Signal drift correction

Hi all,

Perhaps a silly question, but is signal drift compensated for by any MRtrix tool? I know there is a functionality in ExploreDTI, specifically called “signal drift correction”, but in MRtrix, there is only dwinormalise, if I am not mistaken. Does dwinormalise do the trick?


Hi Daniel,

dwinormalise only applies a global scaling factor to all DWI data for an individual subject, so doesn’t account for signal drift.

The coming update has a new script dwicat that attempts to account for differences in intensity scaling between distinct acquisition protocols; not “drift” per se, moreso trying to correct for scanner re-calibration between acquisitions. It would however be somewhat of a natural extension of this concept to estimate smooth signal drift within each input series in addition to between-series scaling; hence why I already added it as a wishlist item. Indeed this would be a pretty neat project for anyone interested in doing some development contribution, since it can almost entirely be achieved by wrapping up existing MRtrix3 commands in a Python script…