Noddi analysis part 2 : Noddi computation

  • Note: Documentation for NODDI toolbox is available here: http://mig.cs.ucl.ac.uk/index.php?n=Tutorial.NODDImatlab

Get the toolbox

Run the script

  • Let’s open Matlab on CHEAHA, open a new terminal and type:

module load rc/matlab/R2020a
  • Note: Other matlab version are available on CHEAHA, R2020a is working fine but feel free to change if needed

  • Then type matlab in terminal to launch MATLAB

  • Open the matlab script you previously downloaded.

  • The part to change for your needs is highlighted at the beginning of the script but basically you need to change 3 things: - Where all the toolboxes are - Path to your data - Search for subjects ID

  • Once everything is changed, just start the script and wait!!

  • A Noddi_files folder will be created containing all NODDI files for each subject !

Run the script on slurm (Cheaha)

  • See SLURM section on preprocessing to learn about Job creation and use.

  • Here we use a single-subject version of the Matlab script.

Warning

The Noddi toolbox script needs to be modified for this method to work

  • Two files are therefore needed to make this work:

#!/bin/bash
#SBATCH --partition=medium
#SBATCH --cpus-per-task=20
#SBATCH --mem-per-cpu=12000
#SBATCH --time=15:00:00
#SBATCH --array=0-41
module load rc/matlab/R2020a
cd /data/user/rodolphe/Data/MRST/NODDI/Preprocessed/
readarray -t FILES < <(find . -maxdepth 1 -type d -name 'CBDm7*' -printf '%P\n')
cd /data/user/rodolphe/Scripts/Origin/Szaflarski\ lab/MRST/NODDI/Create_noddi_files/
srun matlab -nodisplay -nodesktop -r \
"MRST_NODDI_single_subject_SLURM('${FILES[$SLURM_ARRAY_TASK_ID]}',20); quit;"