TIGRLab/enigmaDTI

Name: enigmaDTI

Owner: Kimel Family Translational Imaging-Genetics Research Lab

Description: A DTI QC Pipeline

Created: 2014-10-08 16:10:59.0

Updated: 2017-07-11 17:21:09.0

Pushed: 2017-07-11 17:32:08.0

Homepage: null

Size: 77454

Language: Shell

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

ENIGMA QC Pipeline for DTI Acquisitions

See here http://enigma.ini.usc.edu/ongoing/dti-working-group/ for details, and the originators of this project.

This pipeline takes an input FA map from a DTI acquisition and reports (!!!Spooky Various Things!!!) about your DTI data.

This project was mostly produced by David Rotenberg, the Rotenator.

Instructions

This code will print out the average FA value along a skeleton and ROI values on that skeleton according to an ROI atlas, for example the JHU-ROI atlas (provided with this code).

The look up table for the atlas should be tab-delimited – the first column should refer to the voxel value within the image and the second column should refer to the desired label of the region.

./singleSubjROI_exe look_up_table.txt skeleton.nii.gz JHU-WhiteMatter-labels-1mm.nii.gz OutputfileName Subject_FAskeleton.nii.gz

Example:

./singleSubjROI_exe ENIGMA_look_up_table.txt mean_FA_skeleton.nii.gz JHU-WhiteMatter-labels-1mm.nii.gz Subject*_ROIout Subject*_FA.nii.gz

DTI Quick Start

To properly generate FA maps, TBSS, and ENIGMA QC for each DTI acquisition, you need the following:

./SUBJECT/T2.nii.gz
./SUBJECT/DTI_60DIR/DTI.nii.gz
./SUBJECT/DTI_60DIR/DTI.bvecs
./SUBJECT/DTI_60DIR/DTI.bvals

Then, run something like this (from the folder containing all the subjects):

1) ENIGMA_pre -s ${SUBJECTNAME} -m FNIRT 2) ENIGMA_main.sh

NOTE

This is a WIP (hardcore) and there is lots of junk in /BKUP that is to-be-determined useful. Don't venture in there if you are faint of heart.

Full Pipeline

1) DWI Preprocessing You can call DIST_CORR.sh as follows:

./DIST_CORR.sh -s (SUBJECT) -m (FNIRT) -b (FSL)

Here, SUBJECT can be anything, however the script expects a sub-directory DTI_DIR60.

This subdirectory should contain one .nii.gz and a .bval and .bvec file. It doesnt matter if there are DICOMS present as well.

In addition to the DTI_DIR60 directory, there must be a T2.nii.gz file directly within the subjects directory.

Other than this, you should have the correct modules for FSL and MINC loaded that currently work in your environment for the registrations commands required by the script.

NOTE: If this gives you any trouble at all, we can use a more rudimentary eddy_correct script

2) ENIGMA

The ENIGMA master script can be first referenced using a new module (ENIGMA) which you can load into your environment. This script should be called directly, from any directory that contains (only) the FA maps in .nii.gz or .nii format. TBSS should run to completion and the ENIGMA regions of interest should be extracted.

3) EXTRACTION Script

The extraction script has a higher level, but I will ignore this for the time being as that is the part you will be modifying. I have attached one of the current extraction scripts for DTIG1MR, that uses the included text file to match up the series and scan types. It may not be useful in and of itself, but it should indicate how things are currently structured.

Let me know if you have any questions!

ENIGMA DTI extraction protocol (April 10, 2014)

REQUIRED:

ENIGMA FA average, mask, skeleton, skeleton mask, and distance map (are all found in ?/home/dan/Downloads/enigmaDTI?).

Each subject to be analyzed must be preprocessed using /dti/Merged_data/Scripts/TBSS_Preproc (for 3 repetitions) or /dti/Merged_data/Scripts/tbss_preproc (for single repetition), as the input for step 1 requires an FA map for each subject.

1. Perform TBSS on all subjects FA maps as follows:

tbss_1_preproc *.nii.gz

tbss_2_reg -t ENIGMA_DTI_FA.nii.gz

tbss_3_postreg -S

1. Extract registered FA maps from ?FA/? directory (FA_to_target.nii.gz) and place in new

directory (?FA_to_target/?)

1. cd into ?FA_to_target/? and skeletonize each registered FA map as follows:

for a in *; do tbss_skeleton

-i ENIGMA_DTI_FA.nii.gz

-p 0.049

ENIGMA_DTI_FA_skeleton_mask_dst.nii.gz

/usr/share/fsl/4.1/data/standard/LowerCingulum_1mm.nii.gz

${a}

${a}_FAskel

-s ENIGMA_DTI_FA_skeleton_mask.nii.gz ;

done

NOTE: the ENIGMA* templates in the above script may not be in the current working directory (i.e.

?FA_to_target/?), so ensure that where they are referenced (lines 3, 5, and 9), you provide a full

pathname so that the script can find the template in your file system.

1. mv new FA skeletons (_FAskel.nii.gz) to a new directory (?FA_skels/?)

2. Perform ROI extraction using modified run_ENIGMA_ROI_ALL_script.sh found

in ?/projects/dan/ROIextraction_info?, by changing datapath (line 15 in run_ENIGMA_ROI_ALL_script.sh) to ?FA_skels? (make sure to include full path name)

1. .csv files containing FA values for each ENIGMA ROI will be contained in the ?ENIGMA_ROI_part1? directory in ?/projects/dan/ROIextraction_info? This code will print out the average FA value along a skeleton and ROI values on that skeleton according to an ROI atlas, for example the JHU-ROI atlas (provided with this code). The look up table for the atlas should be tab-delimited – the first column should refer to the voxel value within the image and the second column should refer to the desired label of the region.

   ./singleSubjROI_exe look_up_table.txt skeleton.nii.gz JHU-WhiteMatter-labels-1mm.nii.gz OutputfileName Subject_FAskeleton.nii.gz

   example – ./singleSubjROI_exe ENIGMA_look_up_table.txt mean_FA_skeleton.nii.gz JHU-WhiteMatter-labels-1mm.nii.gz Subject*ROIout Subject*FA.nii.gz