2024-25 Project (Rudnicka & Lewin & Owen)
AI methodology for investigating age-related cognitive changes and retinal imaging
SUPERVISORY TEAM
Supervisor
Professor Alicja Rudnicka at SGUL
Email: arudnick@sgul.ac.uk
Co-Supervisor
Dr Alex Lewin at LSHTM
Email: alex.lewin@lshtm.ac.uk
Co-Supervisor
Professor Christopher Owen at SGUL
Email: cowen@sgul.ac.uk
PROJECT SUMMARY
Project Summary
This project will use AI/machine learning methods to investigate the association between age-related cognitive changes and age-related changes in retinal vessels.
The shape and size of vessels from high-dimensional retinal images and Retinal Nerve Fibre Layer (RNFL) thickness measures from Optical Coherence Tomography (OCT) images have been found to be strongly related to age. The object of this project is to use these morphological features in machine learning models for cognitive changes over time, investigating relations between cognitive measures, retinal imaging data and age. Cognitive change will be quantified using multivariate observations (prospective memory, numeric and verbal reasoning), measured at repeated time points in around 55,000 UK Biobank participants. Morphological features derived from retinal images are available at baseline for the same sample.
Examples of research questions include:
1. Are age-related changes in retinal images predictive of poorer cognition in later life?
2. Can we detect particular areas of the retinal vessels which drive associations between age and cognition?
3. Are cognitive changes with age mediated through changes in the retinal vessels?
Project Key Words
Big data, Bayesian modelling, Machine learning, Spatial modelling
MRC LID Themes
- Global Health = Yes
- Health Data Science = Yes
- Infectious Disease = No
- Translational and Implementation Research = Yes
Skills
MRC Core Skills
- Quantitative skills = Yes
- Interdisciplinary skills = Yes
- Whole organism physiology = No
Skills we expect a student to develop/acquire whilst pursuing this project
These questions will require the use of high-dimensional supervised learning models. These will be extended to incorporate spatially-dependent predictors for finding which parts of the vessels are important, and to mediation approaches for investigating the three-way dependence between retinal images, age and cognition.
The student will join an established team of investigators, including statisticians, epidemiologists, image scientists, and clinicians, working at Moorfields/UCL, LSHTM, St George’s and Kingston Universities.
Routes
Which route/s is this project available for?
- 1+4 = Yes
- +4 = Yes
Possible Master’s programme options identified by supervisory team for 1+4 applicants:
- LSHTM – MSc Health Data Science
- LSHTM – MSc Medical Statistics
Full-time/Part-time Study
Is this project available for full-time study? Yes
Is this project available for part-time study? Yes
Eligibility/Requirements
Particular prior educational requirements for a student undertaking this project
- SGUL’s standard institutional eligibility criteria for doctoral study.
- Strong quantitative background.
Other useful information
- Potential CASE conversion? = No
PROJECT IN MORE DETAIL
Scientific description of this research project
This project will use AI/machine learning methods to investigate the association between age-related cognitive changes and age-related changes in the retina.
Retinal vasculometry features from high-dimensional retinal vessel images and Retinal Nerve Fibre Layer (RNFL) thickness measures from Optical Coherence Tomography (OCT) images have been found to be strongly related to age. The object of this project is to use these morphological features in machine learning models for investigating cognitive changes over time, relations between cognitive measures, retinal imaging data and age. Cognitive change will be quantified using multivariate observations (prospective memory, numeric and verbal reasoning), measured at repeated time points in around 55,000 UK Biobank participants who also underwent retinal imaging. Retinal vasculometry features derived from retinal images (and RNFL measures in a sub-set) are available at baseline for the same sample.
Primary questions to be answered:
(a) Are age-related changes in retinal images predictive of poorer cognition in later life?
(b) Can we detect particular areas of the retinal vascular tree which drive associations between age and cognition?
(c) Are cognitive changes with age mediated through changes in the retinal vessels?
TECHNIQUES
(a) Multivariate supervised learning approach.
(b) Feature selection in multivariate supervised learning, extension to spatially-structured predictors.
(c) Extension to mediation models with multiple mediators, including spatial structure.
DATA AVAILABILITY
The project will use UK Biobank data. The supervisory team already has access to all data to be used in the project.
RISK MITIGATION
Data are already in place. The project could involve methodology and software development for mediation analysis, which can be a lengthy process. However, existing software could be used for the feature selection and spatial analysis questions.
Further reading
(Relevant preprints and/or open access articles)
Additional information from the supervisory team
- The supervisory team has provided a recording for prospective applicants who are interested in their project. This recording should be watched before any discussions begin with the supervisory team.
Rudnicka-Lewin-Owen Recording
MRC LID LINKS
- To apply for a studentship: MRC LID How to Apply
- Full list of available projects: MRC LID Projects
- For more information about the DTP: MRC LID About Us