I am conducting a seed-to-voxel analysis using several ROIs with comparisons between groups.  I recently added a new ROI and ran it through processing, choosing "do not overwrite (skip already-processed subjects/seeds)" at every step in order to save time. 

I then re-ran some analyses I had previously completed with existing ROIs that had already been run through the pipeline and found different results, despite having done nothing to change these ROIs.  Is there some explanation for why these results would have changed despite my not having modified the images or the old ROI?

Thank you in advance for your response.