Using neuroimaging data to map dysfunctional brain networks and predict symptom severity in Parkinson's disease progression.

点击翻译按钮获取中文摘要

A major challenge in Parkinson's disease (PD) research is heterogeneity in disease progression and severity of cognitive involvement. Conventional neuroimaging uses MRI to assess volume loss caused by neuronal cell death. However, grey matter atrophy is poorly sensitive in PD and fails to track cognitive change or motor progression. Techniques sensitive to brain tissue microstructure and to changes in network dynamics are better suited to do this. One such technique is quantitative susceptibility mapping (QSM), which estimates the absolute magnetic susceptibility of different brain tissues as a proxy for the distribution of brain iron. Oxidative stress secondary to excess brain iron accumulation is a pathomechanism in PD. Iron generates free radical species that interact with a-synuclein to promote Lewy-related pathology and produces neurotoxic by-products via catalysation of dopamine oxidation reactions. It is also linked with mitochondrial dysfunction, neurotoxicity and chronic inflammation. This PhD will use two complementary neuroimaging approaches to mapping disease activity in PD. The first section will use QSM data already acquired in 100 deeply phenotyped patients with Parkinson's disease to detect brain iron changes that correlate with disease activity in PD. An extension to this section will be to combine these maps of iron deposition with whole-brain transcriptional data to identify the relative expression of genes of interest in implicated regions. The second section will leverage resting state functional MRI data already acquired in the same cohort of patients. This will allow us to characterise the functional connectome, a detailed mapping of the functional integration of the brain, and how this changes in PD. An emerging approach to understanding the functional connectivity of resting state networks is spectral dynamic causal modelling (spectral DCM). This technique has the advantage of using Bayesian modelling to discover the most likely model from fully connected graphs, enabling us to characterise the connectivity patterns in different stages of PD. Together, this project aims to utilise novel neuroimaging techniques to track symptom severity in early stage PD and predict the rate of cognitive decline. By further relating our findings to whole-brain transcriptional data, we aim to shed light on the mechanisms underlying selective vulnerability and network dysfunction in PD.