Project title
Automated retinal imaging analysis systems for prediction of cognitive and neurodegenerative outcomes in very large UK population cohorts
Supervisory team
About the Project
Early detection of cognitive decline allows identification of those at high risk of developing dementia when medical treatments may be effective in preventing disease onset.
Our group have developed an AI-enabled system which can extract detailed retinal vasculometry characteristics from colour fundus photographs (CFPs) in a fully automated way, allowing application to large population-based studies. Using data from 11 million vessel segments from over 200,000 CFPs, we have shown that retinal vasculometry can predict cardiovascular disease as well as established risk scores and predict novel outcomes such as circulatory mortality. Retinal vasculometry (a neurovascular biomarker) as a predictor of cognitive/neurodegenerative status is yet to be examined at scale. While end-to-end AI approaches have recently been used to detect Alzheimer’s disease from CFPs among those with established disease (in case-control studies), the use of such approaches to predict disease (i.e., in early ‘prodromal’ stages) is yet to be established in large community settings.
This PhD project will examine the effectiveness of AI-based analysis of eye images in predicting cognitive/neurodegenerative status, in exceptionally large population-based studies (e.g., UK Biobank and EPIC Norfolk cohorts with over 100,000 study participants, with repeat assessment in a sub-set). Findings will provide an early warning system of dementia, to be exploited within existing healthcare pathways to trigger early intervention.
Skills acquisition
The studentship will provide training in data-analytics, statistics including predictive modelling, computer vision and epidemiology.
The student will join an established team of investigators, including statisticians, epidemiologists, image scientists, and clinicians, working at Moorfields/UCL, St George’s and Kingston Universities.
Entry requirements
Applicants are expected to hold (or about to obtain) a minimum upper second class undergraduate honours degree (or equivalent) in a biomedical science. Experience in neuroscience, data analytics / statistics and AI is desirable.
Funding
The studentship provides funding for four years full-time and includes Home tuition fees plus a tax-free stipend in line with UKRI rates. We welcome applications from international students (EU and rest of the world), but they will normally be required to cover the difference between Home and International tuition fee rates.
Application process
Please send the completed application form to stgeorgesphd@sgul.ac.uk by no later than Wednesday 4 September 2024, 17.00 BST. An equal opportunities form should also be submitted as a separate document. References will be requested should you be successful in being offered the studentship.
Applications will undergo shortlisting and successful applicants will then be invited to interview during the week commencing 9 September 2024.
The successful candidate will be given a verbal offer and once it has been accepted, will be sent a formal offer letter and, in due course, a registration pack with joining information.
Unsuccessful candidates will be contacted with their outcomes at the earliest opportunity and will be able to request feedback if required.