Data Science approaches to investigating the vascular footprint of Alzheimer's disease for early disease detection

Project Description: Using Machine Learning models to find early biomarkers in the microvascular footprint of the retina to perform early Alzheimer's disease detection.

The specific aims are:
- Develop deep learning approaches to vessel segmentation in OCTA images that preserve network connectivity for topological network characterisation.
- Investigate biomarkers across the whole depth of the retina (superficial capillary plexus, deep capillary plexus, choriocapillaris).
- Develop multi-modality retinal vascular phenotyping based on OCTA and the current standard of care fundus camera and OCT.
- Establish whether the approach contributes predictive biomarkers indicative of the risk of developing AD in asymptomatic individuals or prognostic biomarkers to supplement or even replace existing clinical trial enrolment criteria currently based on expensive (e.g. PET) or invasive (e.g. CSF sampling) biomarkers.