I'm working with SPM fMRI analysis. I have recognized that after the realignment step SPM sets nii.hdr.dime.pixdim(5) to zero which is meant to be the TR I think. This item is not zero before preprocessing. This is also happening after normalising and smoothing. Why is that?