For some data, population_template's seems to crop the FOV. Deactivating the scale and drift correction via average_inv = None seems to fix it:

• https://community.mrtrix.org/t/population-template-fov-cropped-issue/5681
• https://community.mrtrix.org/t/population-template-fov/4199

I think I'll need test data where this is an issue to debug this. It'd be great if someone could share with me (maximilian.pietsch@kcl.ac.uk) the data (including masks and the exact command used) that reproduce this issue.

If you can not share the ODFs, could you try to reproduce the issue when using the masks instead of ODFs as input. If that leads to a cropped FOV, then please just share the masks. You do not need to wait for population_template to finish but should see the FOV problem when inspecting the template images located in the scratch directory.

If you can share ODFs, to reduce file size, I suspect you can use a subset of the data, can downsample the images and only use the first volume of the ODFs (lmax = 0). For instance, assuming odfs contains your ODFs, you can downsample to 1cm voxel grid and extract only the first volume with:

mkdir odfs_10mm
for f in odfs/*; do mrconvert "$f" -coord 3 0 - | mrgrid - regrid -voxel 10 odfs_10mm/"$(basename $f)"; done