There's an issue that precedes most of your questions:

To make sure the raw_dwi was actually getting divided, I went back and did 'fslmaths raw_dwi.nii.gz -div 10 raw_dwi_div10' and checked the intensity of the raw_dwi_div10, and it's the same intensity range as that of the input image. I'm not sure why fslmaths division isn't working here.


FSL eddy's movement output has two columns; each value in the first column corresponds to absolute movement relative to the first volume, and each value in the second column corresponds to movement relative to the previous volume. These movement estimates are in mm. So, I'm trying to exclude subjects with >2mm of movement between volumes.