Skip to content

antsCorticalThickness and antsLongitudinalCorticalThickness output

Philip Cook edited this page Mar 11, 2021 · 8 revisions

antsCorticalThickness.sh

antsCorticalThickness.sh output is to a single directory, with multiple files.

Definitions

Subject space - The native space of the first anatomical input (usually T1).

Template space - The space of the template image provided at run time. Optionally, the segmentation template (used for brain extraction and priors) and registration template (used for final, high-quality registration on brain-extracted images) may be different. In this case, the output that is in template space will be to the registration template. Often, the registration template will be the same as the segmentation template, but with the brain extracted.

Some files are produced only if optional input parameters are set. The full file name will be ${outputPrefix}description${outputSuffix}, the prefix and suffix are set at run time by the user.

  • BrainExtractionMask- Brain extraction mask in subject space.

  • BrainNormalizedToTemplate - Extracted brain image normalized to the template space.

  • BrainSegmentation0N4 - Input to the segmentation algorithm. It is not brain extracted, but is bias-corrected. If multiple images are used for segmentation, there will be BrainSegmentation1N4 and so on. The brain extracted version of this is ExtractedBrain0N4.

  • BrainSegmentation - Segmentation image, one label per tissue class. The number of classes is determined by the input priors.

  • BrainSegmentationPosteriors1 - Posterior probability of class 1 (usually CSF). A similar image is produced for all classes. The numbering scheme matches the input priors.

  • CorticalThickness - Cortical thickness image in subject space.

  • CorticalThicknessNormalizedToTemplate - Cortical thickness image in template space.

  • ExtractedBrain0N4 - Brain-extracted version of BrainSegmentation0N4.

  • SubjectToTemplate1Warp, SubjectToTemplate0GenericAffine.mat - Transforms to be used when warping images from the subject space to the template space (see below).

  • SubjectToTemplateLogJacobian - Log of the determinant of the Jacobian, quantifies volume changes in the subject to template warp.

  • TemplateToSubject0Warp, TemplateToSubject1GenericAffine.mat - Transforms to be used when warping images from the template to the subject space (see below).

Applying the warps

The warps produced by antsCorticalThickness.sh are named differently to the usual antsRegistration convention. The inverse affine transform is written to disk, so users do not need to invert the affine transform explicitly in the call to antsApplyTransforms. The template is the fixed image in this registration so the files containing SubjectToTemplate are the forward warps and the Jacobian derived from the forward warps.

To warp an image from the subject to template space:

antsApplyTransforms -d 3 -i imageInSubjectSpace.nii.gz -o imageInTemplateSpace.nii.gz \
-t outputPrefixSubjectToTemplate1Warp.nii.gz -t outputPrefixSubjectToTemplate0GenericAffine.mat -r template.nii.gz

To warp an image from the template to subject space:

antsApplyTransforms -d 3 -i imageInTemplateSpace.nii.gz -o imageInSubjectSpace.nii.gz \
-t outputPrefixTemplateToSubject1Affine.mat -t outputPrefixTemplateToSubject0Warp.nii.gz -r template.nii.gz

Note that the ordering of the affine and warp field are reversed, but the numbering is consistent (1 then 0, reading left to right).

antsLongitudinalCorticalThickness.sh

antsLongitudinalCorticalThickness.sh first builds a single-subject template (SST) out of all available time points, and then runs antsCorticalThickness.sh on the SST as though it was an individual image, normalizing it to a group template. The SST and the output of antsCorticalThickness.sh are used to generate segmentation priors in the SST. Then antsCorticalThickness.sh is run on the individual time point images, using the SST as the template.

The outputs are organized in one folder per input time point, plus one folder for the SST. The output within each folder is similar to that of antsCorticalThickness.sh

There are additional warps, suffixed "SubjectToGroupTemplateWarp.nii.gz" and "SubjectToTemplate0GenericAffine.mat", that can be used to warp each time point image to the group template. These are a combination of the subject to SST warp, and the SST to group template warp.

Clone this wiki locally