MRIght is a standardized MRI preprocessing and conversion pipeline. It organizes DICOMs, converts them to BIDS format, performs quality control, and runs FreeSurfer recon-all.

This pipeline allows you to go from raw DICOMs directly from the scanner to BIDS-compliant data with FreeSurfer recon-all outputs. It's designed to be used directly from the terminal.

• Exporting MRI images
• Step-by-Step Workflow
  • Install the virtual environment
  • 1. Sort the DICOMs
  • 2. Convert to BIDS
  • 3. Quality Control (QC)
  • 4. Run FreeSurfer recon-all
• License

1. Log in using the usuario predeterminado (no password required).
2. Open the patient navigator.
3. Select the subject you want to export (this automatically selects all images).
   • Tip: You can sort by project name (e.g., sort by “BBS Lab” subjects first).
4. Click Exportar to open the export pop-up.
5. In the pop-up:
   • Choose your hard drive as the Destino using the Examinar button.
   • Under Propiedades, set:
     • Conversión de imagen: mejorada
     • Hacer anónimo como: mantenimiento
     • Tick only the hacer anónimo como box
     • Enter the subject ID (e.g., 4132)

• conda env create -f 0-env_config/linux_environment.yml
  conda activate mright-env

• python 1-sort/Sort_DICOMS.py

  Prompts for: DICOM directory, timepoint folder name (e.g., TP2)

  Output: Organizes images into subfolders by subject and sequence.

• python 2-convert/DICOM_to_BIDS.py

  Prompts for: Path to DICOM directory, timepoint folder name, path to shared BIDS directory (for determining subjects who have already been processed), path to local (temporary) BIDS output directory, path to project heuristic file (i.e., 2-convert/heuristic_general.py).

  Output: Creates a BIDS-compliant dataset in the specified temporary output directory.

• python 2-convert/move_and_merge.py

  Prompts for: Path to local (temporary) BIDS output directory, path to shared BIDS destination directory.

  Output: Moves the BIDS compliant data from a local (temporary) directory to a shared destination directory.

• python 3-bidsqc/DICOMS_BIDS_inventory.py

  Prompts for: Path to shared BIDS directory, path to DICOM directory, path to recon-all directory, path to preprocessed BOLD images (enter ‘.’ if none exist), path to QC output directory, session label (e.g., 02).

  Output: Creates an inventory of which subjects have completed each processing step.

  Note: Run this script before generating GIFs; otherwise, the GIF script will throw an error.

• python 3-bidsqc/anat_animate.py

  Prompts for: Path to shared BIDS directory, path to QC output directory, session label (e.g., 02).

  Output: Creates T1 and T2 GIFs for manual QC inspection.

• python 3-bidsqc/play_anat_gifs.sh

  Behavior: Automatically opens each GIF for manual inspection. After closing a GIF, you’ll be prompted to enter a QC rating and any comments. T1 QC ratings: 1, 2A, 2B, 3, or 4 and T2 QC ratings: 1, 2, 3, 4.

  Output: Creates a .txt file with all QC ratings.

• python 3-bidsqc/rename_runs_qc_anat.py

  Prompts for: Path to shared BIDS folder, path to QC folder, path to project heuristic file, session label (e.g., 02), sequence to rename (e.g., T1).

  Output: Renames T1/T2 files when a subject has two runs and run 2 is better than run 1, or when a run is rated as unusable (i.e., rated as a 4).

  Note: You need to run this script after doing QC, before running recon-all, to ensure the highest quality runs are used in recon-all.

• python 3-bidsqc/rename_runs_qc_others.py

  Prompts for: Path to list of individual files that need to be renamed.

  Output: Renames BIDS files that are not T1 nor T2, from a given list that contains their paths.

  Note: This script is relevant for updating the metadata files such that BIDS standard compliance is maintained.

• bash 4-reconall/recon-all_parallel.sh -i <BIDS_DIR> -o <FREESURFER_SUBJECTS_DIR> -p <N_CORES> [-s <session>] [-l <list_file>]

  Arguments: -i shared BIDS directory, -o recon-all output directory, -p number of cores per job, -s optional session ID (e.g., ses-02), -l optional list file with subject IDs; if not included, the list is automatically generated.

  Example:

  bash 4-reconall/recon-all_parallel.sh -i /pool/guttmann/institut/UB/Superagers/MRI/BIDS -o /pool/guttmann/institut/UB/Superagers/MRI/derivatives/freesurfer-reconall -p 3 -s 02

  Output: Generates recon-all outputs in the specified directory.

MRIght © BBSLab, University of Barcelona.

Please cite appropriately if using this pipeline in publications.