Modelling brain ageing using neuroimaging to improve brain health in older adults

As we get older, our brain undergoes changes to both its structure and its function, so-called 'brain ageing'. These changes cause impairments to reaction time, memory and concentration and have a negative impact on people's ability to work or look after themselves. Older people are also more at risk of certain brain diseases, particularly dementia. However, brain ageing does not affect people uniformly; some people experience the negative consequences of brain ageing much earlier than others. Understanding the reasons for this variability will allow us to identify risk factors for negative brain ageing. To understand these individual differences we need a way of measuring them. To do this I have developed a 'biomarker' (i.e., biological measure) of brain ageing, which is based on magnetic resonance imaging (MRI) scans of people's brains. This allows us to accurately predict a person's actual age from the structure of their brain. I can then measure how typical an individual's brain looks for their age and identify individuals whose brains look unusually old or young for their actual age. Using this approach, I have shown that having an 'older-appearing' brain is linked to poorer cognitive and physical functioning, as well as a higher risk of death. I will measure brain-age in the UK Biobank study. This is the largest MRI study ever conducted and provides a unique opportunity to identify factors that influence the way brains age. My proposed work has four main goals:
1. To identify personal (age, gender, ethnicity), lifestyle and environmental (e.g., diet, smoking, exercise) factors related to positive or negative brain ageing.
2. To test whether specific genetic factors relate to brain ageing, helping us to understand why brain health problems can be inherited.
3. To explore the interactions between genes and the environment, where genes only cause poorer brain health under certain environmental conditions.
4. Design software for estimating brain-ageing in different regions of the brain, or in different aspects of brain function, such as blood flow and neural connectivity.
This study will help us understand what influences the way brains age, to establish whether these factors are modifiable and to determine important genetic influences. In the long-term, my approach could screen people to determine individual risk for cognitive decline and dementia, and help design better clinical trials of treatments aimed at protecting the ageing brain.

Older age is associated with cognitive decline, decreased well-being and increased risk of neurodegenerative disease and dementia. Fostering healthier 'brain ageing' would benefit for society by reducing these risks. However, the factors that specifically affect the brain are uncertain and brain ageing does not affect people uniformly. Hence, I have developed a neuroimaging-based biomarker of brain ageing, that quantifies individual deviation from healthy brain ageing; brain-predicted age difference (brain-PAD). I have shown that brain-PAD relates to cognitive performance, the presence of neurological disease and predicts future cognitive decline and dementia. Brain-PAD also relates to physiological ageing and predicts mortality risk in older adults, showing its utility as surrogate measure of future neurological and general health. Here, I will use brain-PAD to identify the genetic and environmental factors that influence brain ageing. This will provide mechanistic insights into causes of poor brain health and inform clinical efforts to promote healthier brains as people age. I will achieve this by measuring brain-PAD in the UK Biobank. This unprecedentedly large study currently includes high-resolution, multi-modality neuroimaging data from over 19,000 adults aged >40. Using the detailed demographic, behavioural, physiological, medical and genetic information collected from this cohort, I will:
1) Identify personal (age, sex), lifestyle and environmental (e.g., diet, smoking) factors that correlate with brain-PAD.
2) Determine genetic factors relating to brain-PAD, resulting in a polygenic risk score for brain ageing.
3) Explore gene-environment interactions in relation to brain-PAD, moving towards personalised risk assessments for poorer brain health during ageing.
4) Design a machine-learning software tool for estimating brain-predicted age at a regional level or in different components of brain structure and function.