Background: Cardiovascular diseases (CVDs) and dementia are the two age-related leading causes of morbidity and mortality worldwide. Increasing evidence shows that they share common risk factors and overlapping pathobiological mechanisms. This convergence is of growing concern given the global ageing of the population, the frequent coexistence of cardiovascular and neurodegenerative disorders in the same patients, and the lack of effective therapies for dementia.
This PhD project will pioneer a data-driven, generative framework to model the heart–brain axis, integrating cardiovascular and neuroimaging data to uncover potential mechanistic links between cardiovascular ageing and neurodegeneration. By leveraging large-scale population datasets such as the UK Biobank and TwinsUK, advanced computational methods, and causal inference approaches, this work aims to transform our understanding of how cardiovascular health influences brain ageing trajectories.
Objective 1: Develop a machine-learning model to estimate biological brain age (BrainAge) from MRI-derived brain phenotypes using gradient boosting (XGBoost) and identify cardiovascular traits associated with accelerated brain ageing.
Objective 2: Build a deep learning generative model (“Digital Heart–Brain”) to capture statistical and biophysical dependencies between cardiac function and brain phenotypes, enabling simulation of disease progression across systems.
Objective 3: Apply genome-wide association (GWAS) and Mendelian Randomisation (MR) analyses to establish causal relationships between cardiovascular ageing and neurodegenerative changes.
Deliverable: By combining deep learning, causal inference, and digital twin technologies, this project will provide a novel in-silico framework for predicting brain health from cardiac data. The outcomes could inform early diagnosis, risk stratification, and the development of new preventive or therapeutic strategies targeting both CVD and dementia.