FEIT Research Project Database

Machine learning prediction of biological ageing of the brain and body


Project Leader: Andrew Zalesky
Collaborators: Dr Ye Tian (Medicine) Dr Vanessa Cropley (Medicine)
Primary Contact: Andrew Zalesky (azalesky@unimelb.edu.au)
Keywords: machine learning; magnetic resonance imaging MRI; neuroimaging; neuroimaging and neuroinformatics
Disciplines: Biomedical Engineering,Electrical & Electronic Engineering
Domains:

Your brain and other organs may be older (or younger) than your chronological age! Why?

Aging is a progressive, generalised deterioration and loss-of-function across multiple organ systems. When you age by a year, your body an particular body organs may show signs of ageing that appear greater (or less than) what would be considered normal for a year, as determined relative to population norms. The apparent age of your heart, lungs and other organs is referred to as your biological age and can vary markedly from your chronological age.

Critically, organ systems age much faster in some individuals compared to others and advanced biological ageing is associated with increased risk of many age-related body and brain disorders such as cancer, coronary artery diseases and dementia and therefore, decreased life expectancy.

This project aims to establish cutting-edge machine learning models to predict the biological age of an individual's brain and bodily organs. Medical imaging data wil be used to train the model, using images from 30,000 individuals participating in the UK Biobank.

Whereas twin and family studies suggest moderate genetic influences on biological ageing, to what extent heritability can be explained by common genetic variants has not been well characterised. The extent to which environmental factors can modify the pre-determined genetic impact on ageing and whether interactions between genetic and environmental effects on ageing act differently across different organ systems is unknown.

Further research and key questions

  • Identify common genetic variants that associate with biological ageing using genome-wide meta-analyses
  • Subtype individuals based on their genetic variants and investigate whether individuals with certain genetic profiles are more susceptible/resilient to environmental effects on ageing. This would facilitate the development of intervention programs that could potentially delay the process of biological ageing

Further information: Check out our lab website for further details: www.sysneuro.org

Machine learning will be used to predict biological changes of ageing by organ system; Soto-Perez-de-Celis et al, 2018.
test