This Github repo consists of codes used in my thesis project: Testing differences in brain structure and functional connectivity between children with and without a family history of Alcohol Use Disorder, data from the Adolescents Brain Cognitive Development (ABCD) study.

• Step 1: Identify subjects with and without a family history of Alcohol Use Disorder
• Step 2: Download the data and organize in BIDS format
• Step 3: Run MRIQC
• Step 4: Run BRAINSTools
• Step 5: Run Freesurfer
  • QC with Quoala-T
• Step 6: Run fMRIPrep and Resting State analysis

Download data and organize in BIDS format

Instruction here: https://github.com/DCAN-Labs/abcd-dicom2bids

To run their scripts on our machines, make sure you follow what is listed in the file below /Shared/tientong_scratch/abcd/code/00abcd2bids_convert/abcd2bids_github/notes_abcd2bids.txt

Also, because I only care about rs fMRI and sMRI, I went through all of the scripts and config files in my local abcd-dicom2bids repo and delete everything related to task fMRI or validating BIDS (running BIDS validation needs docker, which can't be run on IFT machines). Below is an example of how to run this script

python <path to script>/abcd2bids.py \
 /Shared/pinc/sharedopt/apps/fsl/Linux/x86_64/5.0.8 \
 /Shared/tientong_scratch/abcd/code/00abcd2bids_convert/R2016b/v91 \
 --username tientong --password <your password> 2>&1 | tee abcd2bids_log.txt 

What I would recommend is submitting this as job on Argon

# run this on argon
qsub /Shared/tientong_scratch/abcd/code/run_abcd2bids.sh 

Move data

mv /Shared/tientong_scratch/abcd/code/00abcd2bids_convert/abcd2bids_github/data/* /Shared/tientong_scratch/abcd/rawdata

Run BIDS Validator

http://bids-standard.github.io/bids-validator/

Some file wasn't associated with json files, e.g., /Shared/tientong_scratch/abcd/rawdata/sub-NDARINV12F1NWCG/ses-baselineYear1Arm1/fmap/vol0002.nii.gz

Right now not sure what happened with the abcd2bids script, move all of those files to /Shared/tientong_scratch/abcd/nojson with the code in the cell below:

for i in /Shared/tientong_scratch/abcd/rawdata/[a-zA-Z0-9]*/[a-zA-Z0-9]*/fmap ; do 
  sub=(`echo $i | awk '{gsub("/"," "); print $5}' | awk '{gsub("-"," "); print $2}'`)
  ses=(`echo $i | awk '{gsub("/"," "); print $6}' | awk '{gsub("-"," "); print $2}'`)
  [ -f /Shared/tientong_scratch/abcd/rawdata/sub-${sub}/ses-${ses}/fmap/vol*2*.nii.gz ] && \
  mkdir -p /Shared/tientong_scratch/abcd/nojson/sub-${sub}/ses-${ses}/fmap && \
  mv /Shared/tientong_scratch/abcd/rawdata/sub-${sub}/ses-${ses}/fmap/vol*.nii.gz \
  /Shared/tientong_scratch/abcd/nojson/sub-${sub}/ses-${ses}/fmap
done

To run a different batch of participants

Download the ABCD Fasttrack QC Instrument file (https://nda.nih.gov/) with the list of participants you're interested in. You can then make changes to this file (e.g., exclude task fMRI, DTI) by running getStructandRS.R

then, copy the output to

/spreadsheets/abcd_fastqc01.txt

Multiple images of the same imaging module of the same session

Have to pick the run that is better. How? running MRIQC

Run MRIQC

MRIQC Subject Level

Create a template file (currently put on argon

/Users/tientong/job_scripts/mriqc/abcd/mriqc_TEMPLATE.job

Content of the template file pasted below

#!/bin/sh
#$ -pe smp 16
#$ -q PINC,CCOM,UI
#$ -m e
#$ -M [email protected]
#$ -o /Users/tientong/logs/abcd/mriqc/out
#$ -e /Users/tientong/logs/abcd/mriqc/err
OMP_NUM_THREADS=8
singularity run -H /Users/tientong/singularity_home \
  /Users/tientong/mriqc_0.14.2.sif \
     /Shared/tientong_scratch/abcd/rawdata \
     /Shared/tientong_scratch/abcd/derivatives/mriqc \
     participant --participant_label SUBJECT \
     --no-sub --verbose-reports --write-graph \
     -w /Shared/tientong_scratch/abcd/derivatives/mriqc \
     --n_procs 8 --mem_gb 36000

Create a list of available subjects

outputFile=/Shared/tientong_scratch/abcd/log/subjectlist/200129list.txt
for i in /Shared/tientong_scratch/abcd/rawdata/[a-zA-Z0-9]*/[a-zA-Z0-9]*/anat ; do 
  info=(`echo $i | awk '{gsub("/"," "); print $5}'`)
  sub=(`echo $info | awk '{gsub("-"," "); print $2}'`)
  echo $sub
done > ${outputFile}

Then run this on argon to create individual subjects mriqc job script, then submit the jobs

for sub in $(cat /Shared/tientong_scratch/abcd/log/subjectlist/200129list.txt | tr '\n' ' ') ; do
    sed -e "s|SUBJECT|${sub}|" /Users/tientong/job_scripts/mriqc/abcd/mriqc_TEMPLATE.job > \
    /Users/tientong/job_scripts/mriqc/abcd/mriqc_sub-${sub}.job
done

for sub in $(cat /Shared/tientong_scratch/abcd/log/subjectlist/200129list.txt | tr '\n' ' ') ; do
    qsub /Users/tientong/job_scripts/mriqc/abcd/mriqc_sub-${sub}.job
done 

After subject level mriqc

Jan 30 2020 Completed running mriqc all subjects in the list below

/Shared/tientong_scratch/abcd/log/subjectlist/200129list.txt

EXCEPT
sub-NDARINV93KR583V - T2w
sub-NDARINVA01DRNBZ - resting state
sub-NDARINVCVK8YHVR - T2w
sub-NDARINVJXT9LVJP - T2w
sub-NDARINVRT0YKWJM - T2w
sub-NDARINVZL6CFV5G - T2w

Detailed logs are at /Shared/tientong_scratch/abcd/derivatives/mriqc/logs

MRIQC group level

# after finish subject level analysis, run the group analysis (concatnate subjects' files)

singularity run \
   -H /Users/tientong/singularity_home \
   /Users/tientong/mriqc_0.14.2.sif \
      /Shared/tientong_scratch/abcd/rawdata \
      /Shared/tientong_scratch/abcd/derivatives/mriqc \
      group --no-sub --verbose-reports --write-graph \
      -w /Shared/tientong_scratch/abcd/derivatives/mriqc

Once the group level MRIQC is completed, move all files to

/Shared/tientong_scratch/abcd/derivatives/mriqc/200129list_group

Choosing the best T1 or T2

Code: baw00ChooseBestStruct.R