ANTsLongCT utilizes the output of ANTsSST and ANTsPriors to get volume, cortical thickness and gray matter density values for every region in the DKT atlas. Cortical thickness is computed using ANTs' cortical thickness algorithm.
You must install Docker to use the ANTsLongCT Docker image.
After Docker is installed, pull the ANTsLongCT image by running the following command:
docker pull pennbbl/antslongct:0.0.10
.
Typically, Docker is used on local machines and not clusters because it requires root access. If you want to run the container on a cluster, follow the Singularity instructions.
Here is an example from one of Ellyn's runs:
docker run --rm -ti -e projectName="ExtraLong" -e subLabel="bblid" \
-v /Users/butellyn/Documents/ExtraLong/data/singleSubjectTemplates/antssst5/sub-10410:/data/input/sub-10410 \
-v /Users/butellyn/Documents/ExtraLong/data/groupTemplates/antspriors:/data/input/antspriors \
-v /Users/butellyn/Documents/ExtraLong/data/corticalThickness/antslongct3/sub-10410:/data/output \
pennbbl/antslongct:0.0.10
- Line 1: Specify environment variables: the name of the project without any spaces
(
projectName
), and what you want the subject label column to be called (subLabel
). - Line 2: Bind a subject's ANTsSST output directory
(
/Users/butellyn/Documents/ExtraLong/data/singleSubjectTemplates/antssst5/sub-100079
) to the subject's ANTsSST directory in the container (/data/input/antssst/sub-100079
). Note that theantssst
directory outside of the container must start with the stringantssst
, but after that can contain any other characters. Ellyn has it asantssst5
because she got good output on her fifth try. - Line 3: Bind the group template directory (
/Users/butellyn/Documents/ExtraLong/data/groupTemplates/antspriors
) to its spot in the container (/data/input/antspriors
). - Line 4: Bind the directory where you want your ANTsLongCT output to end up
(
/Users/butellyn/Documents/ExtraLong/data/corticalThickness/antslongct3/sub-10410
) to the output directory in the container (/data/output
). - Line 5: Specify the Docker image and version. Run
docker images
to see if you have the correct version pulled.
Substitute your own values for the files/directories to bind.
You must install Singularity to use the ANTsLongCT Singularity image.
After Singularity is installed, pull the ANTsLongCT image by running the following command:
singularity pull docker://pennbbl/antslongct:0.0.10
.
Note that Singularity does not work on Macs, and will almost surely have to be installed by a system administrator on your institution's computing cluster.
Here is an example from one of Ellyn's runs:
SINGULARITYENV_projectName=ExtraLong SINGULARITYENV_subLabel=bblid singularity run --writable-tmpfs --cleanenv \
-B /project/ExtraLong/data/singleSubjectTemplates/antssst5/sub-10410:/data/input/sub-10410 \
-B /project/ExtraLong/data/groupTemplates/antspriors:/data/input/antspriors/ \
-B /project/ExtraLong/data/corticalThickness/antslongct3/sub-10410:/data/output \
/project/ExtraLong/images/antslongct_0.0.10.sif
- Line 1: Specify environment variables: the name of the project without any spaces
(
projectName
), and what you want the subject label column to be called (subLabel
). - Line 2: Bind a subject's ANTsSST output directory
(
/project/ExtraLong/data/singleSubjectTemplates/antssst5/sub-10410
) to the subject's ANTsSST directory in the container (/data/input/antssst/sub-100079
). Note that theantssst
directory outside of the container must start with the stringantssst
, but after that can contain any other characters. Ellyn has it asantssst5
because she got good output on her fifth try. - Line 3: Bind the group template directory (
/project/ExtraLong/data/groupTemplates/antspriors
) to its spot in the container (/data/input/antspriors
). - Line 4: Bind the directory where you want your ANTsLongCT output to end up
(
/project/ExtraLong/data/corticalThickness/antslongct3/sub-10410
) to the output directory in the container (/data/output
). - Line 5: Specify the Singularity image file.
Substitute your own values for the files/directories to bind.
See this script for an example of building a launch script per subject.
- For details on how ANTsLongCT was utilized for the ExtraLong project (all longitudinal T1w data in the BBL), see this wiki.
- Set home directory in Dockerfile.
- Get volumes of subcortical labels (will require tinkering with ANTsPriors).
- Evaluate whether method of computing GMD is comparable to Stathis' method.
- Improve overlap between cortical thickness image and DKT labels.
- Fix ANTs seed.