Investigation of Parkinson's Cognitive Subtypes

The goal of this project is to aid our understanding of Parkinson's heterogeneity, which has been deemed, in the absence of any consensus throughout the past two decades, a priority in the field. The project will have three stages, each equivalent to one year of the PhD programme, in this order: identifying Parkinson's cognitive subtypes in a data driven way by distilling the literature and replicating the findings in a clinical cohort, carrying out neuroimaging on a subset of the cohort and finally relating the findings to short term fluctuations in motor symptoms.

The first stage will involve using novel computational methods (such as natural language processing) to distil the currently available literature. The aim is to use previous findings to infer a subtyping classification that can then be replicated in a clinical population. Subsequently, the classification system will be verified through the use of computerised cognitive tasks to determine cognitive phenotypes within that cohort. This will be carried out by analysing cognitive data coming from Parkinson's patients recruited from specialty clinics.

The second stage will involve carrying out functional brain network imaging using classical MRI, as well as ultra-high field MRI on a cohort of patients belonging to one selected subtype identified during the first stage of the project. Despite the use of neuroimaging being previously limited by logistic constrains, the potential posed by functional neuroimaging as a phenotyping tool makes this study particularly important to pursue. The goal would be to identify brain network phenotypes associated with one Parkinson's cognitive subtype, as well as volumetric changes associated with it.

Neuroimaging data collection will be ongoing alongside cognitive and behavioural data collection, allowing us to integrate the two into the same model. By monitoring the behaviour and cognition of patients using smart watches and computerised cognitive tasks during a short term longitudinal study, we could deepen our understanding of the factors that are variable within a certain subtype, and of factors which fluctuate and may result in misclassification if only clinical assessment was carried out.

The main outcomes of this project would be:
1) A more stable classification of Parkinson's disease rooted in cognitive phenotyping
2) Functional imaging analysis of one of the identified subtypes
3) Finer grain analysis of the cognitive and motor symptoms short term variability for one of the identified subtypes

There is also room for expanding the purpose of the project. Such an avenue could be creating a diagnosis and prognosis tool that uses computerised cognitive tasks to make inferences about the disease status of Parkinson's patients. Moreover, there would be the possibility of using more advanced computational modelling methodology to design a way of predicting how patients will change over time.

The impact of the project would be identifying cases that have a high susceptibility to cognitive decline early. This would be providing a platform for developing individualised treatments based on the cognitive manifestation of the disease. Ultimately this would result in a decreased socioeconomic burden, as dementia could be detected earlier.