Brain imaging modelling and inference at biobank scale
Stephen Smith, FIMRIB (Functional MRI of the Brain Centre)
Magnetic resonance imaging (MRI) of the brain has become an essential tool for probing the underlying causes of neurological and psychiatric diseases. However, tools for MRI analysis were developed for at most hundreds of subjects. This research project will focus on building statistical and computational techniques needed to scale up existing methods, in particular motivated by the UK Biobank’s brain MRI data on (ultimately) 100,000 subjects.
RESEARCH EXPERIENCE, RESEARCH METHODS AND TRAINING
After benchmarking existing methods for mass univariate modelling of functional MRI (fMRI) and structural MRI (sMRI) data, evaulations will be conducted on a number of techniques to accelerate these methods. Approaches will include subsetting space, subsetting subjects, data transformations and approximations.
FIELD WORK, SECONDMENTS, INDUSTRY PLACEMENTS AND TRAINING
Work will be carried out in close collaboration with the Wellcome Centre for Integrative Neuroscience (WIN), FMRIB Centre. There will be a possibility of secondment to another relevant research group internationally, e.g. ENIGMA Consortium at University of Southern California.
This project is ideally suited for a student with a degree in statistics, biomedical engineering, computer science or related field, with strong computational skills.