PhD Studentship Probabilistic prediction of neurodegenerative disorders from neuroimaging data using machine learning techniques |2021-2022 Scholarships Grants

To start:Â October 2013

Project:

This is a four year project to develop and apply novel statistical and machine learning techniques to a wide range of magnetic resonance (MR) neuroimaging data (e.g. structural MRI, diffusion weighted MR imaging, cerebral blood flow maps and magnetic resonance spectroscopy), both individually and in combination. These methods will be applied in the first instance to discriminate healthy volunteers from patients diagnosed with Alzheimerâs disease (AD) and those deemed at high risk of developing the disorder due to mild cognitive impairment (MCI) as well as to predict future disease outcome (e.g. conversion from MCI to AD).Â

In a second stage these methods will be applied to additional clinical data sets to assess their robustness and validity.Â

During the course of the project, a variety of methods for extracting discriminative information from the data will be developed, in addition to methods to integrate prior clinical knowledge in a principled fashion. Behavioural, genetic and other potential biomarkers (e.g. proteomics) will be integrated into the assessment in order to increase accuracy and improve interpretability of the predictive models.

Entry requirements:

Applicants should have (or be expected to obtain) a 2:1 or 1st class honours degree in a numerate discipline (e.g. engineering, mathematics, computer science or physics) and have experience of writing computer software (e.g. in Matlab, C/C++, R, Python or Java). Prior experience of working with neuroimaging data and/or statistical methodology is also desirable. If applicants already possess (or expect to obtain) a research-based MSc degree, a merit or distinction level is required.