Search

Parkinson's Disease Psychosis (PDP) is common, extremely challenging for both patients and carers, and yet poorly understood. Symptoms include illusions, hallucinations, and delusions. They negatively affect patients' quality of life and are a significant predictor of dementia. Nevertheless, basic questions about PDP remain unanswered due to limitations in existing research, including small sample sizes and a lack of diverse datatypes (i.e., clinical, cognitive, and from different neuroimaging modalities). The research in this doctoral thesis therefore aims to advance our understanding of the neuro-cognitive factors associated with important components of PDP - visual hallucinations and related perceptual symptoms - by combining innovative and robust methods with different clinical and neuroimaging datasets. To achieve this, I first explore the visuo-cognitive profile of PDP patients and employ novel meta-analytic methods in a large cohort of over 7,000 patients with and without hallucinations across 99 published studies. The results highlight the necessity of updating existing hallucinations models of Parkinson's Disease to account for impairments in a wider range of cognitive areas than previously addressed, which has significant implications for both clinical treatment and experimental research in the field. Secondly, I investigate changes in functional brain networks associated with PDP using a combination of graph-theoretical and data-driven methods. Different chapters address three main questions: what are the changes in functional brain networks associated with PDP? How can we best define and evaluate network impairments in this population? And how do these changes relate to cognitive and clinical variables of interest? These questions are investigated in the context of data from the Parkinson's Disease Progression Markers Initiative (PPMI), a large study of individuals recently diagnosed with Parkinson's Disease. Finally, I use a data-driven approach to investigate the changes in structural brain networks associated with psychosis in PD both cross-sectionally and longitudinally in an imaging cohort that is significantly larger than those previously utilized in longitudinal investigations of PDP. To achieve this and leverage the power of multimodal imaging, I employ a novel method (Morphometric Similarity Networks - MSNs) that can jointly synthesise data from multiple imaging modalities and provide a more comprehensive characterization of the changes in brain structure associated with psychotic symptoms in Parkinson's Disease. Taken together, this thesis not only advances our knowledge of important visuocognitive and network-related changes underlying psychosis in Parkinson's Disease, but also provides a steppingstone for future research investigations with direct clinical applications for these patients.

The role of the medial temporal lobes (MTL) in visuo-spatial learning has been extensively studied and documented in the neuroscientific literature. Numerous animal and human studies have demonstrated that the parahippocampal place area (PPA), which sits at the confluence of the parahippocampal and lingual gyri, is particularly important for learning the spatial configuration of objects in visually presented scenes. In current visuo-spatial processing models, the PPA sits downstream from the parietal lobes which are involved in multiple facets of spatial processing. Yet, direct input to the PPA from early visual cortex (EVC) is rarely discussed and poorly understood. This thesis adopted a multimodal neuroimaging analysis approach to study the functional anatomy of these connections. First, the pattern of structural connectivity between EVC and the MTL was explored by means of surface-based 'connectomes' constructed from diffusion MRI tractography in a cohort of 200 healthy young adults from the Human Connectome Project. Through this analysis, the PPA emerged as a primary recipient of EVC connections within the MTL. Second, a data-driven clustering analysis of the PPA's connectivity to an extended cortical region (including EVC, retrosplenial cortex, and other areas) revealed multiple clusters with different connectivity profiles within the PPA. The two main clusters were located in the posterior and anterior portions of the PPA, with the posterior cluster preferentially connected to EVC. Motivated by this result, virtual tractography dissections were used to delineate the medial occipital longitudinal tract (MOLT), the white matter bundle connecting the PPA with EVC. The properties of this bundle and its relation to visual configuration learning were verified in a different, cross-sectional adult cohort of 90 subjects. Finally, the role of the MOLT in the visuo-spatial learning domain was further confirmed in the case of a stroke patient who, after bilateral occipital injury, exhibited deficits confined to this domain. The results presented in this work suggest that the MOLT should be included in current visuo-spatial processing models as it offers additional insight into how the MTL acquires and processes information for spatial learning.

Visual snow is a neurological condition that was first described and defined clinically a little over five years ago. The main clinical feature is an unremitting, positive visual phenomenon present in the entire visual field and characterized by uncountable tiny flickering dots. The pathophysiology of this disorder is largely unknown, making treatment of affected patients extremely challenging. The aim of this PhD was to understand more about the biology of visual snow, first through its clinical characterization and then with the use of neuroimaging. The first objective has been addressed with a large web-based survey that involved visual snow subjects (n = 1104), who were defined according to the recently proposed diagnostic criteria. This study (chapter 2) allowed recognition of the common clinical presentation of the syndrome and of its principal comorbidities, as well as prediction of its severity. The second aim was carried out with a multimodal structural and functional magnetic resonance imaging study, comparing twenty-four visual snow patients with an equal number of healthy volunteers (chapter 3). Voxel based-morphometry (chapter 4) was used to study structural changes in the brain; arterial spin labelling (chapter 5) to analyse regional cerebral blood flow; functional connectivity (chapter 6) and task-based functional magnetic resonance imaging (chapter 0) to study localised brain responses at rest and in response to a ‘visual snow-like’ stimulus; finally magnetic resonance spectroscopy (chapter 0) was employed to study neurochemical properties of the brain. The imaging data show that visual snow is characterized by subtle morphological, as well as widespread functional, changes that involve important brain networks regulating attention, salience, sensory processing and cognition.