Age is the strongest risk factor for Alzheimer’s disease (AD) and other neurodegenerative diseases. A better understanding of mechanistic links between age and AD is an urgent priority to develop effective strategies to deal with their rising burden amid an ageing population. Recently, machine learning techniques have gained popularity as brain-age prediction models due to their ability in identifying relevant data-driven patterns within complex data. These models learn the association between chronological age and cerebral morphological features derived from magnetic resonance imaging (MRI) in healthy individuals, yielding a predicted brain-age for each individual. Individuals with a predicted brain-age higher than their chronological age may have an “older” brain than expected, whereas an individual with an estimated brain-age lower than their
chronological age has a “younger” brain. Recent literature has shown the adequacy of using brain-age in the assessment of the clinical severity of AD and other diseases. Moreover, we previously implemented a brain-age prediction model that estimates brain-age from T1-weighted structural MRI. We showed that brain-age delta is positively associated with biomarkers of AD, as well as with unspecific markers of neurodegeneration (https://pubmed.ncbi.nlm.nih.gov/37067031/ ).
Given that aging affects multiple aspects of brain structure and function, developing brain-age prediction models from other neuroimaging modalities can provide a more comprehensive explanation of the mechanisms underlying individual differences in brain aging. The aim of this project is to implement a white matter brain-age from Diffusion weighted imaging (DWI) measurements and to evaluate its associations with AD biomarkers and risk factors. DWI is a technique for monitoring white matter microstructural impairment. Recent literature has studied the relationship between white-matter brain-age and lifestyle factors and cognition.

The student will implement, train and validate a machine or deep learning model from diffusion weighted imaging. The model will be trained using measurements from the UK BioBank cohort (N= 22,661) and tested on measurements from the ALFA cohort (N=418). Then, the associations between the white matter brain-age and different biomarkers of AD will be studied.

This project is in collaboration with Fundació Pasqual Maragall.