I have done two group analyses for a contrast of interest in fsl. when I looked more carefully in fsleyes at the cope and zstats files, I saw that there are some voxels, in both analyses, where the cope (and zstats) values equal zero. These voxels are concentrated at the edges between the two hemispheres, but they are definitely part of the brain. When I look at the files from level-1 for different subjects, there is nothing special in these voxels and everything looks normal.
Does anyone have an idea what does this mean?