responding voxels within the V1 ROI. Mean absolute beta value was used in order to include all instances of cross-modal response, as tactile stimulation was found to evoke both negative and positive activities. These measures jointly provide a comprehensive characterization of the tactile-evoked BOLD response, including in subjects whose responses were spatially extensive but weakly modulated and those with both strong positive and negative modulations. ANOVA, multiple regression, and a linear mixed effects model (described in section 3.2.6) were used to identify statistically significant relationships (α = 0.05) between these dependent measures and the two measures of a subject’s visual function (acuity and preserved visual field area in V1) across the three tasks. Putative primary somatosensory cortex (S1) was also identified anatomically for each subject, extending from the middle of the central sulcus to the peak of the postcentral gyrus, and extending from the medial longitudinal fissure to the lateral sulcus. The percentage of modulated voxels and mean absolute beta value of those voxels were similarly calculated within the S1 ROI. 3.2.6. Statistical Modeling A linear mixed effects model was used to analyze the relationship between visual function and V1 tactile-evoked responses (dependent variables: extent or strength; covariates: fractional preserved visual field in V1 or visual acuity; repeated variables and factors: Task – Shapes, Braille, and Sandpaper). Task and the covariates (without interaction term) were entered into the model as fixed effects. This same linear mixed effects model was also used to determine the relationship between visual function and tactile-evoked responses in S1. 47 | P a g e To determine the relationship between vision loss and the V1 response when controlling for the influence of S1, a combined linear regression and linear mixed effects analysis was used to compute a pseudo-partial correlation amongst 1) V1 responses (strength or extent), 2) S1 responses, and 3) the subject’s visual function (i.e. preserved visual field in V1 or visual acuity). With this analysis, S1 responses were controlled as a contributing factor in order to study the relationship between visual function and V1 responses, while appropriately combining the effects across the three tasks that were run for each subject. As an example, the following steps were used to study the between-subjects relationship between strength of the response in V1 and visual acuity, while controlling for strength of the response in S1 (“partialling out” the S1 response): First, for each task, separate linear regression models were fitted to the subjects’ response strength in V1 and visual acuity (with response strength in S1 as the regressor) in order to estimate the residuals of the V1 response and visual acuity after accounting for variation in S1 response strength across subjects. For a given task, the correlation between these residuals would be the partial correlation between V1 response strength and visual acuity across subjects, controlling for S1 response strength. A mixed effect model was then used to combine these partial correlations across task, resulting in a pseudo-partial correlation. Specifically, to calculate the pseudo-partial correlation between V1 response strength and visual acuity while controlling for S1 response strength, the residuals from the aforementioned linear regression model for each task were modeled together using a mixed effects linear regression of the V1 residuals on the visual acuity residuals. The subject variable was modeled as a random intercept and the task variable was modeled as a slope, with the assumption that the covariance structure between the subject and task follows a compound symmetry structure (exchangeable). The pseudo-R 2 of the model was computed using the formula proposed by Cox & Snell (Cox et al., 1971). The square-root of this pseudo-R 2 was considered the pseudo-partial correlation between the strength of V1 response and visual acuity, while controlling for S1 response strength. Unlike models estimated using least squares regression, pseudo-R 2 does not represent proportion of variance explained, but rather the improvement of the fitted model compared to a null model using maximum likelihood. 48 | P a g e 3.3 Results 3.3.1. Extent and strength of V1 BOLD responses to tactile stimulation in RP and sighted subjects A summary of responses elicited with tactile stimulation is presented on an inflated cortical surface of V1 for four representative RP subjects and two representative sighted subjects (Figure 20). Individuals exhibited a range of vision loss (Table 1), which allowed us to partially account for the inter-subject variability of V1 responses to the tactile tasks. For each subject, V1 BOLD responses were quantified in terms of cortical extent, which was calculated as the percentage of voxels in V1 significantly modulated by the tactile stimuli during each task. A large degree of variability was found in the extent of tactile-evoked V1 responses among RP subjects [M = 31.63%, SD = 19.25%] (Figure 21A). A repeated measures ANOVA (between-subjects factor: Vision Level – Blind, Low Vision, and Sighted; within-subject factor: Task – Shapes, Braille, Sandpaper) revealed a highly significant effect of vision level, coarsely categorized into the three levels, on extent of the tactile-evoked BOLD response in V1 [F(2,14) = 16.758, p < 0.0001]. The spatial spread of the modulated voxels increased significantly with the severity of vision loss (Figure 20). The effect size was 0.705 (partial Eta Squared), and post hoc analyses (Tukey) showed that this effect of vision level on extent of the BOLD response was present between the Blind and Sighted groups [p < 0.0001], Blind and Low Vision groups [p = 0.017], and Low Vision and Sighted groups [p = 0.047]. An effect of task was also found on the extent of V1 BOLD activity [F(2,28) = 4.116, p = 0.027], where the Sandpaper task seemed to elicit the strongest response; interactions between vision level loss and task were not significant [p = 0.189]. Levene’s test for equality of variances demonstrated equal variance between the Blind and Low Vision groups for all tasks. 49 | P a g e Figure 20. V1 BOLD responses to the three tactile tasks in four representative RP subjects and two sighted control subjects. Significant responses (FDR < 0.05) were color-coded, with warm colors denoting increases in BOLD responses relative to rest. For each subject, the response patterns were projected onto an inflated representation of the occipital lobe; the outer white line represents the assumed V1/V2 boundary while the center white line represents the calcarine sulcus (CAS). Goldmann visual field results for both eyes (right eye on right) are presented in the first column and represent the subject’s visual field loss (black) and sighted field (white). Tactile-evoked V1 BOLD responses were further quantified in terms of the strength of the response, measured using the mean absolute beta value for each subject during each task (Figure 21B). The absolute beta value was used in order to include all instances of cross-modal response, as tactile stimulation was found to evoke both negative and positive activities (for 1 sighted and 5 V1 BOLD Responses to Tactile Stimuli Braille-Dot Counting Task Sandpaper Task Shapes Task Visual Cortex Responses Subject Goldmann Visual Field Increasing visual field 50 | P a g e RP subjects, at least 50% of the V1 voxels were significantly suppressed during tactile stimulation). A large degree of variability was found in the strength of V1 BOLD responses among RP subjects (in units of parameter estimate) [M = 3.41, SD = 0.63] (Figure 21C). However, a repeated measures ANOVA (between-subjects factor: Vision Level – Blind, Low Vision, and Sighted; within-subject factor: Task – Shapes, Braille, Sandpaper) did not reveal any significant effect of vision level on strength of the tactile-evoked BOLD response in V1 (p = 0.093). Similarly, no significant effect of task was found [p = 0.547] and interactions between vision level and task were not significant [p = 0.562]. Figure 21. Extent and strength of tactile-evoked responses in V1. A: The extent of tactile-evoked BOLD responses in V1, measured in terms of the percentage of modulated voxels (FDR < 0.05) in V1 for each subject and each task. RP subjects are ranked along the x-axis in descending order of severity of visual field loss. Sighted controls are grouped against a gray background. B: The strength of tactile-evoked BOLD responses in V1, measured in terms of mean absolute beta value of the significantly modulated V1 voxels for each subject and each task. C: Boxplots illustrating the distributions of the percentage and mean absolute beta value of modulated V1 voxels in RP and sighted control groups across all tasks. The red line indicates the mean within each group, the edges of the boxes indicate the 25th and 75th percentiles, and the whiskers illustrate the most extreme data points. 51 | P a g e Several studies have suggested that early blind individuals use the occipital cortex for language processing, resulting in visual cortex responses to reading Braille (Amedi et al., 2003; Bedny et al., 2011; Watkins et al., 2013). We explored if our late-blind RP subjects’ ability to read Braille had a significant effect on V1 responses during the Braille task. For RP subjects only, repeated measures ANOVA (between-subjects factor: Braille and non-Braille reader; within- subjects factor: Task – Shapes, Braille, and Sandpaper) revealed a significant effect [F(1,6) = 26.988, p = 0.002] of an ability to read Braille on extent of the V1 response (but not strength, p = 0.291). However, there was no interaction between the ability to read Braille and task [p > 0.10]. In other words, the effect of Braille reading ability was not more pronounced during the Braille task. 3.3.2. Relationship between cross-modal V1 BOLD responses and degree of preserved visual functions The BOLD response to tactile stimulation was characterized in greater detail by comparing activity in V1 to each subject’s preserved visual functions, quantified in terms of acuity and the fractional area of V1 cortex that corresponded to the preserved visual field (as described in section 2.4). A linear mixed effects model was used to analyze the relationship between a measure of visual function (visual acuity or preserved visual field in V1) and a measure of V1 tactile-evoked responses (extent or strength), with task as a random effect (see section 3.2.6). Visual inspection of residual plots for RP subjects alone did not reveal any obvious deviations from homoscedasticity and a Shapiro-Wilk test confirmed normality of the data [p’s > 0.10]. The modeled effects of visual function on V1 response are shown for RP subjects alone (Figure 22 bold lines), as well as for RP and sighted subjects combined (Figure 22 dashed lines). 52 | P a g e Figure 22. Predictive margins from a linear mixed effects model relating visual function to tactile- evoked responses in V1. For all plots, lines indicate linear predictions for the fixed portion (task) of a linear mixed effects model. The observed data (for each subject and task) is overlaid for comparison. The solid lines illustrate predictive margins among RP subjects only (“RP”), while dashed lines illustrate predictive margins among RP and sighted subjects combined (“All”). Statistics (p-values) above and below the lines correspond to dashed and bold lines, respectively, and indicate significance of the effect of visual function on the V1 response based on the mixed effects model. Sighted control subjects have by definition a fractional preserved visual field in V1 of 1 and average acuity of 0 logMAR, but are plotted separately (open symbols) from patient data and highlighted in gray for comparison. A: Percentage of tactile- modulated voxels in V1 (extent) versus fractional preserved visual field in V1. Fractional preserved visual field in V1 was determined by calculating the areal cortical magnification factor for sighted regions of the visual field. This value was then normalized using the maximum possible area. B: Mean absolute tactile- evoked BOLD modulation amplitude (strength) of responding V1 voxels versus fractional preserved visual field in V1. C: Percentage of modulated voxels in V1 versus visual acuity (logMAR) (subjects RP1 and RP3 were assigned a visual acuity of logMAR = 3 for analysis purposes, as these subjects’ had minimal light perception only and their acuity could not be measured). D: Mean absolute BOLD amplitude in modulated V1 voxels versus visual acuity. When the analysis was restricted to just the nine RP subjects, we found that visual acuity significantly affected the extent of V1 responses ([Parameter estimate β(d.f. = 8.085) = 8.992, SE = ±2.983, p = 0.017], Figure 22C). A trending effect of preserved visual field in V1 was found on the cortical extent of tactile-evoked responses in V1 ([β(d.f. = 6.930) = -27.710, SE = ±11.904, p = 0.053], Figure 22A). No effect of visual acuity or fractional preserved visual field in V1 was found on the strength of V1 responses [p’s > 0.07] (Figure 22B and D). When sighted subjects were included in the model, visual acuity and preserved visual field in V1 had a highly significant effect on extent of the V1 response (visual acuity: [β(d.f. = 17.269) 53 | P a g e = 12.154, SE = ±2.349, p < 0.0001]; preserved visual field: [β(d.f. = 16.339) = -28.720, SE = ± 5.421, p < 0.0001], Figure 22A and C). The effect of visual acuity on the strength of the response was also significant ([β(d.f. = 17.575) = 0.209, SE = ± 0.091, p = 0.034], Figure 22D), while the effect of preserved visual field in V1 on strength of the response remained insignificant [p = 0.263]. A previous tactile study by Merabet et al. (2008) found a correlation between the duration of vision deprivation and cross-modal responses in the occipital cortex. However, a linear mixed effects model (using years since onset of blindness as a covariate) revealed no significant effect of duration of blindness on either the extent or strength of the cross-modal V1 BOLD response in RP subjects [p’s > 0.70]. 3.3.3. Comparison of tactile-evoked BOLD responses in V1 and S1 V1 responses to tactile stimulation prompted the question of whether S1 was a source. With early blind individuals, repetitive TMS stimulation to S1 activated V1 (Wittenberg et al., 2004). Similarly, an effective connectivity study by Fujii et al. (2009) suggests that an indirect cortico- cortical feedback pathway from S1 to V1 exists that is modulated by vision loss, resulting in an expansion of tactile processing into visual areas. 54 | P a g e Figure 23. S1 BOLD responses to the three tactile tasks in four representative RP subjects and two sighted control subjects. Significant activations (FDR < 0.05) were color-coded, with warm colors denoting increase in BOLD responses relative to rest. For each subject, the response patterns were projected onto a partially-inflated representation of the cortex; the white-bolder regions represent S1, extending from the middle of the central sulcus to the peak of the postcentral gyrus. Goldmann visual field results for both eyes (right eye on right) are presented in the first column and represent the subject’s visual field loss (black) S1 BOLD Responses to Tactile Stimuli Braille-Dot Counting Task Sandpaper Task Somatosensory Cortex Responses Subject (Handedness) Goldmann Visual Field Shapes Task Increasing visual field LH RH RP4 (R/L) RP5 (R) RP8 (R) RP9 (L) S2 (R) S3 (R) 55 | P a g e and sighted field (white). Subject handedness is given in parentheses in the first column. All subjects used their dominant hand to complete the tasks (subject RP4, who is ambidextrous, used his left hand to complete the tasks). A summary of responses elicited with tactile stimulation is presented on an inflated cortical surface of S1 for four representative RP subjects and two representative sighted subjects (Figure 23). A large degree of variability was found in the extent of tactile-evoked S1 responses [M = 30.07%, SD = 13.71%] and strength of the response [M = 4.18, SD = 0.69] among RP subjects (Figure 24). We first determined the effect of vision loss on tactile-evoked BOLD responses in S1, similar to what we did with responses in V1: a repeated measures ANOVA (between-subjects factor: Visual Ability – Blind, Low Vision, and Sighted; within-subject factor: Task – Shapes, Braille, Sandpaper) revealed no significant effect of visual ability on either the extent or the strength of tactile-evoked S1 BOLD responses [p’s > 0.10] and no significant effect of task [p’s > 0.20]. 56 | P a g e Figure 24. Extent and strength of tactile-evoked responses in S1. A: The extent of tactile-evoked BOLD responses in S1, measured in terms of the percentage of modulated voxels (FDR < 0.05) in S1 for each subject and each task. RP subjects are ranked along the x-axis in descending order of severity of visual field loss. Sighted controls were grouped against a gray background. B: The strength of tactile-evoked BOLD responses in S1, measured in terms of mean absolute beta value of the significantly modulated S1 voxels for each subject and each task. C: Boxplot illustrating the distributions of the percentage and mean absolute beta value of activated S1 voxels in RP and sighted control groups across all tasks. The red line indicates the median within each group, the edges of the boxes indicate the 25th and 75th percentiles, and the whiskers illustrate the extreme data points, excluding outliers (red data points). However, when visual ability was further quantified in terms of preserved visual field and visual acuity, a significant relationship was found between tactile-evoked responses in S1 and vision loss. Among RP subjects only, a linear mixed effects model including RP and control subjects found a significant effect of preserved visual field in V1 on the extent and strength of the S1 response (extent: [β(d.f. = 18.799) = 22.252, SE = ±6.741, p = 0.004]; strength: [β(d.f. = 17.759) = 1.262, SE = ± 0.249, p < 0.0001], Figure 25A and C). A significant effect of visual acuity was also found on the extent of the S1 response [β(d.f. = 19.366) = -4.852, SE = ± 1.833, p = 0.018, 57 | P a g e Figure 25B], while a trending effect of visual acuity was found on S1 response strength [p = 0.059, Figure 25D]. When sighted subjects were included in the mixed effects analysis, the effect of visual acuity and preserved visual field in V1 on the strength of S1 response was significant [p’s < 0.01], while the effect of visual acuity and visual field on the extent of the S1 response did not reach significance [p’s > 0.07]. This difference in the effect of vision loss on S1 responses when compared to RP subjects alone may be attributed to variability in S1 responses among the sighted subjects (Figure 24). Figure 25. Predictive margins from a linear mixed effects model relating visual function to tactile- evoked responses in S1. For each plot, the lines indicate the marginalized predicted (fixed) effect of preserved visual field (left panels) or visual acuity (right panels) by the linear mixed effects model. The observed data (for each subject and task) is overlaid for comparison. The solid lines illustrate marginalized effects among RP subjects only (“RP”), while dashed lines illustrate marginalized effects among RP and sighted subjects combined (“All”). Statistics (p-values) above and below the lines correspond to dashed and bold lines, respectively, and indicate significance of the effect of visual function on the S1 response based on the mixed effects model. Sighted control subjects have by definition a fractional preserved visual field in S1 of 1 and average acuity of 0 logMAR, but are plotted separately (open symbols) from patient data and highlighted in gray for comparison. A: Percentage of tactile-modulated voxels in S1 (extent) versus fractional preserved visual field in V1. B: Mean absolute tactile-evoked BOLD modulation amplitude (strength) of modulated S1 voxels versus fractional preserved visual field in V1. C: Percentage of modulated voxels in S1 versus visual acuity (logMAR) (subjects RP1 and RP3 were assigned a visual acuity of logMAR = 3 for analysis purposes, as these subjects’ had minimal light perception only and their acuity could not be measured). D: Mean absolute BOLD amplitude in modulated S1 voxels versus visual acuity. 58 | P a g e It is worth noting that vision loss was found to result in an increased tactile-evoked response in V1 but a decrease in S1 responses. This trend in S1 may suggest that individuals with greater vision loss did not press as hard on the tactile elements when compared to their sighted counterparts, possibly as a result of increased tactile sensitivity with blindness (Goldreich et al., 2006). Given that different individuals may have employed diverse tactile strategies when completing the tasks, it is conceivable that these strategic differences could account for variation in the V1 BOLD response. We used the tactile-evoked responses in S1 as a proxy for tactile strategy differences across individuals and controlled for it statistically. Specifically, when we controlled for task-evoked responses (i.e. extent and strength) in S1 across RP and sighted subjects, the (pseudo-partial) correlations between visual function (indexed by either the preserved visual field in V1 or visual acuity) and the extent or strength of tactile-evoked responses in V1 became stronger (Table 2A) when compared to the effect of visual function on V1 responses found without controlling for responses in S1. Among RP subjects only, controlling for the extent and strength of the S1 BOLD response also resulted in a more significant pseudo-partial correlation between visual acuity and the extent and strength of the V1 BOLD response when compared to results not controlling for the S1 response (Table 2B). 59 | P a g e Table 2. The effects of visual function (preserved visual field in V1 or visual acuity) on V1 BOLD responses (extent or strength) with and without controlling for S1 BOLD responses (including both extent AND strength of the tactile-evoked response in S1). The marginalized effect column was duplicated from the linear mixed effects model results shown in Figure 25. The column showing the effect when controlling for S1 responses represents the results of the pseudo-partial correlation between visual function and V1 BOLD responses, excluding the influence of S1 responses (see section 3.2.6). (*) indicates significant relationships (p > 0.05). A: Results including both sighted and RP groups. B: Results including RP subjects only. Tactile-evoked responses were further compared between V1 and S1 across subjects. While vision loss was found to modulate responses in both regions, a linear mixed effects model found no relationship between either the extent or strength of BOLD responses in V1 and S1 across all subjects [p’s > 0.40]. Since vision loss was found to have a different effect on responses in V1 and S1, it was possible that any underlying relationship between the responses in both regions would be masked by vision loss. We therefore further explored the relationship between the signals in both regions after removing the influence of visual function. When a pseudo-partial correlation was performed to control for either visual acuity or preserved visual field, the relationship between the extent and strength of responses in V1 and S1 remained insignificant (p’s > 0.10, Figure 26). This suggests that V1 tactile-evoked activity is not a direct consequence of S1 activity in late-blind individuals. 60 | P a g e Figure 26. Graphical illustration of pseudo-partial correlations between visual function (preserved visual field in V1 and visual acuity) and tactile-evoked responses in V1 and S1 across all subjects. The pseudo-partial correlation was determined by modeling the residuals of V1 and S1 responses (in the absence of visual function) from each task using a mixed effects linear regression of the V1 residuals on the S1 residuals while adjusting for task. The pseudo-partial correlation between two quantities connected by an edge in each diagram represents the relationship between two quantities while excluding the effect of a third (e.g. the bottom edge in each orange diagram represents the pseudo-partial correlation between V1 and S1 responses while controlling for the effect of visual acuity. Cox-Snell pseudo-R 2 values are indicated as “coxs” along with their corresponding p-values. A: Pseudo-partial correlations between preserved visual field in V1, extent of responses in V1, and extent of responses in SI. B: Pseudo-partial correlations between preserved visual field in V1, strength of responses in V1, and strength of responses in SI. C: Pseudo-partial correlations between visual acuity, extent of responses in V1, and extent of responses in SI. D: Pseudo- partial correlations between visual acuity, strength of responses in V1, and strength of responses in SI. Significant pseudo-partial correlations are indicated with a bold edge. 3.4 Discussion Individuals with retinitis pigmentosa are the target population for recent sight restoration technologies, including retinal implants and gene therapies. We focused solely on RP patients in 61 | P a g e the current study in order to build a foundation for future studies that investigate the cortical effects of sight-restoration treatments in RP patients. Previous studies, using a diverse subject population, have demonstrated that the visual cortex becomes responsive to tactile input with vision loss (Amedi et al., 2003; Buchel et al., 1998; Burton 2003; Cheung et al., 2009; Merabet et al., 2006, 2008; Ptito et al., 2005; Sadato et al. 1996, 2004; Sathian 2005). Here, we sought to expand upon these findings by determining if a relationship exists between severity of vision loss and the extent and strength of tactile-evoked V1 responses in late-blind individuals with RP. Our results indicate that a significant correlation exists between vision loss and amount of cross-modal modulation: as visual acuity and visual field loss become more severe, V1 becomes more responsive to tactile stimulation. We found that the pseudo-partial correlation between tactile-evoked responses in V1 and visual function (preserved visual field in V1 and visual acuity) across subjects became highly significant after controlling for tactile-evoked responses in S1. This suggests that an underlying relationship exists between variation in V1 tactile-evoked responses and vision loss, but that this relationship is partially masked by variations in S1 activity. Several factors may result in a variable S1 response, including differences in subject tactile exploration strategies and subject differences in sensitivity to tactile elements; these tactile strategy and sensitivity differences may be especially pronounced between sighted and blind subjects. Removing these variances thus increases the association between vision loss and V1 cross-modal activity. This result, in addition to the insignificant relationship between V1 and S1 responses after controlling for vision loss, suggests that tactile-evoked activity in V1 is not a direct consequence of S1 activity. Several theories exist as to the cause of these cross-modal activations. Merabet et al. (2008) suggest that preexisting multisensory pathways remain suppressed in sighted individuals and are “unmasked” with vision deprivation. Others have endorsed the notion that vision deprivation results in the creation of new neural networks and sensory associations that support cross-modal responses (Burton 2003), while others suggest that the occipital cortex may act as an operator of a function based on the best-suited input available (Pascual-Leone et al., 2001). However, recent studies on humans and primates have questioned these functional claims and instead suggest that the observed activations may be unmasked feedback signals driven by task demands that are otherwise suppressed in the presence of visual inputs (Smirnakis et al., 2007; Wandell and Smirnakis, 2009). If tactile-evoked responses in V1 are due to “unmasking” of otherwise 62 | P a g e suppressed connections from S1, it is expected that the correlation of task-evoked responses between S1 and V1 would be higher for individuals with more severe vision loss. We did not observe this simple version of unmasking from S1. Overall, the observed cross-modal modulation is highly variable across late-blind individuals. The degree of preserved visual functions, particularly when expressed in terms of the fractional areal size of the V1 cortex that corresponds to the preserved visual field, explains some of this variance. Other factors may include preserved visual functions beyond the ones we quantified, variations in functional connectivity between visual and other cortical areas (Fujii et al., 2009; Leo et al., 2012), differences in tactile sensitivity, and unspecific individual differences. No significant effect of years since onset of blindness was found on tactile-evoked responses in V1. This is counter to a finding of Merabet et al. (2008), which suggests that visual cortex responses to tactile stimulation become increasingly stronger over time after the onset of vision deprivation. The number of subjects presented in this study may not be sufficient to observe this effect. Alternatively, our results may be specific to late-blind individuals and/or retinitis pigmentosa patients (whose photoreceptor layer deteriorates at varying rates in different individuals). For RP subjects, it may be the amount of vision loss that has occurred over time— and not the duration of vision deprivation—that affects cross-modal changes. Since the pace of vision deterioration varies significantly among RP patients, years since onset of blindness may be insufficient to infer the degree of cross-modal modulation that has occurred. 3.5 Conclusions In summary, vision loss was found to have a significant effect on tactile-evoked V1 BOLD responses in late-blind individuals. Our findings indicate that while tactile-evoked V1 responses are variable among late-blind individuals and partially depend upon the type of tactile task being performed, the correlation between tactile responses in V1 and vision loss is reliable across subjects, particularly after controlling for tactile-evoked S1 responses. Cross-modal modulation may be a useful biomarker for assessing progress and identifying bottlenecks in different visual areas following sight restoration treatments. In particular, if pre-treatment cross-modal responses are found to correlate with an individual’s ability to adapt to sight restoration treatments, such a biomarker—relating vision loss to tactile-evoked responses—could be used to predict how a late- blind RP patient will respond to treatment given their severity of blindness. 63 | P a g e Chapter 4: A case study of tactile-evoked V1 BOLD responses in patients with RP, AMD, and optic nerve hypoplasia 4.1 Overview The significant relationship between visual function and V1 tactile-evoked responses in RP patients prompts the question of whether retinal eccentricity maps could further be used to predict the location of cross-modal responses on the visual cortex. Several theories have been proposed on the mechanism driving cross-modal plasticity in low vision populations, including: the formation of new connections to process non-visual input, the unmasking of already existing cross- modal pathways, and the characterization of the visual cortex has an operator that computes the best-suited input available. Based on studies by those who believe cross-modal responses in the visual cortex serve a functional purpose (and are not solely the results of unmasked feedback signals driven by task demands), we can hypothesize that regions no longer receiving visual input undergo some form of reorganization that then causes them to respond to other sensory information. Regardless of the underlying mechanism, one might expect a relationship to exist between the absence of visually-evoked activity and the presence of tactile-evoked responses in 64 | P a g e the visual cortex, where regions that no longer respond to visual stimulation are more likely to produce tactile-evoked responses in blind subjects. We explored the existence of this relationship by reviewing visual and tactile-evoked V1 responses in 2 RP subjects, 1 individual with AMD (a retinal degenerative disease characterized by progressive central vision loss), and 1 individual with optic nerve hyploplasia (a congenital disorder symptomatically similar to RP in which the optic nerve is underdeveloped). By comparing cases of central versus peripheral vision loss, we can gain a general understanding of how vision loss across the retina manifests itself in the visual cortex. This case study accomplished the following objectives:  We compared BOLD responses to a flickering checkerboard stimulus and tactile stimulation in V1 of 1 RP and 1 AMD subject  We further compared the retinotopic maps of 1 RP and 1 optic nerve hypoplasia (ONH) patient with the location of tactile-evoked responses on the cortex 4.2 Study Design 4.2.1. Participants 6 subjects were included in the following study and divided into 4 categories: RP, AMD, ONH, and sighted groups. 3 RP subjects (RP4, RP5, and RP8) were taken from the original group of 18 subjects who had previously completed our fMRI tactile-discrimination study and had varying degrees of tunnel vision. However, subject RP4 showed no cortical response to the visual stimulus and was therefore excluded from all analyses due to uncertainty of him completing the task correctly. Subject AMD1 presented with severe central vision loss while subject ONH1 exhibited upper peripheral vision loss (Table 3). 65 | P a g e Table 3. RP, AMD, ONH, and Sighted Subject Demographics. Goldmann visual field results are displayed for the left eye (on left) and right eye (on right). Black regions indicate visual field loss based on subject responses to a 15 dB, 64 mm 2 light stimulus. The stripped region in ONH1’s visual field maps indicate regions responding to a smaller 15dB, 1 mm 2 light stimulus. L.P. = light perception only. 4.2.2. Image Acquisition MR images were acquired in a 3 Tesla Siemens MAGNETOM TIM Trio scanner using a 12-channel Matrix head array coil. Anatomical images were obtained using a T1-weighted sequence (MPRAGE) with TR/TE/flip angle/slice thickness = 2.3s/2.98ms/9°/1.0mm. Functional images with BOLD contrast were acquired using an echo-planar imaging (EPI) sequence with TR/TE/flip angle = 2s/25ms/60° and PACE. 36 slices with isotropic voxels of 3x3x3 mm 3 were axially oriented and covered the entire cerebral cortex except for the tip of the temporal lobe for some subjects. 66 | P a g e 4.2.3. Experimental Stimuli and Tasks All subjects completed the same three tactile tasks as described in Chapter 3 (see section 3.2.2), including: 1) a shapes task requiring subjects to determine if any of a series of raised-line shapes was bilaterally symmetric, 2) a Braille-dot counting task in which subjects counted the number of dots in a series of random Braille letters, and 3) a sandpaper task requiring individuals to determine the relative roughness between a strip of sandpaper and the sandpaper disc surrounding it. Subject RP5 also completed a visual task in which a flickering checkerboard was presented during scanning. The subject was instructed to gaze directly upwards towards a clip-on mirror that was placed on the head coil directly above the subject’s eyes. This mirror was oriented towards a rear-projection screen at the back of the scanner, allowing the subject to view the stimulus presentation. The screen size using the clip-on mirror was 23.6 ○ x 18.4 ○ with a viewing distance of 85 cm. Two alternating black and white checkerboards were flickered at a frequency of 0.10 Hz (Figure 27). The stimulus was presented in a block design paradigm, in which the checkerboard flickered continuously during active blocks and a black screen was presented during rest blocks. Each run was composed of ten 16s active blocks and eleven 16s rest blocks. These blocks were interleaved, with the run starting and ending on a rest block. Two runs were completed for the checkerboard task, each lasting a total of 336s. This stimulus allowed us to determine the responsiveness of each subject’s visual cortex to visual input. Figure 27. Example of the checkerboard stimulus. Two alternating checkerboards were presented during active blocks, followed by a blank black screen during resting blocks. Active Block Rest Block … 67 | P a g e Retinotopic mapping was completed in subjects RP8, ONH1, and S3 by Pinglei Bao, PhD. Two visual stimuli were presented during scanning as subjects fixated on a point at the center of the screen: a flickering checkerboard configured in a ring that expanded through increasing eccentricities and a wedge that rotated around a central point. 4.2.4. fMRI Data Analysis Image data was analyzed using BrainVoyager QX (Goebel et al., 2006) in subjects’ native space (as opposed to normalizing to a standard space). Anatomical data underwent inhomogeneity correction and were reoriented via rigid-body rotation and translation to place the origin at the Anterior Commissure and the Posterior Commissure on the y-axis. All functional data was preprocessed with 3D motion correction (PACE and post hoc), slice timing correction, and temporal filtering—spatial smoothing was not applied to the functional data. For both tactile and visual tasks, whole-brain voxel-wise BOLD modulation was obtained by estimating the signal level during the active blocks with respect to that during the resting blocks using a GLM, with head-motion parameters as covariates. Significant voxel-wise activations were identified at FDR < 0.05 with a cluster threshold of 25mm 2 . Comparison of visually-evoked and tactile-evoked responses in V1 For subject RP5 who completed the checkerboard stimulus task, a “Sighted Region” ROI was created in V1 that included all voxels with a positive response to the visual stimulus. A second “Vision Loss Region” ROI was defined as voxels that had either a negative response to the stimulus or no response at all. Within the two ROIs, the areal extent of cross-modal activation from the three tactile tasks was defined as the percentage of significantly modulated voxels in a region, while the strength of the response was calculated as the mean absolute parameter estimate (beta value) of the responding voxels within the V1 ROI. The extent and strength of tactile-evoked voxels in each ROI were then compared to determine if the cross-modal responses in the Sighted Region were significantly different from those in the Vision Loss Region. 68 | P a g e 4.3 Results 4.3.1. Visual Comparison of Visually- and Tactile-Evoked Responses in V1 There is some evidence that the amount of spatial vision loss may be related to the observed tactile activation patterns in V1: of those RP subjects enrolled in the tactile discriminations study, several exhibited tactile responses in regions of the cortex corresponding to their vision loss. Subject RP4 had both central and peripheral blindness and experienced activation of both foveal and peripheral regions of V1 during the sandpaper task, while subjects RP8 and RP9 only exhibited peripheral blindness and had a greater amount of activation in the peripheral regions of V1 when compared to the foveal region (Figure 20). Given that subjects RP5 and AMD1 had opposing forms of retinal degeneration (loss of peripheral vision in the former case and loss of central vision in the latter), both subjects’ tactile- evoked responses were compared to determine if regions no longer responsive to visual input were now responsive to tactile input. Visual inspection of RP5 and AMD1’s responses to the sandpaper task seem to show tactile-evoked responses being contained to a particular region of V1: tactile- evoked responses were absent from the occipital poles in RP5 (who still retained some central vision) but present at the occipital poles of subject AMD1 (who’s central vision was lost, Figure 28). However, AMD1 also exhibited tactile-evoked responses in the peripheral regions of V1. Figure 28. Comparison of V1 responses to the sandpaper task in one RP patient and one AMD patient. White lines indicate the boundary of V1 and the central sulcus (CAS) through the center. Description of inflated mesh and Goldmann field maps Visual Cortex Responses to the Sandpaper Task AMD1 RP5 V1 V1 CAS V1 V1 CAS LH RH 69 | P a g e V1 tactile-evoked activity in RP5 were further compared with his cortical responses to a visual stimulus. Positive visually-evoked responses were localized to the foveal and early peripheral regions of V1, while a strong negative response was observed in the peripheral regions of V1 (Figure 29). This was consistent with RP5’s tunnel vision (i.e. degeneration of his peripheral visual field). Overlaying responses to the visual stimulus on a cortical map of the tactile-evoked activity showed significant overlap between tactile responses and positive BOLD responses to the visual task (Figure 29, Top). The extent and strength of RP5’s tactile-evoked responses were further compared between two regions: regions with a positive response to visual input and regions with voxels that had either a negative response to the visual stimulus or no response at all. Across the three tactile tasks, a one-tailed t-test (unequal variances) found no significant difference between the strength or extent of tactile-evoked responses in regions still responding to visual input versus regions with a suppressed or absent response to the checkerboard task (p’s > 0.10, Figure 29 Bottom). It is interesting to note, however, that responses to the Braille-counting task were only present in regions of V1 that no longer responded to visual stimulation. 70 | P a g e Figure 29. Top: Comparison between V1 BOLD responses to a visual stimulus (left) and responses to the three tactile tasks in subject RP5. The yellow-white overlay (3 rd column from left) shows visually-evoked (positive) BOLD responses against the activation maps of tactile-evoked activity, while the blue overlay (4 th column from left) shows regions with a suppressed or absent response to the visual input against a tactile-evoked activation map. Bottom: Comparison of tactile-evoked responses to the three tasks (extent on left and strength or right) in areas with a positive visual response (“Sighted Region”) and areas with a non-positive visual response (“Vision Loss Region”). Visual Cortex Responses Checkerboard Stimulation Tactile Task Tactile Task overlaid with Positive Responses to Visual Stimulus Tactile Task overlaid with Non- Positive Responses to Visual Stimulus Sandpaper Task Shapes Task Braille Task V1 V1 V1 V1 CAS CAS V1 V1 CAS V1 V1 CAS CAS CAS V1 V1 V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 CAS V1 V1 71 | P a g e 4.3.2. Comparison of Tactile-Evoked Responses in V1 to Retinotopic Maps Retinal eccentricity maps (provided by Pinglei Bao, PhD) were further used to explore whether a relationship may exist between responses to tactile input in V1 and the representation of the retina on the cortex. In accordance with their level of vision loss, sighted subject S3 showed a representation of both foveal and peripheral regions of the retina on the cortex, while subject RP8 had a more extensive foveal representation (indicated by the light blue and yellow regions) and less peripheral representation on the cortex. Subject ONH1 with upper visual field loss also had a greater representation of the fovea in V1, particularly on the left hemisphere (Figure 30). For subject ONH1, the extent of foveal versus peripheral representation on the cortex seems to coincide with the regions of V1 that are responsive to tactile input—less represented eccentricities correspond to regions with the least amount of tactile-evoked responses. The right hemisphere visual areas of subject ONH1 are more dedicated to the foveal region of the retina and similarly shows a larger tactile response at the occipital pole when compared to the left hemisphere, which has a smaller foveal representation and no tactile responses at the occipital pole. Subjects RP8 and S3, however, did not exhibit this pattern. 72 | P a g e Figure 30. V1 BOLD responses to a tactile task in 1 subject with optic nerve hypoplasia, 1 RP subject, and 1 sighted control subject. Significant activations (FDR < 0.05) were color-coded, with warm colors denoting increase in BOLD responses relative to rest. Goldmann visual field results for both eyes (right eye on right) are presented in the first column and represent the subject’s visual field loss (black) and sighted field (white). Activation patterns (shown in the second column) were projected onto an inflated representation of the occipital lobe; the outer white line represents the V1/ V2 boundary while the center white line represents the calcarine sulcus (CAS). The third column shows the retinotopy maps in the occipital lobe for the three subjects. Color shows the retinal eccentricity (red [fovea] to magenta [periphery]). 4.4 Discussion In an exploratory case study of one subject (RP5), we found no significant difference between the strength and extent of tactile-evoked responses in areas of V1 still responding to vision versus unresponsive regions. Tactile and visually-evoked regions often overlap with one another, suggesting that multisensory pathways exist in V1 that are responsive to both visual and tactile stimulation. However, a visual comparison of RP and AMD subjects’ visual field maps and BOLD activation maps seem to indicate a general localization of tactile responses to regions with the most 73 | P a g e severe vision deprivation (e.g. peripheral regions in individuals with tunnel vision)—more subjects should be evaluated to verify this theory. These two findings are not in conflict, but instead may be further evidence that the adult visual cortex is both plastic and stable, as described by Wandell et al. 2009. The presence of tactile- evoked activity in sighted control subjects (see section 3.3.1) suggests that multisensory pathways already exist in sighted individuals that become increasingly responsive to tactile input as vision loss advances. Some studies have suggested that cross-modal responses are not reversible and may therefore be detrimental to visual prostheses: since those areas no longer process visual stimuli, they will prevent artificial vision from being completely effective. However, the observed form of plasticity does not necessarily require the dramatic creation of new neural networks, but may simply use pre-existing ones in a new way such that these pathways might still be available to process vision once it is restored. The stable component would consist of tactile-evoked responses that are not functional or a result of cortical reorganization, but instead are a consequence of task-related cortical signals that are present in both visually-responsive and visually-unresponsive regions of V1 (as described by Masuda et al., 2010). Comparison of tactile-evoked responses in V1 with retinotopic maps in 3 subjects furthers this notion. Subjects RP8 and S3 exhibited tactile responses that were located across V1, irrespective of how the retina was represented on the cortex. Subject ONH1’s response maps revealed an interesting pattern: tactile responses were greater in regions that corresponded to the highly represented eccentricities. More subjects should be evaluated to determine if this response pattern is specific to this patient population (resulting from optic nerve defects) or if it will also appear in RP subjects. If this result is reproducible and consistent across subjects, and if cross-modal responses are shown to have an adverse effect on the brain’s ability to process the signal from a retinal prosthesis, retinal eccentricity maps may be used to better understand where electrodes should be placed on the retina to optimize cortical processing of the device’s signal. 74 | P a g e Chapter 5: The effect of vision loss on functional connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams 5.1 Overview The existence of increased tactile-evoked BOLD responses in V1 with vision loss prompts the question of whether changes in connectivity between the visual areas and other regions of the brain (such as the somatosensory cortex) are a source of this activity. Functional connectivity between SI and the visual areas has been established in early blind patients, with subjects exhibiting V1 activation after rTMS stimulation to S1 (Wittenberg et al., 2004), while effective connectivity studies suggest that an indirect cortico-cortical feedback pathway from S1 to VI exists that is modulated by vision loss, resulting in an expansion of tactile processing into the visual areas (Bavelier et al., 2002 and Fujii et al., 2009). These studies concluded that the dorsal visual stream (specifically visual association areas of the parietal lobe) may be a region where visuo-tactile 75 | P a g e information is integrated in sighted individuals, while blindness may cause both dorsal and ventral visual streams to process tactile stimuli. Understanding how vision loss affects these functional pathways is a first step towards accounting for performance variability in Argus II patients. If the functionality of essential visual pathways is altered, it may hinder an individual’s ability to process the implant’s signal. The following study sought to understand the functional relationship between the visual cortex and other regions of the brain, and the effect of vision loss on these functional connections, by addressing the following objectives:  We determined if a relationship exists between the task-state residual activities of V1 and S1 (where tactile processing originates) that would shadow a functional connection between the two regions  Significant resting-state correlations were identified within the dorsal visual pathway (including posterior parietal association areas), ventral visual pathway, and map a significant resting-state network between V1 and SI  We determined if the strength of these resting-state connections are affected by changes in visual function 5.2 Study Design 5.2.1. Participants Two sets of subjects were used in the following analyses. A task-state connectivity analysis was completed using the same set of 18 subjects described in Table 1 who completed our fMRI tactile-discrimination study. The second set of subjects underwent additional resting-state scans and included seven late-blind RP patients and 4 sighted subjects recruited from the original group of 18 subjects (Table 4). Subjects within the second set had a mean ± SD age of 40.09 ± 13.01 years (range: 21-58 years); sighted control subjects were gender-matched and had a similar age range (24-57 years) to the RP subjects (21-58 years). Within the second set of subjects, RP patients had a range of vision loss, from minimal light perception only to a visual acuity of 20/20, allowing us to account for variability among the subject group. 76 | P a g e Table 4. Resting-State Analysis Subject Demographics. Normally sighted subjects S4-S8 served as sighted, gender-matched control subjects for RP subjects RP1-RP9. L.P. = light perception only. 5.2.2. Image Acquisition MR images were acquired in a 3T Siemens MAGNETOM TIM Trio scanner using a 12- channel Matrix head array coil. Anatomical images were obtained using a T1-weighted sequence (MPRAGE), while functional (resting-state) images with BOLD contrast were acquired using an EPI sequence with TR/TE/flip angle = 2s/30ms/80° and PACE. 37 slices with isotropic voxels of 3x3x3 mm 3 were axially oriented and covered the entire cerebrum. Subjects laid head first and supine in the scanner. Foam padding was placed around the head to minimize movement during scanning, while earplugs and sound-attenuating headphones were provided to dampen scanner noise. All subjects were blindfolded and asked to rest passively in the MRI machine during the resting-state scans. 5.2.3. fMRI Data Analysis Image data was analyzed using BrainVoyager QX (Goebel et al., 2006) in subjects’ native space (as opposed to normalizing to a standard space). Anatomical data underwent inhomogeneity correction and were reoriented via rigid-body rotation and translation to place the origin at the Anterior Commissure and the Posterior Commissure on the y-axis. All functional data was preprocessed with 3D motion correction (PACE and post hoc), slice timing correction, and 77 | P a g e temporal filtering—spatial smoothing was not applied to the functional data. Significant voxel- wise activations were identified at FDR < 0.05 with a cluster threshold of 25mm 2 . Task-state correlation analysis The relationship between the underlying activities of V1 and S1 were first estimated using functional data obtained from all 18 subjects during the tactile tasks (see section 3.3). The residual time courses of the modulated voxels in both regions was defined as the BOLD response left after regressing out activities attributable to task, head movements, ventricle and white matter activity. A second-order AR(2) model was used to remove serial correlations from the residual time courses. The correlation, in terms of R 2 , between the residual time courses of V1 and S1 was taken to index the “task-state correlation” between the two regions in each subject. This preliminary estimate of functional connectivity was then entered into a multiple regression model, along with visual acuity, to determine if it was related to the amount of tactile-evoked BOLD activity in V1. Resting-state fMRI analysis ROI templates for the resting-state analysis were obtained from FSL’s MNI Structural Atlas (Smith et al., 2004). Eleven ROIs were selected to include voxels in S1 and along both ventral and dorsal visual pathways, including: V1-V5 of the visual cortex, middle temporal cortex (MT), inferior parietal area (IP), superior temporal cortex (ST), inferior temporal cortex (IT), and the lateral occipital cortex (LOC). These templates were first thresholded in FSL by 5-45% to ensure that the regions did not overlap with one another on a standard MNI152 T1 2mm resolution template brain. The thresholded ROIs were then transformed from MNI space to subject ACPC space in BrainVoyagerQX (Table 5). 78 | P a g e Table 5. List of ROIs used in the following resting-state and DTI analyses. The thresholded ROIs used in the following analyses are displayed in red-orange on a standard MNI152 template. 79 | P a g e Image data from the resting-state scans was used to assess the functional connectivity between ROI pairs in each subject. We defined the residual time course for each ROI as the BOLD response remaining after having removed activities attributable to head movement, white matter activity, non-neuronal responses in the ventricles, and temporal correlation in the residual data (via an AR(2) model). The correlation, in terms of r, between the residual time courses of each ROI in a pair was taken to define the functional connectivity between the two ROIs. The resulting r-values were evaluated as a function of vision loss (i.e. visual acuity and fractional preserved visual field in V1), and the extent and strength of tactile-evoked responses in V1 (as described in Chapter 3). The significance threshold was corrected for multiple comparisons (α = [0.05/55 comparisons] = 0.001). A series of partial correlations were calculated to determine the functional connectivity between two ROIs in the absence of any influence from a third ROI. Specifically, partial correlations were determined among the 11 ROIs when controlling for residual responses in either SI, IP, MT, IT, ST, or V1. These results were then combined to estimate a significant resting-state network between V1 and SI. For example, previous studies have suggested that the parietal lobe may act as a tactile processing hub zone between the visual and somatosensory cortices. We therefore evaluated the correlation, in terms of r, between the residual time courses of V1 and S1 when controlling for residual responses in IP. If there was originally a significant correlation between V1 and SI, a non-significant partial correlation between V1 and S1 would suggest that area IP is an essential component of the functional network between V1 and SI, while a still significant partial correlation would suggest that IP is nonessential and should not be included in a V1—S1 functional map. The resulting r-values were evaluated as a function of visual acuity, fractional preserved visual field in V1, and the extent and strength of tactile-evoked responses in V1. To evaluate differences in partial correlations between the RP and sighted subject groups, the r-value for a partial correlation (e.g. between V1 and S1, controlling for IP) for each subject was converted to a z-score using Fisher’s transform. These z-scores were averaged across all subjects in the group. The group averaged z-score was then normalized by dividing by the standard error of the mean (SEM); p-values for each group were determined based on this normalized averaged z-score. 80 | P a g e 5.2.4. Statistical Modeling In order to understand how the functional relationship between S1 and both dorsal and ventral visual pathways differs between sighted and late-blind RP groups, a path analysis was conducted using Lavaan—a latent variable modeling program for R (Yves et al., 2012)—to map a network that included all 11 ROIs. Due to errors that would arise from comparing groups with an unequal number of subjects, the RP group only included 4 RP subjects with the most severe vision loss (RP1, RP2, RP4, and RP5). A structural equation model was used to model the covariances between the residual time courses of each ROI for each subject group, where each ROI was included as an endogenous variable. The resulting beta weights were used as path coefficients in a resting-state network model. Separate resting-state maps were created for both subject groups for comparison. 5.3 Results 5.3.1. Task-state Correlation Between Tactile-Evoked Responses in V1 and SI The insignificant pseudo-partial correlation between tactile-evoked responses in V1 and S1 when controlling for vision loss (see section 3.3.3) suggests that V1 tactile-activities are not a result of some direct connection between V1 and S1. To further evaluate this idea, we estimated the underlying (task-state) correlation between V1 and S1 for each subject by correlating the residual BOLD time courses of the modulated voxels in these regions, after having removed the task-evoked activities, head motion, and temporal correlation in the residual data (via an AR(1) model). The task-state (i.e. residual) correlation between V1 and S1 was highly significant [p’s < 0.001] for the majority of RP and sighted subjects for each task (Table 6). 81 | P a g e Table 6. Residual correlations (r) between V1 and S1 for each subject and task. Significant correlations (p < 0.05) are indicated by an asterisk (*). A linear mixed effects model found a significant effect of visual acuity on the residual correlation between V1 and S1, quantified by R 2 [p = 0.012, (dependent variables: V1—S1 residual correlation; covariates: fractional preserved visual field in V1 or visual acuity; repeated variables and factors: Task – Shapes, Braille, and Sandpaper)]. No effect of preserved visual field was found on the V1—S1 residual correlation [p = 0.172]. We asked if the task-state correlation between V1 and S1 and visual acuity would together provide a better account of the extent and strength of tactile-evoked responses in V1. In the context of a linear multiple regression model that included visual acuity and the residual correlation between V1 and S1, we found a main effect of both on the extent and strength of the V1 response [extent: F(2,52) = 20.140, p < 0.001; strength: F(2,52) = 3.696, p = 0.032]. This suggests that the strength in any preexisting connection between V1 and S1, in conjunction with an individual’s visual acuity, may be used to predict tactile-evoked responses in V1 following vision loss. 5.3.1. Mapping Resting-State Connectivity between V1, SI, and along the Visual Stream The significant residual correlation between V1 and S1 suggests that the two regions are functionally connected. This notion was further evaluated by determining the functional resting- 82 | P a g e state connectivity between the two regions, as well as amongst 9 additional ROIs located along the ventral and dorsal visual streams (Table 5). The functional connectivity between two regions was defined as the correlation between each ROI’s residual signal obtained from a resting-state scan (during which no sensory stimulus or task was presented). A significant resting-state correlation was found among all 55 ROI pairs for the majority of RP subjects and all sighted subjects (p’s < 0.001 after correction for multiple comparisons, Figure 31). The strongest correlations existed in the striate and extrastriate cortex (V1—V4): given that these regions comprise the visual cortex, it is expected that they would be highly functionally connected. 83 | P a g e Figure 31. Color maps depicting strength of resting state correlations (r) between pairs of ROIs for 4 representative RP subjects and 4 sighted control subjects. Stripped pale regions indicate insignificant correlations (p > 0.001, correcting for multiple comparisons). 84 | P a g e No significant correlation was found between visual function (i.e. preserved visual field or visual acuity) and functional connectivity (in terms of r) between any of the 55 ROI pairs (p’s > 0.001). Similarly, no relationship was found between the extent or strength of tactile-evoked responses in V1 and functional connectivity amongst these ROIs. A path analysis comparing both sighted and RP groups revealed that both groups had a statistically identical visual stream network—all relationships among the 11 ROIs were significant (p’s < 0.001). However, path weights were stronger (but not significantly so) between some ROIs in the RP group when compared to the sighted group—including within the temporal cortex, and between the temporal cortex and visual areas (Figure 32)—that may be a target for future analyses. Figure 32. Resting-state network including all 11 ROIs among RP subjects (left) and sighted subjects (right). Arrows indicate the covariance between two ROIs based on a structural equation model. The thickness of the arrows illustrate the strength of the residual correlation between ROIs (where thicker arrows indicate a stronger functional connection). Within this network, we determined which ROIs were essential to a functional pathway between V1 and SI. A significant resting-state correlation was found in both RP and sighted groups (p’s < 0.01, Figure 33A). However, a partial correlation removing the effect of area IP decreased 85 | P a g e the correlation between these two regions among the RP subjects (p = 0.269, Figure 33C), but not sighted subjects (Figure 33B). This suggests that a significant resting-state connection between V1 and S1 exists that is mediated by the parietal area in late-blind subjects. In addition, controlling for the effect of S1 in sighted subjects rendered the relationship between areas V1 and IP insignificant (p = 0.278, Figure 33C). Figure 33. A: Illustration of the functional connectivity between V1 and S1 for both RP and sighted subjects. Correlations are indicated by a normalized averaged z-score for each group and corresponding p- value. B and C: Illustration of pairwise correlations between residual time courses of areas S1, IP, and V1. The partial correlation between the two quantities connected by an edge in each diagram represents the correlation between the two quantities while excluding the effect of the third (e.g. C: the bottom bold line represents the correlation between V1 and IP resting-state activities while excluding the effect of activities in area S1). The three ROIs are displayed on a cortex to the right. Consistent with this finding, a significant positive correlation was found between the strength of the partial correlation between V1 and S1 (controlling for area IP) and visual field across both RP and sighted subjects (r = 0.655, p = 0.029), where more severe vision field loss results in IP playing a larger mediating role between V1 and S1. A significant negative correlation 86 | P a g e was also found between the strength of the partial correlation between V1 and IP (controlling for area S1) and visual field across both RP and sighted subjects (r = -0.664, p = 0.026), where more severe vision loss results in S1 playing a smaller role in the relationship between V1 and IP. 5.4 Discussion Our findings indicate that the dorsal and ventral visual streams are significantly functionally connected. A significant functional connection also exists between areas V1 and S1 that is more greatly mediated by the inferior parietal area as vision loss progresses. This is consistent with the known architecture of the dorsal visual stream, which has been implicated in contributing to tactile processing in V1. However, the overall functional connectivity of the visual streams was not found to be affected by vision loss in late-blind RP patients, and was not significantly related to tactile-evoked BOLD responses in V1. This finding leads to the question of whether a separate mechanism exists that causes the responsivity of V1 to non-visual information to change in such a way that is not reflected in spontaneous resting-state BOLD activities. Bavelier et al. 2002 outlines several theories that currently exist to explain how multisensory association areas (like the parietal lobe) reorganize following sensory deprivation, including changes in: local synaptic connectivity, long-range subcortical connectivity that typically occurs during the early-developmental stages of life, and cortico-cortical feedback that may be influenced by sensory input. The mechanisms mediating cross-modal plasticity can have a significant impact on an individual’s ability to adapt to a retinal prosthesis. If these mechanisms allow the functional system to quickly adapt to the signal generated by the device, cross-modal plasticity need not be seen as a hindrance to sight restoration treatments. However, if these changes are not easily reversible, there is a greater likelihood that they will interfere with an Argus II patient’s progress following treatment. In this study’s case, a lack of large-scale functional changes with vision loss suggests that the same long-range connectivity involved in processing normal vision is still available to process the retinal signal generated by a visual prosthesis. Cross-modal responses thus likely result from small-scale changes in synaptic connectivity or cortico-cortical feedback that are both highly plastic in young and adult subjects alike. This is promising for both retinal prosthetic devices and other vision restoration treatments that assume that connectivity of the visual streams remain intact. 87 | P a g e Chapter 6: The effect of vision loss on structural connectivity between primary visual and somatosensory cortices, and within the dorsal and ventral visual streams 6.1 Overview Several studies have determined that functional connectivity generally follows some underlying structure (Greicius et al., 2008). We therefore used several diffusion tensor imaging techniques to determine if the strength of functional connections along the dorsal and visual streams is correlated with the structural connections in the same areas. While a stable functional connectivity could be beneficial to sight restoration treatments, changes in white matter integrity could make it difficult for an individual to adapt to restored vision: any signal generated by a retinal prosthetic device could not follow the same cortical pathways as normal vision if their vision loss has compromised those channels. Understanding the effect of vision loss on white matter integrity 88 | P a g e in the cortex could provide additional information to help us predict how an individual will respond to treatment. A DTI study conducted with late-blind glaucoma patients found decreased diffusivity along the optic radiation (Chen et al., 2012). However, other studies have demonstrated that structural connectivity changes are minor or absent in individuals who lose their sight later in life. Li et al., 2012 found that age of onset of blindness is negatively correlated with the mean number of connections throughout the entire cortex, with late-blind patients having a greater connectivity density (similar to that of sighted individuals) than early or congenitally blind patients. Other findings have demonstrated that late-blind RP subjects exhibit changes in the cortical thickness of these regions when compared to their sighted counterparts that are correlated with years since onset of blindness (Park et al., 2009). Given that RP patients typically begin to lose their sight as adults, we do not expect to find significant changes in white matter integrity of pathways associated with visual processing (excluding the optic radiation). This will confirm that both functional and structural pathways remain intact and are available to process the signal from a retinal prosthesis. Here, we sought to determine how the structure of the dorsal and ventral visual pathways, and pathway between V1 and SI, are affected by vision loss and if any changes in structural connectivity are the source of cross-modal responses in V1. We further determined if functional connectivity along the visual pathway follows any underlying structure. This was accomplished through the following goals:  DTI was used to estimate diffusion tensor parameters and probabilistic connectivity of the dorsal visual pathway, ventral visual pathway, and pathway between V1 and S1 in both subject groups.  We used tract-specific analyses (TSA, Zhang et al., 2010) to determine if vision loss also leads to localized changes in visual stream white matter fiber tracks  The cortical thickness of 11 ROIs was also determined and correlated with visual function.  We determined if a relationship exists between functional and structural connectivity across all ROI pairs 89 | P a g e 6.2 Study Design 6.2.1. Participants Two sets of subjects were used in the following analyses. A cortical thickness analysis was completed using the same set of 18 subjects described in Table 1 from whom we collected an anatomical data set. The second set of subjects consisted of seven RP and 2 sighted subjects who had previously participated in our fMRI tactile-discrimination study and from whom we acquired resting-state data (as described in Chapter 4). Subjects in the second set had a mean ± SD age of 38.22 ± 13.10 years (range: 21-58 years); sighted control subjects had an age range of 24-57 years, while RP subjects ranges from 21-58 years old (Table 7). Within the second set of subjects, RP patients had a range of vision loss, from minimal light perception only to a visual acuity of 20/20, allowing us to account for variability among the subject group. Table 7. DTI Analysis Subject Demographics. Subjects comprise the second set of subject who underwent a series of DTI scans, as described above. 6.2.2. Image Acquisition MR images were acquired in a 3.0T Siemens MAGNETOM TIM Trio scanner using a 12- channel Matrix head array coil. Anatomical images were obtained using a T1-weighted sequence (MPRAGE). Diffusion-weighted images were acquired using a diffusion-weighted imaging sequence (TR/TE = 10s/88ms) and includes 60, 2mm thick slices with isotropic voxels of 2x2x2 mm 3 . A DTI protocol with 30 diffusion directions, each having a b-value of 900 s/mm 2 , was used Subject ID Age, Gender Visual Acuity Years since onset of symptoms Can Subject Read Braille? Diagnosis Additional Description of Vision RP1 41, F L.P. 22 Yes RP Light perception only RP2 52, M 20/60 46 No RP Partial tunnel vision; cataract removal from both eyes RP4 43, M 20/800 38 Yes RP Partial tunnel vision in right eye RP5 58, M 3/200 23 No RP Cataract removal from both eyes RP7 24, F 20/25 11 No RP Partial tunnel vision RP8 51, F 20/40 44 No RP Loss of peripheral vision, night blindness, blurred vision in right eye RP9 21, M 20/20 1 No RP Beginning loss of peripheral vision and some night blindness S4 24, M 20/20 --- No --- Sighted S5 30, M 20/20 --- No --- Sighted 90 | P a g e that included 4 repetitions of a 5:40 minute whole-brain scan. Subjects laid head first and supine in the scanner. Foam padding was placed around the head to minimize movement during scanning, while earplugs and sound-attenuating headphones were provided to dampen scanner noise. 6.2.3. MRI Data Analysis Cortical thickness of regions along the visual stream The cortical thickness of each ROI (Table 5) was calculated using BrainVoyagerQX. V1 and S1 ROIs were manually drawn for each subject according to section 3.2.5. Cortical thickness maps were measured by first estimating the gray matter—white matter and white matter—CSF (corticospinal fluid) boundaries in both cortical hemispheres for each subject. Tissue classification was accomplished using BrainVoyager’s segmentation algorithm which assigned one of the three tissue types depending on a voxel’s intensity value i: CSF (i < 75), GM (75 ≤ i ≤ 125) and WM (i > 125 ). A thickness map was then generated based on the number of voxels in each tissue group. The average grey matter thickness (in mm) was calculated for each of 11 ROIs in each subject and evaluated as a function of vision loss and tactile-evoked responses in V1. After preprocessing, all subjects’ cortical meshes were aligned to a standard mesh sphere, allowing individual thickness maps to be compared on a single cortically aligned cortex mesh. ROIs were similarly registered to the standard mesh. For each ROI, a t-test was conducted to determine if any significant differences in average thickness existed between sighted and RP subjects. Analysis of white matter integrity parameters and probabilistic connectivity White matter integrity was measured by estimating three structural parameters using DTI: mean diffusivity (MD), fractional anisotropy (FA), and the probability that fiber bundles exist between two regions. DTI is used to measure the diffusivity of water molecules in the brain, with water typically diffusing more quickly along the length of an axon versus across an axon. This diffusion is modeled using tensors to describe both the strength and direction of diffusion (where isotropic diffusion indicates diffusion in all directions, and anisotropic diffusion describes diffusion of molecules along one direction—a larger FA suggests that fiber bundles are present in a region). Tensors may be compared among voxels in a region to determine the general pathway of fibers. These can also be used to obtain maps describing the probability that fibers exist between two regions of interest. 91 | P a g e All subjects completed 3-4 repetitions of a DTI scan—these separate data sets were concatenated in time to create a single diffusion weighted data set for each subject. This data was then processed using FSL’s diffusion toolbox (Behrens et al., 2003b, 2007 and Woolrich et al., 2009), and included: eddy current correction, fitting tensors to the corrected data, using a Bayesian estimation of the diffusion parameters, and fitting of a probabilistic model to the corrected data. The resulting diffusion tensor images were then co-registered to each subject’s anatomical data in MNI space. This process yielded FA and MD values throughout the cortex. The same set of 11 visual stream ROIs used in the resting-state analysis were selected— these ROIs were also in MNI space. Probabilistic tractography was used to calculate the probability of white matter tracts existing between two ROIs, based on the diffusion parameters estimated above (including principal diffusion direction estimations). Each ROI was used as a seed mask in MNI space, so that pathways only included tracts that initiated in one mask and terminated in a second or vice versa. Pathways were thus estimated for all 55 ROI pairs and were threshold to only include tracts with a 50% chance or more of existing between the two ROIs. From the resulting white matter tract files for each ROI pair, we calculated the probability that two ROIs are connected by fibers according to the following equation: 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 = 𝑊𝑎𝑦𝑡𝑜𝑡𝑎𝑙 𝑇𝑜𝑡𝑎𝑙 # 𝑆𝑡𝑟𝑒𝑎𝑚𝑙𝑖𝑛𝑒𝑠 Equation 3. where the waytotal is the total number of tracts generated between ROIs A and B, and the total number of streamlines is the number of samples per voxel (5000) across all voxels in both ROIs: 𝑇𝑜𝑡𝑎𝑙 # 𝑠𝑡𝑟𝑒𝑎𝑚𝑙𝑖𝑛𝑒𝑠 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑅𝑂𝐼 𝐴 𝑎𝑛𝑑 𝐵 = 5000 ∗ [(# 𝑣𝑜𝑥𝑒𝑙𝑠 𝑖𝑛 𝑅𝑂𝐼 𝐴 ) + (# 𝑣𝑜𝑥𝑒𝑙𝑠 𝑖𝑛 𝑅𝑂𝐼 𝐵 )] Equation 4. The non-zero voxels represented voxels in the 3D image file that had a non-zero probability of being part of a tract between the specified ROIs. The average FA and MD values along each of the 55 pathways, and p values representing the probability of tracts existing between two ROIs were calculated for each ROI pair. These parameters were evaluated as a function of vision loss (visual acuity and fractional preserved visual field in V1) and tactile-evoked BOLD responses in V1. The average probabilistic connectivity (𝑝 ̅) between two ROIs and the average resting-state correlation (𝑟 ̅) were calculated across subjects for each ROI pair after controlling for the distance 92 | P a g e in mm between ROIs. These two parameters were then correlated to determine if structure and function are related to one another along the visual streams. Tract-Specific Analysis (TSA) of regions related to the visual areas Diffusion weighted data was used after eddy current correction and tensor fitting in FSL (Smith et al, 2004). Diffusion tensor images for all subjects were registered to a common adult atlas using DTI-TK (Yushkevich et al., 2008) that included medial-representations of 4 white matter tracts: the corpus callosum (CC), inferior fronto-occipital tracts (IFO), inferior longitudinal tracts (ILF), and superior longitudinal tracts (SLF), all of which have connections in the occipital lobe. Tract-specific regression analyses were conducted to identify significant relationships between subjects’ visual function (i.e. visual acuity and fractional preserved visual field in V1) and mean radial diffusivity (RD), apparent diffusion coefficient (ADC), axial diffusivity (AD), and maximum fractional anisotropy (FA) along the 4 tracts. We further defined the functional connectivity along a tract as the correlation (r) between the time courses of 2 ROIs (one at each end of a tract) after regressing out head-motion parameters. 6.3 Results 6.3.1. Cortical Thickness of Regions along the Dorsal and Ventral Visual Streams The mean gray matter thickness across all 11 ROIs ranged from 3.87mm—4.30mm for RP subjects and from 3.77mm—4.94mm for sighted subjects (Figure 34). Among the 11 ROIs chosen, no significant changes in cortical thickness were found as a function of preserved visual field in V1 or visual acuity (p’s > 0.05). A t-test describing significant differences in the thickness of each ROI between sighted and RP subject groups found that the two groups were similar for every ROI except the right hemisphere of V4 (p = 0.024) and V5 (p = 0.04). No significant correlation was found between cortical thickness and the age of each subject (in both sighted and RP groups). Among RP subjects only, no significant correlation was found between cortical thickness and years since onset of blindness for any of the 11 ROIs. 93 | P a g e Figure 34. Comparison of cortical thickness (mm) in 2 representative RP subjects and 1 representative sighted subject. Color gradient represents gray matter thicknesses from 0.50mm—5.00mm. Cortical thicknesses (averaged across both hemispheres) are given for each of 11 ROIs. Subjects are placed in order of increasing visual function. Increasing Visual Function 94 | P a g e 6.3.2. Analysis of White Matter Integrity using FA and MD Values White matter integrity was generally defined in terms of fractional anisotropy and the mean diffusivity between two regions of interest. No significant relationship was found between the FA values of tracts connecting two ROIs in each pair and visual function (i.e. preserved visual field or visual acuity) after correcting for multiple comparisons (p’s > 0.001, Figure 35). However, uncorrected results revealed a possible significant positive correlation between preserved visual field and FA for the ROI pairs shown in Table 8 (p’s < 0.05). Similarly, visual acuity was found to be significantly negatively correlated with FA values along the ROI pairs shown in Table 8 (uncorrected p’s < 0.05). 95 | P a g e Subject FA Maps RP1 RP4 RP8 RP9 S4 S5 Figure 35. FA maps for 4 representative RP subjects and 2 representative sighted subjects. Principle diffusion directions are represented by color: red indicated fibers that run right-left, blue indicated inferior- superior, and green indicates anterior-posterior. Subjects are placed in order of increasing visual function. Increasing Visual Function 96 | P a g e No significant correlations were found between MD values along the 11 tracts and visual function after correcting for multiple comparisons. However, similar to the above FA correlations, uncorrected results suggested that a negative relationship between preserved visual field and MD values exists along tracts connected to regions V1, V2, V3, V4, V5, LOC, IT, IP, and S1 (p’s < 0.05, Table 8). Visual acuity may positively affect MD values along tracts connected to regions V1, IT, LOC, IP, and S1 (p’s < 0.03, Table 8). Table 8. Correlation values (r) and p-values for ROI pairs whose FA and/or MD values are significantly correlated with preserved visual field or visual acuity results across subjects. Significance is indicated by p-values that have not been corrected for multiple comparisons (α = 0.05). 6.3.3. Probabilistic Connectivity of the Visual Streams No significant correlations were found between the probabilistic connectivity between each ROI in a pair and visual function (i.e. preserved visual field or visual acuity) across RP and sighted subjects when correcting for multiple comparisons (p’s > 0.001, Figure 36). However, uncorrected results suggest that preserved visual field may be positively correlated with probabilistic connectivity between V1—MT and V2—MT (p’s < 0.05). A trending correlation was also found 97 | P a g e for V3—MT (p = 0.069) and V4—MT (p = 0.055). Visual acuity may negatively affect the probabilistic connectivity between V1 and V2 (p = 0.045). Figure 36. Probabilistic connectivity of ROI pair IP—MT for 4 representative RP subjects and 2 sighted subjects. Left: Maps are thresholded to only include voxels with a 50% chance or more of being part of a fiber tract between the two regions. Subject are placed in order of increasing visual function. Right: Image of the ROIs used in the analysis overlaid on a standard MNI152 cortex template. 6.3.4. Tract-Specific Analysis of Visual Stream Connectivity Regression analyses with FWER correction revealed no significant relationship between visual function (i.e. preserved visual field in V1 or visual acuity) and mean FA, RD, ADC, AD, or maximum FA values along the 4 white matter tracts (FWER corrected min. p = 0.181, Figure 37). However, uncorrected results show a possible significant effect of visual field on ADC values (p’s < 0.01). 98 | P a g e Figure 37. Tract-specific regression analysis results. Left: Illustration of regions in each tract with a significant linear correlation (p < 0.01) between mean FA values and preserved visual field across subjects. Significant regions are outlined in white. Right: Illustration of regions in each tract with a significant linear correlation (p < 0.01) between ADC values and preserved visual field across subjects. The left side of each column displays significance based on FWER correction, while the right side of each column displays uncorrected p-values. All tracts are arranged with the anterior region on the left and posterior region on the right. A significant resting-state correlation (indicating functional connectivity along the tract) was found between the two ends of each tract for the majority of subjects (Table 9). A regression analysis of resting-state connectivity and visual function revealed a trending correlation between the functional connectivity of SLF and subjects’ preserved visual field (r = -0.749, FWER p = 0.078). No significant correlations between functional connectivity and visual function were found for the remaining 3 tracts. 99 | P a g e Table 9. Resting-state correlations (r) between the two ends of each tract and corresponding p-values for each subject. Significant correlations (p < 0.05) are indicated by an asterisk (*). 6.3.5. Comparison of Structural and Functional Connectivity along the Visual Streams While vision loss does not generally seem to be related to changes in functional or structural connectivity along the cortical visual streams, we explored whether resting-state connectivity followed some underlying structure. A significant positive correlation was found when comparing the average resting-state connectivity and average probabilistic connectivity of each ROI pair (p < 0.0001, Figure 38): a stronger functional connection between two ROIs means a stronger white matter connection also exists. It is conceivable that the distance between the ROIs in each pair could result in a false correlation: the closer together two ROIs are, the stronger their resting-state or structural connectivity may be. However, this correlation remained highly significant even after controlling for (partialling-out) the distance between the ROIs in each pair (p < 0.0001). 100 | P a g e Figure 38. Correlation between average resting-state connectivity and average probabilistic connectivity of each ROI pair. For each ROI pair, resting-state and probabilistic connectivity values were averaged across subjects. 6.4 Discussion The absence of changes in gray matter thickness or white matter integrity as vision loss becomes more severe suggests that those regions are not plastic in adulthood and thus not affected by late-onset blindness, or that the regions are still being used to process information. Previous studies have suggested several possible mechanisms of cross-modal plasticity, including the reallocation of the visual areas for processing other types of sensory information. Our findings may support the notion that regions along the dorsal and ventral visual streams retain their thickness because the underlying neurons are still in use to process non-visual information (e.g. tactile and auditory stimuli), as well as participating in feedback processes from other regions of the brain. This consistent gray matter thickness and white matter integrity across RP and sighted subjects may also be a result of RP patients going blind in adulthood: previous studies have found that late-blind individuals exhibit fewer structural changes in vision-related regions when compared to early-blind or congenitally blind individuals (Li et al., 2012). In addition, cortical thickness was not found to vary with age among our subjects. This seems contrary to several studies that have found a decrease in the thickness of gray matter as an individual ages (Tisserand et al., 2004 and Ge et al., 2002). However, the regions most associated with decreased gray mater thickness (such as the prefrontal lobe) were not included in this analysis 101 | P a g e (Thambisetty et al. 2010). It is also possible that the age range of our sighted and RP subjects was not large enough to account for this phenomenon. And if you have read my thesis up until this point, yay! As a token of thanks for opening my humble dissertation, here is a $10 amazon gift card—please visit the link below 9 and enter in the following claim code: T8GZ-7JSQJZ-28DA. In accordance with these findings, we found no significant correlations between visual function and white matter integrity along 55 vision-related tracts. However, a significant relationship was found between structure and function along the visual stream, coinciding with previous studies showing that functional connectivity is related to some underlying structure in healthy individuals (Greicius et al., 2008). In all cases, a lack of significance may be due to a small sample size: additional subjects should be studied to verify these results. However, it is equally likely that vision is not related to structural changes in our late-blind subject group. If this holds true with additional subjects, this result is promising for blindness restoration treatments, particularly retinal prosthetic devices like the Argus II. These treatments seek to stimulate other retinal layers by by-passing the degraded photoreceptors—it is necessary that the underlying cortical structure of the visual stream remain intact to process the novel signals being generated in the retina. 9 www.amazon.com/redeemgift 102 | P a g e PART III: THE EFFECT OF RETINAL PROSTHESIS USE ON THE VISUAL CORTEX If cross-modal plasticity is described in terms of a spectrum, late-blind individuals with tactile-evoked responses in V1 would exist on one end and sighted individuals with little-to-no tactile responses in V1 on the other end (Figure 39). Our findings have allowed us to define both ends of this spectrum, leaving the question of how tactile-evoked responses in V1 are affected by partial sight restoration. Figure 39. Illustration of the differences between sighted and low vision individuals. V1 in RP patients becomes more responsive to tactile-stimulation as the vision loss progresses. We hypothesize that long-term users of a retinal prosthesis will show a decrease in tactile-evoked responses in V1, and therefore begin to more resemble the extent and strength of responses in the sighted population. 103 | P a g e Chapter 7: Effect of extended Argus II prosthesis use on tactile-evoked BOLD responses in V1 7.1 Overview While vision deprivation has been shown to produce tactile-evoked cross-modal responses in primary visual cortex, little is known about the effect of vision restoration on these activations. Studies in sighted individuals (Merabet et al., 2008) have found that tactile-evoked cross-modal responses can develop in the visual cortex after only 5 days of vision deprivation and are reversed as soon as vision is restored. The question remains as to whether tactile responses in the visual cortex of late-blind individuals will similarly decrease after an RP patient has been re-exposed to vision. By determining how cross-modal responses in V1 are affected by short-term and long-term use of the Argus II device, we will have an indirect measure of retinal stimulation’s effect on the visual cortex and can better understand how vision restoration with a retinal prosthesis affects cross-modal plasticity in the brain. If these pre-implantation tactile-evoked responses can be correlated with an individual’s performance with an Argus II prosthesis, cross-modal activity in V1 may be used as a biomarker to help predict patient outcomes following sight restoration treatment. Characterizing these effects, 104 | P a g e as well as how the visual cortex responds to artificial vision, can further help in identifying the sources of patient visual performance variability following receipt of a retinal prosthesis. The following objectives were investigated:  Functional MRI was used to acquire V1 responses to 3 tactile tasks in Argus II patients following 1) a period of consistent device use and 2) a period of no device use.  Responses were compared between these two conditions to determine the effect of retinal prosthesis use on cross-modal responses in V1.  Responses were also compared to those in 18 RP and sighted individuals. 7.2 Study Design 7.2.1. Participants Two female subjects (ages 55 and 79 years old) participated in the study (Table 10). Subject A1 was implanted with the Argus II retinal prosthetic system 6 weeks before our study, while subject A2 received the device 15 weeks prior to our study. In accordance with inclusion and exclusion requirements to receive the device 10 , both individuals had been diagnosed with retinitis pigmentosa and had a visual acuity of 2.3 logMAR or worse at the time of implantation. Subjects were screened by their ophthalmologists prior to recruitment and had previously been advised to use their device for 2-6 hours per day. Table 10. Argus II subject demographics. The study received approval from the University of Southern California’s Health Sciences Campus Institutional Review Board and all subjects provided written informed consent after explanation of the nature and possible consequences of the study. MRI experiments were conducted at the USC David and Dana Dornsife Cognitive Neuroscience Imaging Center, while 10 <http://clinicaltrials.gov/show/NCT00407602> Subject ID Age, Gender Visual Acuity Date of Implantation Implanted Eye Can Subject Read Braille? Diagnosis Average device usage (total time with device) A1 55, F < 2.3 logMAR June 2014 Left No RP 2-5 hours/day (6 weeks) A2 79, F < 2.3 logMAR February 2014 Right No RP 5-6 hours/day (15 weeks) 105 | P a g e indirect ophthalmic exams were conducted by an ophthalmologist from the USC Eye Institute. Subjects received monetary compensation for their participation. This research followed the tenets of the Declaration of Helsinki. 7.2.2. Image Acquisition MR images were acquired in a 3 Tesla Siemens MAGNETOM TIM Trio scanner using a 12-channel Matrix head array coil. Two sets of anatomical images were obtained for the preoperative MRI session (completed by subject A1) using a T1-weighted sequence (MPRAGE) with TR/TE/flip angle/slice thickness = 1.95s/2.26ms/9°/1.2mm and an MPRAGE with TR/TE/flip angle/slice thickness = 2.3s/2.98ms/9°/1.0mm. Postoperative anatomical images were obtained using a T1-weighted sequence with TR/TE/flip angle/slice thickness = 1.95s/4.44ms/12°/1.0mm for increased image resolution given the presence of the implant. Functional images with BOLD contrast for the tactile and visual tasks were acquired using an EPI sequence with TR/TE/flip angle = 2s/25ms/60° and PACE. 36 slices with isotropic voxels of 3x3x3 mm 3 were axially oriented and covered the entire cerebral cortex except for the tip of the temporal lobe for some subjects. Functional (resting-state) images with BOLD contrast were acquired using an EPI sequence with TR/TE/flip angle = 2s/30ms/80° and PACE. 37 slices with isotropic voxels of 3x3x3 mm 3 were axially oriented and covered the entire cerebrum. Diffusion- weighted images were acquired using a diffusion-weighted imaging sequence (TR/TE = 10s/88ms) and included 60, 2mm thick slices with isotropic voxels of 2x2x2 mm 3 . A DTI protocol with 30 diffusion directions, each having a b-value of 900 s/mm 2 , was used that included 3 repetitions of a 5:40 minute whole-brain scan. 7.2.3. Experimental Stimuli and Tasks Both subjects completed the same three tactile tasks as described in Chapter 3 (see section 3.2.2), including: 1) a shapes task requiring subjects to determine if any of a series of raised-line shapes was bilaterally symmetric, 2) a Braille-dot counting task in which subjects counted the number of dots in a series of random Braille letters, and 3) a sandpaper task requiring individuals to determine the relative roughness between a strip of sandpaper and the sandpaper disc surrounding it. All three tasks were conducted in an identical manner to our RP and sighted subject groups, including asking both subjects to use their dominant (right) hand and keep their eyes open while wearing a light excluding eye mask (made of black molded cell foam and nylon interlock 106 | P a g e fabric with a contoured rim) throughout the task. The tasks were completed in the same block design paradigm, in which individuals scanned a column during active blocks and rested their fingers in the empty space between columns during rest blocks. Both subjects also completed a visual task in which a flickering checkerboard was presented during scanning. The subject was instructed to gaze directly upwards towards a clip-on mirror that was placed on the head coil directly above the subject’s eyes. This mirror was oriented towards a rear-projection screen at the back of the scanner, allowing the subject to view the stimulus presentation. The screen size using the clip-on mirror was 23.6 ○ x 18.4 ○ with a viewing distance of 85 cm. Two alternating black and white checkerboards (interleaved with a blank black screen) were flickered at a frequency of 0.10 Hz (Figure 40). Figure 40. Example of the checkerboard stimulus. Two alternating checkerboards were presented during active blocks, followed by a blank black screen during resting blocks. The stimulus was presented in a block design paradigm, in which the checkerboard flickered continuously during active blocks and a black screen was presented during rest blocks. Each run was composed of four 20s active blocks and five 20s rest blocks. These blocks were interleaved, with the run starting and ending on a rest block. Two runs were completed for the checkerboard task, each lasting a total of 180s. This stimulus allowed us to determine the responsiveness of each subject’s visual cortex to visual input. 7.2.4. Experimental Procedure Both subjects completed two sets of MRI scans: one following a period of consistent device use and the other following a period of no device use. Subject A1 was enrolled in the study prior Active Block Rest Block … 107 | P a g e to receiving the Argus II device and so was also able to complete a baseline MRI session the day before her Argus II surgery. This preoperative session consisted of the same set of tactile, resting- state, and DTI scans completed in our RP sighted subjects groups (see sections 3.2.3, 5.2.2, and 6.2.2). In addition, the subject completed the above described checkerboard stimulus task. Both subjects completed two postoperative sessions after receiving the Argus II device. The device currently has FDA approved labeling allowing for its conditional use in a 3T MRI, which specifies that under the conditions on the label, a person with an Argus II can enter an MRI scanner room. As an added safety measure, a procedure was used that gradually added complexity in order to ensure the subject’s device and implanted eye were unaffected by each set of scans (all scanning parameters complied with those specified in the FDA’s approved MRI labeling for the device). Before entering the MRI machine, an ophthalmologist completed a baseline ophthalmic exam during which he examined the externally observable parts of the implanted eye and used an indirect ophthalmoscope to view the inside of the eye after pupil dilation. A baseline device test was then conducted according to standard device assessment protocols designed by the Argus II manufacturer. After verifying that the subject’s implanted eye was normal and that her Argus II device was functioning, the subject entered the MRI scanner for 5 minutes without her external prosthesis coil—no scanning was completed at this time. After exiting the scanner room, a second ophthalmic exam and device functionality test were performed. A brief interview was also conducted to determine if the subject felt any unusual sensations around her implanted eye while in the scanner room or scanner bore, including pressure, pain, phosphenes (spots of light), or heat. The subject then re-entered the scanner without her external prosthesis coil and completed a series of anatomical scans, functional scans (including the tactile tasks, visual stimulus, and resting-state scans), and DTI scans. The same indirect ophthalmoscopic exam, device functionality test, and subject interview were repeated following the anatomical scans, tactile tasks, and after the scanning session was complete (Table 11). 108 | P a g e MRI Session following Device Use MRI Session following NO Device Use BEFORE THE SESSION Baseline device functionality test Indirect ophthalmic exam PART 1 Subject enters scanner without glasses for 5 minutes (no scanning) Indirect ophthalmic exam Device functionality test Subject interview PART 2 Subject re-enters scanner without glasses Anatomical scanning Indirect ophthalmic exam Device functionality test Subject interview PART 3 Subject re-enters scanner without glasses Series of tactile task functional scans Indirect ophthalmic exam Device functionality test Subject interview PART 4 Subject re-enters scanner without glasses Anatomical scanning Resting-state scan Checkerboard functional scans DTI scans Indirect ophthalmic exam Device functionality test Subject interview BEFORE THE SESSION Indirect ophthalmic exam PART 1 Subject enters scanner without glasses Anatomical scanning Series of tactile task functional scans Indirect ophthalmic exam Device functionality test Subject interview PART 2 Subject re-enters scanner without glasses Anatomical scanning Resting-state scan Checkerboard functional scans DTI scans Indirect ophthalmic exam Device functionality test Subject interview Table 11. Summary of Postoperative procedure. 109 | P a g e For 5 days prior to the first postoperative session, subject A2 was asked to use her device for the maximum number of hours recommended by her ophthalmologists (5-6 hours/day). The second postoperative session was conducted 4 days after the first and was identical to the first postoperative session, excluding the initial 5 minutes of no scanning in the MRI machine. Subject A2 was asked to not use her device during the days preceding the second session. Since the device functionality test includes briefly stimulating the subject’s retina, this test was only conducted at the end of the session to prevent confounding the results of this “Device OFF” condition. Subject A1 was asked to not use her device for 7 days prior to the first postoperative session to simulate a “Device OFF” condition. She was then asked to use her device 5-6 hours/day for the 4 days preceding her 2 nd postoperative session—compliance was verified by checking the subject’s VPU activity log. 7.2.5. fMRI Data Analysis Image data was analyzed using BrainVoyager QX (Goebel et al., 2006) in subjects’ native space (as opposed to normalizing to a standard space). Anatomical data (specifically the high resolution MPRAGE images with TR/TE/flip angle/slice thickness = 1.95s/4.44ms/12°/1.0mm obtained during the first scanning session) underwent inhomogeneity correction and were reoriented via rigid-body rotation and translation to place the origin at the Anterior Commissure and the Posterior Commissure on the y-axis. Since both Argus II patients underwent multiple scanning sessions, anatomical scans from each were co-registered to each subject’s first session high resolution MPRAGE. All functional data was preprocessed with 3D motion correction (PACE and post hoc), slice timing correction, and temporal filtering. In cases of excessive head movement, which occurred in subject A2, volumes in which a subject exhibited movement greater than 0.6 mm/degree of motion (based on online PACE estimation) and the corresponding entries in the design matrix were excluded from the analysis. Spatial smoothing was not applied to the functional data. Whole-brain voxel-wise BOLD modulation was obtained by estimating the signal level during the active blocks with respect to that during the resting blocks using a GLM, with head- motion parameters as covariates. For each subject, individual functional data sets of each run were concatenated after normalization (z-transform). Significant voxel-wise activations were identified at FDR < 0.05 with a cluster threshold of 25mm 2 . The activation maps displayed below for each 110 | P a g e subject were constructed by projecting the GLM contrast (t-statistics) obtained from voxels on the cortex onto the reconstructed and inflated cortical surface meshes of the subject. Putative V1 and S1 were identified anatomically for each subject, as described in section 3.2.5. We calculated, for each subject, two complimentary measures (extent and strength) of the unsigned cross-modal response. The areal extent of cross-modal activation in the primary visual cortex was defined as the percentage of significantly modulated voxels on the cortex within the V1 ROI, while the strength of the response was calculated as the mean absolute parameter estimate (beta value) of the responding voxels within the V1 ROI. The percentage of modulated voxels and mean absolute beta value of those voxels were similarly calculated within the S1 ROI. A one-tailed student t-test was used to compare the strength and extent of tactile-evoked responses across tactile tasks following a period of consistent device use versus a period of no device use (i.e. vision deprivation). The t-test was further used to compare Argus II responses to the responses of 18 RP and sighted subjects described in Chapter 3. Statistically significant relationships were identified using a p-threshold of α = 0.05. 7.3 Results 7.3.1. Comparison of Visually and Tactile-Evoked Responses with and without Device Use Neither subject A1 nor A2 exhibited any responses to the checkerboard task—V1 remained unresponsive to visual stimulation following both device ON and OFF periods. For both subjects A1 and A2, a t-test revealed no significant difference between the extent or strength of responses in V1 to the three tactile tasks following a period of consistent device use versus 7 days of no device use for subject A1 (p’s < 0.10, Figure 41) and versus 4 days of no device use for subject A2 (p’s < 0.10, Figure 42). However, a significant change was found in S1 responses for both subjects. In subject A1, a difference was found in the strength (p = 0.004) and extent of the response in S1 (p = 0.011), where the response in S1 increased following 7 days of not using the device (Figure 43). For subject A2, the extent of the response was found to be greater following 4 days of no device use (p = 0.019, Figure 44). 111 | P a g e Figure 41. V1 BOLD responses to the three tactile tasks for subject A1 following 4 days of consistent device use and 7 days of no device use. Top: Significant responses (FDR < 0.05) were color-coded, with warm colors denoting increases in BOLD responses relative to rest. For each subject, the response patterns were projected onto an inflated representation of the occipital lobe; the outer white line represents the assumed V1/V2 boundary. Bottom: The extent of tactile-evoked BOLD responses in V1, measured in terms of the percentage of modulated voxels (FDR < 0.05), and the strength of tactile-evoked BOLD responses in V1, measured in terms of mean absolute beta value of the significantly modulated V1 voxels, for each task. 112 | P a g e Figure 42. V1 BOLD responses to the three tactile tasks for subject A2 following 7 days of consistent device use and 4 days of no device use. Top: Significant responses (FDR < 0.05) were color-coded, with warm colors denoting increases in BOLD responses relative to rest. For each subject, the response patterns were projected onto an inflated representation of the occipital lobe; the outer white line represents the assumed V1/V2 boundary. Bottom: The extent of tactile-evoked BOLD responses in V1, measured in terms of the percentage of modulated voxels (FDR < 0.05), and the strength of tactile-evoked BOLD responses in V1, measured in terms of mean absolute beta value of the significantly modulated V1 voxels, for each task. Device Use NO Device Use Device Use NO Device Use 113 | P a g e Figure 43. S1 BOLD responses to the three tactile tasks for subject A1 following 4 days of consistent device use and 7 days of no device use. Top: Significant responses (FDR < 0.05) were color-coded, with warm colors denoting increases in BOLD responses relative to rest. For each subject, the response patterns were projected onto an inflated representation of the somatosensory cortex; the outer white line represents the S1 boundary. Bottom: The extent of tactile-evoked BOLD responses in S1, measured in terms of the percentage of modulated voxels (FDR < 0.05), and the strength of tactile-evoked BOLD responses in S1, measured in terms of mean absolute beta value of the significantly modulated S1 voxels, for each task. * p < 0.05 114 | P a g e Figure 44. S1 BOLD responses to the three tactile tasks for subject A2 following 4 days of consistent device use and 4 days of no device use. Top: Significant responses (FDR < 0.05) were color-coded, with warm colors denoting increases in BOLD responses relative to rest. For each subject, the response patterns were projected onto an inflated representation of the somatosensory cortex; the outer white line represents the S1 boundary. Bottom: The extent of tactile-evoked BOLD responses in S1, measured in terms of the percentage of modulated voxels (FDR < 0.05), and the strength of tactile-evoked BOLD responses in S1, measured in terms of mean absolute beta value of the significantly modulated S1 voxels, for each task. * p < 0.05 115 | P a g e 7.3.2. Comparison of Argus II Subject Responses to RP and Sighted Groups We compared subject A1 and A2’s BOLD responses to our previous 18 RP and sighted subjects to identify any differences in visual cortex responses with prolonged use of the device. Based on A1 and A2’s visual acuity and visual field, both subjects belong to the “Blind” category (as described in section 3.2.1). Subject A1 completed our study 5 weeks after successful implantation of the Argus II device: both the extent and strength of her V1 responses to the tactile tasks were found to be statistically the same as the blind RP group (p’s > 0.05) and statistically different from the low vision RP group and sighted group (p’s < 0.01, Figure 45). A t-test also found that the extent of subject A1’s S1 responses to the task (following a period of consistent device use) were not statistically similar to any of the three (blind, low vision, sighted) groups (p’s < 0.001, data not shown), while the extent of the S1 response was statistically the same as all groups (p’s > 0.05). It should be noted that no effect of vision loss was found on S1 responses when vision loss was defined categorically as blind, low vision, or sighted—all groups exhibited a similar extent and strength of S1 response to tactile stimulation (see section 3.3.3). The extent of subject A1’s S1 responses thus fell within the general variability seen across the three groups, while the strength of her responses fell outside of this range. Subject A2 completed our study 15 weeks after successful implantation of the device. However, her responses were not similar to the blind group: the extent of A2’s tactile-evoked responses was found to be statistically similar to the sighted group (p = 0.202) and statistically different from both blind (p < 0.001) and low vision (p = 0.029) groups (Figure 46). Similarly, the strength of A2’s responses were statistically similar to the low vision group (p = 0.093) and different from the blind (p = 0.031) and sighted (p = 0.015) groups. A t-test also found that the extent of subject A2’s S1 responses to the task (following a period of consistent device use) were statistically the same as the three (blind, low vision, sighted) groups (p’s > 0.05, data not shown). The strength of the S1 response was only similar to the blind group (p = 0.118) and statistically different from the low vision and sighted groups (p’s < 0.05). 116 | P a g e Figure 45. Comparison of extent and strength of tactile-evoked responses in V1 for Argus II subject A1 to RP and sighted groups. A: The extent of tactile-evoked BOLD responses in V1, measured in terms of the percentage of modulated voxels (FDR < 0.05) in V1 for each subject and each task. RP subjects are ranked along the x-axis in descending order of severity of visual field loss. Subject A2 is grouped against a tan background. B: The strength of tactile-evoked BOLD responses in V1, measured in terms of mean absolute beta value of the significantly modulated V1 voxels for each subject and each task. “A1_2” indicates subject A2’s responses following 7 days of no device use, while “A1_3” indicates responses following 4 days of consistent device use (5-6 hours/day). C: Boxplot illustrating the distributions of the percentage and mean absolute beta value of activated S1 voxels in RP and sighted control groups across all tasks for comparison. The red line indicates the median within each group, the edges of the boxes indicate the 25th and 75th percentiles, and the whiskers illustrate the extreme data points, excluding outliers (red data points).* p < 0.05, ** p < 0.001 A B 117 | P a g e Figure 46. Comparison of extent and strength of tactile-evoked responses in V1 for Argus II subject A2 to RP and sighted groups. A: The extent of tactile-evoked BOLD responses in V1, measured in terms of the percentage of modulated voxels (FDR < 0.05) in V1 for each subject and each task. RP subjects are ranked along the x-axis in descending order of severity of visual field loss. Subject A2 is grouped against a tan background. B: The strength of tactile-evoked BOLD responses in V1, measured in terms of mean absolute beta value of the significantly modulated V1 voxels for each subject and each task. “A2_1” indicates subject A2’s responses following a period of consistent device use (5-6 hours/day), while “A2_2” indicates responses following 4 days of no device use. C: Boxplot illustrating the distributions of the percentage and mean absolute beta value of activated S1 voxels in RP and sighted control groups across all tasks for comparison. The red line indicates the median within each group, the edges of the boxes indicate the 25th and 75th percentiles, and the whiskers illustrate the extreme data points, excluding outliers (red data points). * p < 0.05, ** p < 0.001 7.4 Discussion One of the primary concerns in completing MRI scans of Argus II patients was the amount of image distortion that would result from the presence of the implant. For both MPRAGE and 118 | P a g e EPI scans, our data revealed the presence of an artifact that was only localized around the patient’s implanted eye—this artifact did not extend to our other regions of interest (including the visual cortex, somatosensory cortex, temporal, and parietal regions). This demonstrates that successful data acquisition is possible in retinal implant patients. Results from two Argus II subjects indicate that extended use of a retinal prosthesis (and thus partial vision restoration) may result in a reduction of tactile-evoked responses in V1. Since this decrease was not evident in subject A1 after 5 weeks of using the device, it is possible that a longer period of vision restoration (e.g. up to 4 months as with subject A2) is required before these changes in cross-modal responses occur. Alternatively, repeating this study in a larger group of Argus II patients may reveal that use of the device does not consistently affect tactile-evoked responses in V1. If this is the case, it will be even more essential to determine the effect of cross- modal activity on a patient’s ability to adapt to the device—if the activity does not decrease reliably with device use, the strength and extent of a patient’s tactile-evoked responses before surgery can inform us of how they may respond to the treatment. However, if a larger cohort of Argus II patients shows that these responses do decrease with extended use of the device, this suggests that cross-modal responses in V1 are indeed reversible once vision is recovered—both the strength and extent of subject A2’s responses were reduced compared to other blind RP subjects with a similar visual acuity and visual field. In the case that cross-modal responses are determined to have a negative impact on an individual’s ability to adapt to a retinal prosthesis, our results indicate that these responses may only be temporary. As an individual uses their device over time, these responses will decrease and a patient’s progress with the device may improve. If tactile-evoked responses in V1 do not decrease with prolonged device use, this may indicate that the signal from the implant is not reaching V1, which could be a reason for a subject’s subpar visual performance. This reversal of cross-modal responses within a period of 4 months supports theories that cross-modal responses are driven by changes in local synaptic connections and cortio-cortical feedback, both of which are highly plastic processes in both young and adult humans. A lack of difference in responses following only 6 days of vision deprivation may also be consistent with a study by Merabet et al., 2008 describing the appearance of tactile-evoked V1 responses in sighted subjects after 5 days—a similar increase in tactile-evoked responses may have occurred in our Argus II patients following 6 days of not using the device. However, since the patients already had 119 | P a g e significant cross-modal responses in V1, a measurable increase may only be evident following an even longer period of vision deprivation (i.e. no device use). In addition, an increase in S1 responses was exhibited following 4-6 days of vision deprivation (i.e. no device use). Given that our previous results revealed a negative correlation between vision loss and S1 responses (where the extent and strength of S1 activity decreased with worsening vision), this contradictory result suggests that the subjects’ tactile sensitivity has decreased with consistent use of their Argus II. As a result, their tactile strategy may have changed to adapt to their new level of sensitivity (e.g. pressing harder on the tactile elements), thus causing an increase in S1 responses. If cross-modal responses can be correlated with visual performance in retinal prosthesis recipients, tactile-evoked activities may be used as a predictive measure of patient outcomes following treatment. For example, an increase in performance as cross-modal responses decrease over time with device use can inform rehabilitation strategies: using the device consistently for the first months following surgery may result in a faster decrease in the responses and thus help a patient more quickly adapt to the prosthetic vision. The theory of whether or not cross-modal plasticity will help or hinder retinal prosthesis use can be informed by results in cochlear implant patients. A CI study completed in pre-lingual deaf patients found that individuals with significant cross-modal plasticity in the auditory cortex (which responded to seeing sign language) showed no improvement following successful implantation of a cochlear implant (Lee et al., 2001). This study concluded that these results may be similar in blind patients whose visual cortex responds to reading Braille. However, this study was conducted in early deaf patients—results may differ in late-blind patients if their cross-modal plasticity has come about from mechanisms that are highly plastic even in adulthood (such as changes in synaptic weights or cortical-cortico feedback processes). Cross-modal responses in the congenitally or early-blind may have been the result of mechanisms that become stable with age, and therefore are not easily reversible once a prosthesis has been implanted. Along these lines, a separate study found that late (post-lingual) deaf individuals with a greater extent of audiovisual association were more successful at rehabilitation with a CI than individuals whose visual cortex was less responsive to speech sounds (Giraud et al., 2001). They further concluded that higher performers utilized this enhanced visual cortex activity to compensate for a CI’s imperfect auditory signal. We can expect these latter findings to be translatable to late-blind patients who receive a retinal prosthesis. 120 | P a g e Chapter 8: Development and validation of a protocol for suprathreshold stimulation of the Argus II device in a 3T MRI Machine 8.1 Overview The use of fMRI is crucial in understanding how the visual cortex of Argus II patients responds to suprathreshold electrode stimulation. In order to safely and successfully complete MRI experiments in this population group, it is necessary to ensure that the Argus II device is functional within the static and dynamic magnetic, and RF fields associated with MRI. Safety of an inactive Argus II implant in the scanner has previously been demonstrated (Weiland et al., 2012). However, in order to activate and stimulate the implant during scanning, data and power must be sent from the VPU to the implant without creating excessive artifacts and noise in the resulting MRI images. A stable RF link between the VPU and implant would allow the prosthesis to operate in the MRI machine and enable us to electrically stimulate the device during scanning. The following objectives were investigated:  Phantom testing was used to determine how presence of the Argus II affects the quality of structural and functional MR data 121 | P a g e  We developed a scanning protocol that allows us to record brain activity elicited by the Argus II device  This scanning protocol and associated hardware was validated in an Argus II canine model  Phantom tests were conducted in most our modified Argus II external coil to confirm functional integrity during scanning  The device (including sparse scanning hardware) was re-packaged for easy use in the scanner and with patients 8.2 Study Design 8.2.1. Preliminary Phantom and Canine Experiments Phantom experiments were completed using a simulated implant composed of an internal and external coil and circuit board that served to replicate the internal processing of an actual Argus II device (Figure 47, Left). The device was placed next to a 20 pound water bottle during scanning, which served as the signal bearing substance for our phantom experiments. MR images for all phantom tests were acquired in a 3.0T Siemens MAGNETOM TIM Trio scanner using a 12- channel Matrix head array coil. Anatomical images were obtained using a T1-weighted sequence (MPRAGE, TR = 1950ms), while functional images were acquired using an EPI sequence (TR = 1000ms). 36 slices with isotropic voxels of 3x3x3mm 3 were axially oriented and covered the majority of the water bottle, including the region adjacent to the simulated implant. During an initial phantom experiment, the simulated implant was placed on the scanner table. The implant was connected to a VPU via a BNC cable that ran through a waveguide filter to prevent noise from being transmitted through the cables from the computer area to the scanner room (Figure 47, Right). The static magnetic field generated by an MRI scanner contains inhomogeneities depending on the location inside and outside the scanner bore; such variations in uniformity may induce currents within the cables that interfere with signaling between the VPU and implant. Accordingly, the simulated implant was moved along the scanner table and into the scanner bore to determine locations within the static field where a link could be established. 122 | P a g e Figure 47. Left: Simulated Implant with A. internal receiving coil (red copper coil); B. external transmitter coil and electronics (in black case); C. circuit board. Right: Layout of the MRI scanner suite, showing placement of the VPU outside of the scanner room and implant within the scanner bore. Once a suitable position was found in the scanner bore for establishing a link, the implant was secured to the side of a phantom water bottle (simulating placement next to a patient’s left eye) and an RF link was established. Anatomical and functional scans were completed to determine how well the link was maintained within a rapidly changing gradient magnetic field. The amount of noise and artifact resulting from the simulated implant was determined by comparing images acquired 1) without the implant present, 2) with the implant present but inactive, and 3) with the implant present and active (i.e. RF link on). To determine the extent of MR signal loss due to interference with the prosthesis, a canine implanted with an Argus II device was scanned in a 3.0T GE Signa HDxt scanner using 3D T1- weighted anatomical and EPI functional sequences. The position of the coils (and the canine’s head) was varied within the scanner to determine the locations in the static field where the active prosthesis could maintain an RF link with the VPU (similar to the above-described phantom experiments). The implant remained inactive throughout scanning to prevent noise from masking the amount of artifact in the resulting images. All procedures were approved by the USC Institutional Animal Care and Use Committee (IACUC). 8.2.2. Phantom Experiments with a Sparse Scanning Protocol Preliminary results revealed that an RF link between the internal and external coils could not be maintained during image acquisition periods of scanning. Data acquisition generally occurs simultaneously with changes in frequency and phase gradients and could interfere with the transfer of data and power to the implant, while the RF transmissions of the implant could lead to strong 123 | P a g e RF noise in the MR images. Consequently, a sparse temporal scanning protocol was developed (using an EPI sequence with TR/TE/flip angle = 3s/30ms/90˚) which takes advantage of the relatively slow rate of the hemodynamic response and allows the implant and VPU to communicate in-between periods of image acquisition (Figure 48). This would allow for eventual stimulation of the device during scanning. To allow for transfer of data and power between the two coils, the RF link was turned on prior to image acquisition and off during image acquisition. The retinal stimulations presented to an Argus II patient will initiate a hemodynamic response with a typical time-to-peak of 4-6 seconds; this is slow enough to be recorded by the image acquisition that immediately follows. The sparse scanning protocol was implemented using a Velleman K8055 USB experimental interface board which allowed manipulation of the RF link to be synchronized with the MRI scanner TR. Sparse scanning is an established protocol for studying the auditory system in order to avoid the loud acoustic gradient noise in the MR environment. We have validated this protocol in the visual system of a normally sighted subject using light stimulation. Figure 48. Sparse temporal scanning protocol sequence of events; the RF link is turned on and given time to re-establish during stimulation periods and quickly turned off before acquisition at each TR (this process is repeated for every TR). Phantom experiments were conducted using a modified air-core external coil developed by SSMP that is free of ferromagnetic material. Further modifications were made to the device’s telemetry system and VPU to support proper function and testing within the static field. Phantom experiments were completed in which 1) a gold internal coil and the modified external coil were positioned next to a 20 pound water bottle within the scanner bore, 2) the internal coil was placed in a vial of phosphate buffered saline (PBS), and 3) when the internal coil was wrapped in a package of ground pork to simulate the effect of tissue when the coil is implanted in a human. As a control, the degree of noise in the resulting MR images was compared to functional and anatomical images created with the VPU and RF link turned off. 124 | P a g e 8.2.3. Validation of Sparse Scanning Protocol in a Canine Model The sparse scanning protocol was validated using the modified air-core external coil and a canine implanted with the Argus II device in a 3.0T GE Signa HDxt MRI scanner. In a manner similar to the phantom experiments, the canine was placed at different positions within the scanner to determine the locations in the static field where an RF link could be maintained between the two coils. Once an RF link was established within the bore, 3D T1-weighted anatomical and sparse scanning functional (EPI) sequences were completed. Direct light stimulation was attempted in the canine Argus II subject using the sparse scanning protocol. An optical fiber stimulus was created using a fiber optic cable that was connected to a strong light source on one end and to an eye patch placed over the canine’s non-implanted eye on the other end; this served as a control experiment to verify that direct light stimulation resulted in visual cortex activation. In addition, this allowed us to confirm that artifact from the implant did not obstruct images of the visual cortex. 8.3 Results 8.3.1. Preliminary Phantom and Canine Experiments In both phantom and canine experiments, an RF link between the implant and VPU could only be established at or near the iso-center of the scanner when the coils were oriented at a specific angle to the main static field line; this angle depended on the position of the implant within the scanner bore and may have been due to the static magnetic field inducing currents within the cables that interfered with the backward telemetry between the VPU and implant. Phantom experiments revealed that the presence of an inactive implant induced significant artifact in the resulting MR images when compared to images of the water bottle alone (Figure 49A-B). Scanning with the implant active caused the RF link to fail and interfered with the MR signal, resulting in a noisy image without any intelligible features Figure 49C). 125 | P a g e Figure 49. Phantom experiment with simulated implant resting next to a water bottle. A: Anatomical scan of a water bottle without the simulated implant (TR = 1950 ms); B: Localizer scan of a water bottle placed next to the inactive simulated implant (TR = 20 ms); C: Anatomical scan of a water bottle next to the active simulated implant (TR = 1950 ms). Yellow outline indicates the location of the water bottle. Functional and anatomical scanning in canines similarly demonstrated that the inactive implant resulted in a small degree of artifact in the MR images (Figure 50A-B), while contact of the inactive device with the canine’s retina did not activate the visual cortex (Figure 50C). Figure 50. Canine experiment results. A: Anatomical image of the canine’s head with crosshairs positioned over the location of the inactive retinal prosthesis (green circle indicates a 2.5cm diameter artifact); B: Corresponding functional image (green circle indicates a 4.0cm diameter artifact); C: Functional (T2* weighted gradient EPI) image showing no cortical activity due to retinal stimulation from the device contacting the retina (crosshairs are positioned over the primary visual cortex). 8.3.2. Phantom Experiments with Sparse Scanning Protocol An RF link was successfully established within the scanner’s static magnetic field when 1) the gold internal coil and the modified external coil were positioned next to a 20 pound water bottle, 2) the internal coil was placed in a vial of PBS and 3) when the internal coil was wrapped in a package of ground pork to simulate the effect tissue when the coil is implanted in a human. In all cases, the RF communication link was quickly re-established following periods of image acquisition during the functional scans when sparse scanning was used. A noise pattern present when the implant remained active throughout scanning was no longer present with implementation of the sparse scanning protocol (Figure 51). 126 | P a g e Figure 51. Results of phantom experiments with sparse scanning protocol. Left: 18 image slices of a phantom with the modified external coil and internal gold coil inactive throughout a normal EPI (TR = 3s) without sparse scanning. Center: Same phantom with implant active throughout a normal EPI (TR = 3s) without sparse scanning. A noise pattern is present from simultaneous image acquisition and implant linking. Right: Images of the same phantom with RF link turned ON and OFF by a USB experimental interace board during sparse scanning. Noise is no longer present and images look similar to those acquired when the implant was inactive. 8.3.3. Validation of Sparse Scanning Protocol in a Canine Model Functional (EPI) scans completed in a canine demonstrated successful implementation of the sparse scanning protocol—the RF link was turned off prior to image acquisition and quickly re-established following periods of image acquisition. However, the stability of the link was highly dependent upon the position of the eye (and internal coil) relative to the modified external coil, both inside the scanner bore and outside the scanner room. Direct light stimulation to the non-implanted eye resulted in a BOLD response that could be measured in an area near the reported location of canine visual cortex (Aguirre et al., 2007). While this data is preliminary, it suggests that functional responses can be recorded during sparse scanning in the presence of the modified external system (Figure 52). 127 | P a g e Figure 52. Top: Light-evoked fMRI BOLD signal acquired from a canine with the retinal implant. Z- values of significantly activated voxels (cluster p threshold < 0.05) are overlaid on the EPI images. Bottom: The time course of a voxel in the visual cortex (blue arrow) is shown against model prediction. This result demonstrates that we can observe evoked brain activity in the presence of the implant. 8.3.4. Testing RF link between our Modified External Coil and an Argus II Patient The modified external coil was tested on Argus II subject A1 to determine if we could establish an RF link with her implant. Holding the coil in the typical location of her own coil, an intermittent link was achieved outside of the MRI scanner. We concluded that instability of the link was due to an insufficient range between the patient’s implant and our coil. This may be corrected by increasing the gain on our modified coil’s preamp. 128 | P a g e PART IV: CONCLUSION 129 | P a g e Chapter 9: Context of results and overall conclusions 9.1 Summary of Work and Contribution to the Field Consistent with previous studies, our research found that cross-modal responses in V1 become more extensive as vision loss progresses in late-blind individuals with RP. In addition, our work found a consistent relationship between tactile-evoked activity in V1 and a patient’s visual function: as visual field and visual acuity diminishes, V1 becomes more responsive to tactile- stimulation. These results are evidence that the visual cortex is highly plastic even into adulthood—such plasticity is also evident as a patient’s vision is partially restored. A case study of two Argus II patients revealed decreased cross-modal responses with more than 3 months use of a retinal prosthesis. While several studies in congenitally and early-deaf individuals have demonstrated that audio-visual cross-modal responses in the auditory cortex hinder a patient’s ability to adapt to a cochlear implant, our results suggest that this is not the case in individuals who lose a sense as adults. In fact, cochlear implant studies in post-lingual deaf patients suggest that cross-modal responses are related to improved auditory responses after receiving the device. In order to generate cross-modal processes that are easily reversible in adults, the mechanisms involved must be highly plastic throughout a person’s lifetime. These may include changes to local synaptic weights and cortical-cortico feedback processes, both of which are continually altered throughout adulthood. Since tactile-evoked responses may be reversible, it is unlikely that cross- 130 | P a g e modal activity will hinder an individual’s ability to adapt to a retinal prosthesis if the patient has lost their sight later in life. Instead, these responses may reveal a sensory association network that is easily adaptable to changes in sensory stimulation and can thus help compensate for an imperfect prosthesis signal. While structural connectivity of regions along the visual streams remain unaffected by late- blindness, the inferior parietal lobe was found to more strongly mediate the functional connection between V1 and S1 among RP patients when compared to their sighted counterparts. These findings are consistent with studies that suggest the parietal areas act as a hub in a pathway between the two regions that is functionally modified with blindness, and may offer a glimpse into a mechanism that drives tactile-evoked responses in V1. Through this work, we have come one-step closer to identifying a biomarker that can be used to help predict patient outcomes following a sight restoration treatment. If cross-modal responses in V1 are found to correlate with visual performance results after receiving a retinal prosthesis, tactile-evoked activity may be used as a predictive measure prior to implantation. For instance, if individuals with more extensive cross-modal responses before receiving the device demonstrate greater visual function after successful implantation and prolonged device use, cross- modal responses in RP patients may be incorporated into the initial screening process to determine their likelihood of benefiting from the device. This V1 biomarker may also be used to determine if the signal from the implant is being sufficiently processed by the visual cortex. Since low-level vision restoration has been shown to decrease cross-modal responses in V1, a lack of change in the extent and strength of tactile-evoked activity in V1 after using the device suggests that the signal from the implant is not reaching that region of the brain. In some cases, this may help account for a patient’s subpar visual performance with the device. Similar cross-modal biomarkers may be identified for higher visual areas to determine how the signal from the implant is impacting other regions along the visual stream. 9.2 Suggested Future Work This thesis work has produced as many new and exciting questions as it has results. It is our hope that the findings from this project, in conjunction with the below future work, will lead to a reliable method of predicting outcomes following sight restoration treatment. 131 | P a g e Verify that cross-modal responses decrease with retinal prosthesis use in a larger cohort of Argus II users. This can include measuring the extent and strength of tactile-evoked responses in V1 (and other visual areas) repeatedly over time to identify how the responses change with prolonged used of the device. Use of the device and tactile-evoked responses may also be correlated with changes in the functional relationship among V1, the inferior parietal lobe, and S1, a possible source of cross-modal responses. Correlate Argus II visual performance with the extent and strength of a patient’s tactile- evoked V1 responses over time. This will help us determine whether cross-modal responses in V1 have a positive impact, negative impact, or no effect on visual performance results. It can also reveal possible correlations between variability in visual performance over time and fluctuations in tactile-evoked responses. Correlate an individual’s tactile-evoked responses (in both V1 and association areas like the inferior parietal lobe) before a patient receives the device with their visual performance after receiving and using a retinal prosthesis. Doing so will provide a predictive measure of how pre- implant cross-modal responses are related to visual performance with the device. Measure visual cortex BOLD responses to suprathreshold stimulation with an Argus II device (while presenting a checkerboard pattern to the device’s camera) and compare the responses to those of a sighted individual passively viewing the same checkerboard pattern. This will inform us of how the visual cortex treats the signal from the implant and whether the responses follow a retinotopic pattern similar to that observed in sighted individuals. In order to address the question of whether cross-modal responses in V1 are functional, correlate tactile performance results with the extent and strength of tactile-evoked responses in V1 in sighted and late-blind RP patients with varying degrees of vision loss. Use DTI to analyze the optic radiation and determine: 1. how the integrity of the optic radiation changes with vision loss in late-blind RP patients and 2. if any loss in integrity is restored with prolonged use of an Argus II device. 132 | P a g e REFERENCES Age-Related Eye Disease Study Research Group (2001) A randomized, placebo-controlled, clinical trial of high-dose supplementation with vitamins C and E and beta carotene for age-related cataract and vision loss: AREDS report no. 9. Arch Ophthalmol, 126(9): 1251. Aguirre GK, Komaromy AM, Cideciyan AV, Brainard DH et al. (2007) Canine and human visual cortex intact and responsive despite early retinal blindness from RPE65 mutation. PLoS Medicine, 4(6): 1-12. Amedi A, Raz N, Pianka P, et al. (2003) Early ‘visual’ cortex activation correlates with superior verbal memory performance in the blind. Nature Neurosci, 6(7): 758-766. Amedi A, Raz N, Azulay H, Malach R, Zohary E (2010) Cortical activity during tactile exploration of objects in blind and sighted humans. Restor Neurol Neurosci, 28(2): 143-156. Bavelier D and Neville HJ (2002) Cross-modal plasticity: Where and how? Nature Reviews: Neuroscience, 3: 443-452. Beckmann CF, DeLuca M, Devlin JT, Smith SM (2005) Investigations into resting-state connectivity using independent component analysis. Phil. Trans. R. Soc. B, 360: 1001-1013. Bedny M, Pascual-Leone A, Dravida S, et al. (2011) A sensitive period for language in the visual cortex: Distinct patterns of plasticity. Brain and Language, 122: 162-170. Behrens TEJ, Johansen-Berg H, Woolrich MW, et al. (2003) Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nature Neuroscience, 6(7): 750-757. Behrens TEJ, Johansen-Berg H, Jbabdi S, et al. (2007) Probabilistic diffusion tractography with multiple fibre orientations. What can we gain? NeuroImage, 23:144-155. Benson NC, Buff OH, Datta R, et al. (2012) The retinotopic organization of striate cortex is well predicted by surface topology. Curr Biol, 22: 1-5. Brindley GS (1955) The site of electrical excitation of the human eye. J Physiol, 127: 189-200. 133 | P a g e Brindley GS (1964) A new interaction of light and electricity in stimulating the human retina. J Physiol, 171: 514-520. Brindley GS and Lewin WS (1968) The visual sensations produced by electrical stimulation of the medial occipital cortex. J. Physiol, 196: 479–93. Buchel C, Price C, Frackowiak RSJ, Friston K (1998) Different activation patterns in the visual cortex of late and congenitally blind subjects. Brain, 121: 409-419. Burton H, Snyder AZ, Conturo TE, Akbudak E, et al. (2002) Adaptive changes in early and late blind: a fMRI study of Braille reading. J Neurophysiol, 87(1): 589-607. Burton H (2003) Visual cortex activity in early and late blind people. J Neuro, 23(10): 4005– 4011. Carlson S, Pertovaara A, Tanila H (1987) Late effects of early binocular visual deprivation on the function of Brodmann's area 7 of monkeys (Macaca arctoides). Brain Res, 430(1): 101-11. Chai X, Li L, Wu K, Zhou C (2008) C-Sight Visual Prostheses for the Blind: Optic Nerve Stimulation with Penetrating Electrode Array. IEEE Engineering in Medicine and Biology Magazine, 20-28. Chen Z, Lin F, Wang J, et al. (2012) Diffusion tensor MRI reveals visual pathway damage that correlates with clinical severity in glaucoma. Clinical Science. Cheung S-H, Fang F, He S, Legge GE (2009) Retinotopically-specific reorganization of visual cortex for tactile pattern recognition. Curr Biol, 19: 596-601. Chou C-F, Cotch MF, Vitale S, et al. (2013) Age-related eye diseases and visual impairment among U.S. adults. Am J Prev Med, 45(1): 29–35. Chow AY, Bittner AK, Pardue MT (2010) The artificial silicon retina in retinitis pigmentosa patients (an American Ophthalmological Association thesis). Trans Am Ophthalmol Soc, 108: 120- 54. Cicione R, Shivdasani MN, Fallon JB, Luu CD, et al. (2012) Visual cortex responses to suprachoroidal electrical stimulation of the retina: effects of electrode return configuration. J Neural Eng, 9(3): 036009. Cohen LG, Weeks RA, Sadato N, Celnik P, et al. (1999) Period of susceptibility for cross-modal plasticity in the blind. Ann Neurol, 45(4): 451-60. 134 | P a g e Congdon N, O’Colmain B, Klaver CCW, Klein R (2011) Causes and prevalence of visual impairment among adults in the United States. Arch Ophthalmol, 122: 477-485. Cox DR and Snell E J (1971) On test statistics calculated from residuals. Biometrika, 58.3: 589- 594. Daniel PM and Whitteridge D (1961) The representation of the visual field on the cerebral cortex in monkeys. J Physiol, 159: 203-21. De Volder AG, Bol A, Blin J, Robert A, et al. (1997) Brain energy metabolism in early blind subjects: neural activity in the visual cortex. Brain Res, 750(1-2): 235-44. Dobelle WH and Mladejovsky MG (1974) Phosphenes produced by electrical stimulation of human occipital cortex, and their application to the development of a prosthesis for the blind. J. Physiol, 243: 553–76. Doucet ME, Guillemot JP, Lassonde M, Gagné JP, et al. (2005) Blind subjects process auditory spectral cues more efficiently than sighted individuals. Exp Brain Res, 160(2): 194-202. Facchini S and Aglioti SM (2009) Short term light deprivation increases tactile spatial acuity in humans. Neurology, 60: 1998-1999. Fieger A, Röder B, Teder-Sälejärvi W, Hillyard SA, et al. (2006) Auditory spatial tuning in late- onset blindness in humans. J Cogn Neurosci, 18(2): 149-57. Fine I (2007) The effects of visual deprivation: implications for sensory prosthesis. Artificial Sight, Springer 47-70. Fine I, Wade AR, Brewer AA, May MG, et al. (2003) Long-term deprivation affects visual perception and cortex. Nat Neurosci, 6(9): 915-6. Fornos AP, Sommerhalder J, Pelizzone M (2011) Reading with a simulated 60-channel implant. Front Neurosci, 5: 57. Friedman DS, O'Colmain BJ, Muñoz B, Tomany SC (2004) Prevalence of age-related macular degeneration in the United States. Arch Ophthalmol, 129(9): 1188. Fujii T, Tanabe HC, Kochiyama T, and Sadato N (2009) An investigation of cross-modal plasticity of effective connectivity in the blind by dynamic causal modeling of functional MRI data. J Neurosci Res. 65: 175-186. 135 | P a g e Ge Y1, Grossman RI, Babb JS, Rabin ML, Mannon LJ, Kolson DL (2002) Age-related total gray matter and white matter changes in normal adult brain. Part I: volumetric MR imaging analysis. AJNR Am J Neuroradiol, 23(8): 1327-33. Giraud AL, Price CJ, Graham JM, Truy E, et al. (2001) Cross-modal plasticity underpins language recovery after cochlear implantation. Neuron, 30(3): 657-63. Giriyappa D, Subrahmanyam RM, Rangashetty S, Sharma R (2009) Index finger somatosensory evoked potentials in blind Braille readers. Neurol Neurochir Pol, 43(5): 439-45. Gizewski ER, Gasser T, de Greiff A, Boehm A, et al. (2003) Cross-modal plasticity for sensory and motor activation patterns in blind subjects. Neuroimage, 19(3): 968-75. Goebel R, Esposito F, and Formisano E (2006) Analysis of functional image analysis contest (FIAC) data with BrainVoyager QX: from single-subject to cortically aligned group general linear model analysis and self-organizing group independent component analysis. Hum Brain Mapp. 27: 392-401. Goldreich D and Kanics IM (2003) Tactile Acuity is Enhanced in Blindness. J Neurosci, 23(8): 3439 –3445. Goldreich D and Kanics IM (2006) Performance of blind and sighted humans on a tactile grating detection task. Percept Psychophys, 68(8): 1363-71. Goyal MS, Hansen PJ, Blakemore CB (2006) Tactile perception recruits functionally related visual areas in the late-blind. Neuroreport, 17(13): 1381-4. Greicius MD1, Supekar K, Menon V, Dougherty RF (2008) Resting-state functional connectivity reflects structural connectivity in the default mode network. Cereb Cortex, 19(1): 72-8. Green KM, Ramsden RT, Julyan PJ, Hastings DL (2008) Neural plasticity in blind cochlear implant users. Cochlear Implants Int, 9(4): 177-85. Hamel C (2006) Retinitis pigmentosa. Orphanet Journal of Rare Diseases, 1(40). Hamilton R, Keenan JP, Catala M, Pascual-Leone A (2000) Alexia for Braille following bilateral occipital stroke in an early blind woman. Neuroreport, 11(2): 237-40. Hannula DE, Simons DJ, Cohen NJ (2005) Imaging implicit perception: promise and pitfalls. Nat Rev Neurosci, 6(3): 247-55. 136 | P a g e Henriksson L, Karvonen J, Salminen-Vaparanta N, et al. (2012) Retinotopic maps, spatial tuning, and locations of human visual areas in surface coordinates characterized with multifocal and blocked fMRI designs. PLoS One, 7(5): e36859. Hinds O, Polimeni JR, Rajendran N, et al. (2009) Locating the functional and anatomical boundaries of human primary visual cortex. NeuroImage, 46: 915-922. Humayun MS, de Juan E Jr, Dagnelie G, et al. (1996) Visual perception elicited by electrical stimulation of retina in blind humans. Arch Ophthalmol, 114: 40-46. Humayun MS, Dorn JD, da Cruz L, Dagnelie G (2012) Interim results from the international trial of second sight’s visual prosthesis. Ophthalmology. Hyvärinen J, Carlson S, Hyvärinen L (1981) Early visual deprivation alters modality of neuronal responses in area 19 of monkey cortex. Neurosci Lett, 26(3): 239-43. Jenkinson M, Beckmann CF, Behrens TE, et al. (2012) FSL. NeuroImage, 62: 782-90. Johansen-Berg H, TEJ Behrens, Robson MD, et al. (2004) Changes in connectivity profiles define functionally distinct regions in human medial frontal cortex. Proc Natl Acad Sci, 101(36): 13335- 13340. Kauffman T, Theoret H, Pascual-Leone A (2002) Braille character discrimination in blindfolded human subjects. NeuroReport, 13:571–574. Khandhadia S, Cherry J, Lotery AJ (2012) Age-related macular degeneration. Neurodegenerative Diseases, Landes Bioscience and Springer Science+Business Media: 15-36. Kim M, Ducros M, Carlson T, Ronen I, et al. (2006) Anatomical correlates of the functional organization in the human occipitotemporal cortex. Magn Reson Imaging, 24(5): 583-90. Lauritzen TZ, Dorn JD, McClure K, Greenberg R, et al. (2011) Spatio-temporal effects of inter- electrode discriminability performance in an Argus TM II subject. ARVO poster submission. Lee DS, Lee JS, Oh SH, et al. (2001) Cross-modal plasticity and cochlear implants. Nature, 409: 149-150. Leo A, Bernardi G, Handjaras G, et al. (2012) Increased BOLD variability in the parietal cortex and enhanced parieto-occipital connectivity during tactile perception in congenitally blind individuals. Neural Plast. 137 | P a g e Leporé N, Voss P, Lepore F, Chou YY, et al. (2010) Brain structure changes visualized in early- and late-onset blind subjects. Neuroimage, 49(1): 134-40. Lessard N, Paré M, Lepore F, Lassonde M (1998) Early-blind human subjects localize sound sources better than sighted subjects. Nature, 395(6699): 278-80. Li J, Liu Y, Qin W, Jiang J et al. (2012) Age of Onset of Blindness Affects Brain Anatomical Networks Constructed Using Diffusion Tensor Tractography. Cerebral Cortex. Liu Y, Yu C, Liang M, Li J, et al. (2007) Whole brain functional connectivity in the early blind. Brain, 130: 2085-2096. Marx E, Deutschlander A, Stephan T, et al. (2004) Eyes open and eyes closed at rest conditions: impact on brain activation patterns. NeuroImage, 21: 1818-1824. Masuda Y, Horiguchi H, Dumoulin SO, Furuta A, et al. (2010) Task-dependent V1 responses in human retinitis pigmentosa. Invest Ophthalmol Vis Sci, 51(10): 5356-64. Merabet LB, Swisher JD, McMains SA, et al. (2006) Combined activation and deactivation of visual cortex during tactile sensory processing. J Neurophysiol, 97: 1633-1641. Merabet LB, Hamilton R, Schlaug G, Swisher JD, et al, (2008) Rapid and reversible recruitment of early visual cortex for touch. PLoS ONE, 8:1-12. Mohammadi HM, Ghafar-Zadeh E, Sawan M (2012) An image processing approach for blind mobility facilitated through visual intracortical stimulation. Artif Organs, 36(7): 616-28. Motter BC (2009) Central V4 receptive fields are scaled by the V1 cortical magnification and correspond to a constant-sized sampling of the V1 surface. J Neurosci, 29(18): 5749 –5757. Nanduri D, Fine I, Horsager A, Boynton GM, et al. (2012) Frequency and Amplitude Modulation Have Different Effects on the Percepts Elicited by Retinal Stimulation. Invest Ophthalmol Vis Sci, 53: 205–214. Nassi JJ, Callaway EM (2009) Parallel processing strategies of the primate visual system. Nat Rev Neurosci, 10(5): 360-72. National Institutes of Health (2008) Leading Causes of Blindness. NIH Medline Plus, 3(3): 14-15. 138 | P a g e Park H-J, Lee JD, Kim EY, Park B, et al. (2009) Morphological alterations in the congenital blind based on the analysis of cortical thickness and surface area. NeuroImage, 98–106. Pascual-Leone A, Hamilton R (2001) The metamodal organization of the brain. Prog Brain Res, 134: 427-445. Pascual-Leone A, Torres F (1993) Plasticity of the sensorimotor cortex representation of the reading finger in Braille readers. Brain, 116(Pt 1): 39-52. Potts AM, Inoue J, Buffum D (1968) The electrically evoked response (EER) of the visual system. Invest Ophthalmol Vis Sci, 7: 269-278. Ptito M, Matteau I, Zhi Wang A, Paulson OB, et al. (2012) Crossmodal recruitment of the ventral visual stream in congenital blindness. Neural Plast, 2012: 304045. Ptito M, Moesgaard SM, Gjedde A, and Kupers R (2005) Cross-modal plasticity revealed by electrotactile stimulation of the tongue in the congenitally blind. Brain, 128: 606-614. Ptito M, Schneider FC, Paulson OB, Kupers R (2008) Alterations of the visual pathways in congenital blindness. Exp Brain Res, 187(1): 41-9. Raemaekers M, Bergsma DP, van Wezel RJ, van der Wildt GJ, et al. (2010) Effects of vision restoration training on early visual cortex in patients with cerebral blindness investigated with functional magnetic resonance imaging. J Neurophysiol, 105(2): 872-82. Röder B, Teder-Sälejärvi W, Sterr A, Rösler F, et al. (1999) Improved auditory spatial tuning in blind humans. Nature, 400(6740): 162-6. Sadato N, A, Grafman J, Ibanez V (1996) Activation of the primary visual cortex by Braille reading in blind subjects. Nature, 380(11): 526-528. Sadato N, Okada T, Honda M, Yonekura Y (2002) Critical period for cross-modal plasticity in blind humans: a functional MRI study. Neuroimage, 16(2): 389-400. Sadato N, Okada T, Kubota K, Yonekura Y (2004) Tactile discrimination activates the visual cortex of the recently blind naive to Braille: a functional magnetic resonance imaging study in humans. Neuroscience Letters, 359: 49-52. Sathian K (2005) Visual cortical activity during tactile perception in the sighted and the visually deprived. Dev Psychobiol, 46: 279–286. 139 | P a g e Schira MM, Tyler CW, Spehar B, et al. (2010) Modeling Magnification and Anisotropy in the Primate Foveal Confluence. PLoS Comput. Biol, 6(1). Schmidt EM, Bak MJ, Hambrecht FT, Kufta CV, et al. (1996) Feasibility of a visual prosthesis for the blind based on intracortical microstimulation of the visual cortex. Brain, 119(Pt 2): 507-22. Schoth F, Burgel U, Dorsch R, Reinges MH, et al. (2006) Diffusion tensor imaging in acquired blind humans. Neurosci Lett, 398(3): 178-82. Schwahn HN, Gekeler F, Kohler K, Kobuch K, et al. (2001) Studies on the feasibility of a subretinal visual prosthesis: data from Yucatan micropig and rabbit. Graefes Arch Clin Exp Ophthalmol, 239(12): 961-7. Shimony JS, Burton H, Epstein AA, McLaren DG, et al. (2006) Diffusion tensor imaging reveals white matter reorganization in early blind humans. Cereb Cortex, 16(11): 1653-61. Shivdasani MN, Fallon JB, Luu CD, Cicione R, et al. (2012) Visual cortex responses to single- and simultaneous multiple-electrode stimulation of the retina: implications for retinal prostheses. Invest Ophthalmol Vis Sci, 53(10): 6291-300. Shu N, Liu Y, Li J, et al. (2009) Altered Anatomical Network in Early Blindness Revealed by Diffusion Tensor Tractography. PLoS One, 4(9): 1-13. Smirnakis SM, Schmid MC, Weber B, Tolia As, et al. (2007) Spatial specificity of BOLD versus cerebral blood volume fMRI for mapping cortical organization. J Cereb Blood Flow Metab, 27(6): 1248-1261. Smith SM, Jenkinson M, Woolrich MW (2004) Advances in functional and structural MR image analysis and implementation as FSL. NeuroImage, 23(S1): 208-219. Srivastava NR and Troyk P (2005) A proposed intracortical visual prosthesis image processing system. Conf Proc IEEE Eng Med Biol Soc, 5: 5264-7. Sterr A, Müller MM, Elbert T, Rockstroh B, et al. (1998) Changed perceptions in Braille readers. Nature, 391(6663): 134-5. Stilla R, Hanna R, Hu X, Mariola E, et al. (2008) Neural processing underlying tactile microspatial discrimination in the blind: A functional magnetic resonance imaging study. J Vision, 8(10): 13, 1-19. 140 | P a g e Thambisetty M, Wan J, Carass A, et al. (2010) Longitudinal changes in cortical thickness associated with normal aging. NeuroImage, 52: 1215–1223. Tisserand DJ1, van Boxtel MP, Pruessner JC, et al. (2004) A voxel-based morphometric study to determine individual differences in gray matter density associated with age and cognitive change over time. Cereb Cortex, 14(9): 966-73. Torab K, Davis TS, Warren DJ, House PA, et al. (2011) Multiple factors may influence the performance of a visual prosthesis based on intracortical microstimulation: nonhuman primate behavioural experimentation. J Neural Eng, 8(3): 035001. Vaidya A, Borgonovi E, Taylor RS, et al. (2014) The cost-effectiveness of the Argus II retinal prosthesis in Retinitis Pigmentosa patients. BMC Ophthalmology, 14: 49. Van Boven RW, Hamilton RH, Kauffman T, Keenan JP et al. (2000) Tactile spatial resolution in blind braille readers. Neurology, 54(12): 2230-6. Veraart C, Duret F, Brelen M, Delbeke J (2004) Vision rehabilitation with the optic nerve visual prosthesis. Conf Proc IEEE Eng Med Biol Soc, 6: 4163-4. Volder AG, Bol A, Blin J, et al. (1997) Brain energy metabolism in early blind subjects: neural activity in the visual cortex. Brain Research, 750: 235–244. Walker HK, Hall WD, and Hurst JW (1990) Clinical Methods: The History, Physical, and Laboratory Examinations. “Visual Fields.” 3 rd ed. Boston: Butterworths. Wandell BA, Smirnakis SM (2009) Plasticity and stability of visual field maps in adult primary visual cortex. Nat Rev Neurosci, 10(12): 873-884. Watkins KE, Shakespeare TJ, O’Donoghue MC, et al. (2013) Early Auditory Processing in Area V5/MT of the Congenitally Blind Brain. J Neurosci, 33(46): 18242-18246. Weiland JD, Faraji B, Greenberg RJ, et al. (2012) Assessment of MRI issues for the Argus II retinal prosthesis. Magnetic Resonance Imaging, 30: 382–389. Wittenberg GF, Werhahn KJ, Wassermann EM, Herscovitch P, et al. (2004) Functional connectivity between somatosensory and visual cortex in early blind humans. Euro J Neurosci, 20: 1923-1927. 141 | P a g e Woolrich MW, Jbabdi S, Patenaude B, et al. (2009) Bayesian analysis of neuroimaging data in FSL. NeuroImage, 45:S173-86. World Health Organization (2012) Visual impairment and blindness. WHO Media Centre, Fact Sheet N°282: <http://www.who.int/mediacentre/factsheets/fs282/en/index.html>. Yushkevich PA, Zhang H, Simon TJ, Gee JC (2008) Structure-specific statistical mapping of white matter tracts. NeuroImage, 41(2): 448-61. Yves Rosseel (2012) Lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2): 1-36. Zangaladze A, Epstein CM, Grafton ST, Sathian K (1999) Involvement of visual cortex in tactile discrimination of orientation. Nature, 401(6753): 587-90. Zhang H, Awate SP, Das SR, et al. (2010) A tract-specific framework for white matter morphometry combining macroscopic and microscopic tract features. Medical Image Analysis, 14(5). Zrenner E, Bartz-Schmidt KU, Benav H, Besch D, et al. (2011) Subretinal electronic chips allow blind patients to read letters and combine them to words. Proc Biol Sci, 278(1711): 1489-97. Zrenner E (2012) Artificial vision: solar cells for the blind. Nature Photonics, 6: 344–345.

Abstract (if available)

Abstract Imaging studies show that vision deprivation causes the visual cortex to become responsive to tactile stimulation. The following work determined if this cross-modal activation in the primary visual cortex (V1) is correlated with vision loss in sighted individuals and late-blind patients with retinitis pigmentosa (RP)—an inherited degenerative photoreceptor disease that progressively diminishes vision—and further determined if these tactile-evoked responses were correlated with changes in resting-state and structural connectivity of the cortical visual streams. The effect of sight restoration on these tactile-evoked activities was determined by comparing V1 responses in Argus II patients with sighted and RP subjects. Based on our findings, we determined if tactile-evoked responses can account for visual performance variability among Argus II patients and gauged the feasibility of using cross-modal activation as a predictive measure for assessing outcomes in patients following a sight restoration treatment. ❧ RP patients, sighted control subjects, and recipients of the Argus II retinal prosthesis completed a series of tactile tasks in the scanner. We measured tactile-evoked blood oxygenation level dependent (BOLD) responses using functional magnetic resonance imaging (fMRI) and quantified the extent and strength of activation in V1

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Towards a high resolution retinal implant

PDF

Prosthetic visual perception: retinal electrical stimulation in blind human patients

PDF

Functional models of fMRI BOLD signal in the visual cortex

PDF

Cortical and subcortical responses to electrical stimulation of rat retina

PDF

Functional magnetic resonance imaging characterization of peripheral form vision

PDF

Manipulation of RGCs response using different stimulation strategies for retinal prosthesis

PDF

Prosthetic vision in blind human patients: Predicting the percepts of epiretinal stimulation

PDF

Electrical stimulation approaches to restoring sight and slowing down the progression of retinal blindness

Asset Metadata

Creator Cunningham, Samantha Irene (author)

Core Title Characterization of visual cortex function in late-blind individuals with retinitis pigmentosa and Argus II patients

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Biomedical Engineering

Publication Date 09/22/2015

Defense Date 07/22/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cross-modal plasticity,DTI,fMRI,OAI-PMH Harvest,primary somatosensory cortex,primary visual cortex,resting-state fMRI,retinal prosthesis,retinitis pigmentosa,tactile stimulation

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Tjan, Bosco S. (committee chair), Weiland, James D. (committee chair), Nayak, Krishna S. (committee member)

Creator Email samicunn@gmail.com,sicunnin@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-482340

Unique identifier UC11286800

Identifier etd-Cunningham-2969.pdf (filename),usctheses-c3-482340 (legacy record id)

Legacy Identifier etd-Cunningham-2969.pdf

Dmrecord 482340

Document Type Dissertation

Format application/pdf (imt)

Rights Cunningham, Samantha Irene

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA