Approved Research

Magnetic susceptibility separation application to UK Biobank data

Lay summary

Recently, magnetic resonance imaging-based iron and myelin mapping method, that is, the susceptibility separation method, has been developed. This project aims to apply the susceptibility separation method to the UK Biobank data set using artificial intelligence (AI). Susceptibility separation is a reconstruction method that allows the estimation of iron and myelin contributions from magnetic resonance imaging. The proposed method is expected to have advantages as iron and myelin are known to correlate with various neurological diseases. However, this method requires a very specific kind of magnetic resonance imaging that is not widely used in everyday clinical settings. This requirement for an unusual protocol hampers the widespread application of susceptibility separation. Recently, an artificial intelligence-based methodology for extracting unknown features from magnetic resonance imaging has been proposed. In this study, AI is used to transform magnetic resonance images from the UK Biobank protocol into magnetic susceptibility separation maps. Applying this method to large-scale brain data from UK Biobank is expected to lead to new discoveries for the diagnosis of neurological diseases. The duration of this project is one year.