Conversion

Download

1. Download the appropriate files from UT Southwestern’s AIRC website (https://swlxaircweb.swmed.edu/) to /raid/data/shared/incoming/airc/

   Make the zip file name completely lowercase and should match the following: <airc_id>_<sub>_w<1|2|3>/

   However, if there is no study ID such as the participant being a pilot, it can be: <airc_id>/ or <airc_id>_w<1|2|3>/

Convert

Note

Install the following package if you have not run this before:

module load python/3.8.6
pip install --user git+https://github.com/epongpipat/pyHelperKennedyRodrigue.git

1. Load appropriate modules and set parameters for subsequent code

   module load sge
   module load python/3.8.6
   root_dir='/raid/data'
   airc_id=""
   sub=""
   ses=""
   

2. Convert MRI

   code_dir="${root_dir}/shared/software/scripts/eep170030/mri/convert"
   bash ${code_dir}/convert_uber.sh \
   --airc_id ${airc_id} \
   --sub ${sub} \
   --ses ${ses}
   

   The code performs the following:

   1. unzip files and rename the output directory to lowercase

   2. sort the DICOM files into their appropriate directories by series name

   3. convert files from DICOM to NIfTI using dcm2niix

   After the script is done, check the conversion and move bad files (e.g., incomplete acquisitions) into a bad/ directory

QC Parameters

1. Create QC files

   code_dir="${root_dir}/shared/software/scripts/eep170030/mri/qc_mri"
   bash ${code_dir}/qc_uber.sh \
   --airc_id ${airc_id} \
   --sub ${sub}
   

   The code performs the following:

   1. converts dcm2niix bids output from .json to .csv format

   2. combines dcm2niix bids output (.csv) across participants by acquisition

   3. converts fslhd output to .csv format

   4. combines fslhd output (.csv) across participants by acquisition

   5. combines all .csv output by acquisition

2. Create QC report

Note

The following code needs to be run locally. The server does not currently have a compatible pandoc version

code_dir="${root_dir}/shared/software/scripts/eep170030/mri/qc_mri/rmd"
Rscript ${code_dir}/render-rmds.R

Six (6) QC html reports will be updated/saved in ${root_dir}/shared/incoming/qc/derivatives/sub-differences/rmd-html. Please review and any potential outliers in either qualitative or quantitive values, and bring up any potential concerns to the team.

Copy Files

1. Copy and rename files

   Note

   may need to install the following if running for the first time:

   Rscript -e 'devtools::install_github("epongpipat/rHelperKennedyRodrigue")'
   

   code_dir="${root_dir}/shared/software/scripts/eep170030/mri/modality_specific_conversion"
   bash ${code_dir}/modality_specific_conversion/copy_files_uber.sh \
   --airc_id ${airc_id} \
   --sub ${sub}
   

   The code performs the following:

   1. copy files into an original/ directory

   2. rename and merge/split files volumes to match prior convention