I think the problem I am having with using 3dbucket on .img files is that the .img file name (eg. fred.img) is in the subbrick label and the name of the dataset added to the bucket (clustered img file in this case; fred.cluster.BRIK & .HEAD) is the subbrick keyword. Afer I create a bucket dataset out of several of these .BRIK and .HEAD files created from the original .img file, and subsequently use afni to pick my subbrick, there is no useful label in the subbrick selector as all added datasets in the bucket are simply listed by the original .img file name.
I looked at 3drefit to try to swap the subbrick label and the keyword label, but didnt see that option. Appending the subbrick label wouldnt work because of the limited number of letters viewable when looking at the subbrick selector in afni (i.e. I wouldnt see the appended info anyway, my file name is too long).
I havent had this problem when using 3dbucket on .BRIK datasets.
thanks,
Christine Smith