I created a binary mask by combining two Harvard-Oxford cortical masks to extract values for fALFF analyses (see attached). When I look at the new masks in fsleyes, they are binarized as intended (mask = 1, all other voxels = 0). However, when I go to into CONN after running the fALFF analysis and  "Import Values" and choose one of these ROI masks, it outputs data for 3 clusters. I only combined two masks to make the single mask so this makes no sense at all. I cannot figure out what 3 clusters CONN is "seeing". IAnyone have this happen to them? Any thoughts??? I have recreated the mask using SPM instead of fslmaths and I get the same result. 
Thank you
Amy