I am part of the MR Spectroscopic Imaging (MRSI) group and I am interested in acceleration of reconstruction of whole-brain MRSI data.
My work is focused on using methods of Deep Learning in our fields. The acquisition of whole-brain MRSI data requires advanced acquisition techniques e.q. non-cartesian encoding or undersampling of kSpace, which in combination with multi-channel coils generate a large amount of data (100GB per 15 minutes) and yield prolonged reconstruction time (up to a day for 100GB dataset).
My ultimate goal is to reach online reconstruction on the MR machine, which would provide immediate results of the MRSI scan.