Mindboggle with longitudinal data and a custom template

1

I have a few questions about running mindboggle on data that has already been processed with ANTs & freesurfer

  1. Is it required that antsCorticalThickness was run with the OASIS-30_Atropos_template
  2. Can mindboggle handle the results of data processed longitudinally, i.e antsLongitudinalCorticalThickness (also applies to freesurfer)

I have data with a custom-template that has been processed longitudinally so both of these apply. For ANTs output it should be fairly trivial to create new directory for each subject with the brain mask and segmentation and then generate composed transform files that include the transform from custom-template to OASIS-30 template. I’m curious if there is a more elegant existing solution. I am much less familiar with freesurfer so I’m wondering if I should expect similar issues there with longitudinally processed data?

2

Hello!

  1. Not at all. Mindboggle performs an affine registration from that template to MNI152 space, so you should include a new transform from your template to MNI152 space and use the --my_transform flag when you run the mindboggle command.

From line 423 of Mindboggle’s main (nipype) script:

# ----------------------------------------------------------------------------
# Affine transform from Atropos template to MNI152 space:
# ----------------------------------------------------------------------------
if my_transform:
    affine_template2mni = my_transform
else:
    affine_template2mni = fetch_file_path(atropos_to_MNI152_affine)

(atropos_to_MNI152_affine = ‘OASIS-30_Atropos_template_to_MNI152_affine.txt’)

  1. Mindboggle runs one brain at a time, so it would treat each of the longitudinal samples independently.