Troubles running 3dvolreg.


I am trying to use 3dvolreg to correct for motion in functional MRI data to then be used in a tSNR calculation. As a sanity check to see if 3dvolreg was being used correctly, it was run on a phantom dataset with known tSNR and no motion. Once 3dvolreg ran however, the tSNR changed (went down by almost 1/3). If there should be no motion to correct then how is the dataset being altered in a way that can change tSNR? Perhaps I have a misunderstanding of how 3dvolreg should be used. The command I’ve been running is 3dvolreg -zpad 1 -base 0 ‘dataset’

Any help/advice is appreciated!