brainage/README.md

brainage

get_brainage.sh

This script is used for brain age prediction. Below is a description of each input arguments:

Usage

./get_brainage.sh -input <input_directory> \
                  -output <output_directory> \
                  -BIDS <yes/no> \
                  -CONDOR <yes/no> \
                  -RAW <yes/no> \
                  -RUN_FILE <bash_file_path> \
                  -SAVE_ALL <yes/no> \
                  -CONTAINER_PATH <path/to/containers> \

• input: input directory (can be a folder containing raw or preprocessed T1 MR Image which can be .nii, .nii.gz or .mgz)
• output: output directory where the results will be stored in <subject_name>_<model>.csv
• BIDS: are the input files stored in BIDS format ?
• CONDOR: provides a condor.submit file
• RAW: are the input files raw ?
• RUN_FILE: path to the run_<model>.sh file
• SAVE_ALL: Do you want to save the intermidiate files (for ex: subject_features), if yes then the files will be stored in the folder <subject_name>_<model>
• CONTAINER_PATH: path to the directory containing the container images

'brainageR' folder:

1) Building the container

singularity build brainageR.sif brainageR.def (not working at the moment)

2) Run the get_brainage.sh with the proper files and paths

Example:

./get_brainage.sh -input <input_directory> \
                  -output <output_directory> \
                  -BIDS <yes/no> \
                  -CONDOR <yes/no> \
                  -RAW <yes/no> \
                  -RUN_FILE run_brainageR.sh \
                  -SAVE_ALL <yes/no> \
                  -CONTAINER_PATH <path/to/containers> \

'Enigma' folder:

1) Building the containers

Preprocessing: singularity build --fakeroot cat_preprocessing.simg Singularity_r2042

Prediction: singularity build --fakeroot enigma_stacking_agepred.simg enigma_stacking_pred.def

/!\ It is important to not change the name of the containers

2) Run the get_brainage.sh with the proper files and paths

Example:

./get_brainage.sh -input <input_directory> \
                  -output <output_directory> \
                  -BIDS <yes/no> \
                  -CONDOR <yes/no> \
                  -RAW <yes/no> \
                  -RUN_FILE run_enigma.sh \
                  -SAVE_ALL <yes/no> \
                  -CONTAINER_PATH <path/to/containers> \

'More' folder:

1) Downloading the files

You should first download the following files: Singularity_r2042, BA_predict.recipe and run_more.sh

2) Building the containers

Preprocessing: singularity build --fakeroot cat_preprocessing.simg Singularity_r2042

Prediction: singularity build --fakeroot BA_predict.sif BA_predict.recipe

/!\ It is important to not change the name of the containers

3) Run the appropriate bash script for one subject or several

Examples:

For one subject:

./run_more.sh -input <path/to/input_file> \
              -output <path/to/output_dir> \
              -RAW <yes/no> \
              -SAVE_FILE <yes/no> \
              -CONTAINER_PATH <path/to/container_dir> \
              -MODEL_NAME <model_name>

• input: path to a single T1 weighted MRI file (nii, nii.gz or mgz)
• output: path to the output directory
• raw: is the input file raw or preprocessed ? (default: yes)
• save_file: Do you want to save intermidiate files ? (default: no)
• container_path: path where the container(s) should be stored
• Model_name: name of the trained model that you want to use (default: 4sites.S4_R4_pca.gauss.models)

For several subjects:

./get_brainage.sh -input <input_directory> \
                  -output <output_directory> \
                  -BIDS <yes/no> \
                  -CONDOR <yes/no> \
                  -RAW <yes/no> \
                  -RUN_FILE run_more.sh \
                  -SAVE_ALL <yes/no> \
                  -CONTAINER_PATH <path/to/containers> \
                  -MODEL_NAME <model_name>