Description

This is a docker image including a recipy to spin up a container automatically running freesurfer 7.2.0

• recon-all
• Hippocampal subfield segmentation on all files in an input folder.

Output ($SUBJECTS_DIR) is directed to a folder on the host machine.

Requirements:

You need to have a working docker environment

• Instructions for Mac, Linux and Windows can be found here: https://docs.docker.com/desktop/
• Docker will take care of the environment. It will also download matlab runtime (might take some time initially - please be patient). Matlab runtime is necessary for the hippocampal subfield segmentation.

You need to have a valid freesurfer license file and copy it to the app/ directory.

• This can be obtained from here: https://surfer.nmr.mgh.harvard.edu/fswiki/License

Usage:

Modify 2 things in process_mris.sh:

- v enter/path/to /auto_recon_all/app/input:/app/input

- v enter/path/to /auto_recon_all/app/output:/app/output

Copy input files to input folder on host machine

• Input is expected to be a (3 tesla) T1 MRI sequence
• Subfolder name becomes the subject name, so
• • inputfolder/foobar/foo.nii results in subjectname foobar
• the pipeline expects one of two input formats:
• • a .nii file in inputfolder/foobar/
• • dicom files via inputfolder/1000*/10000*/10000*/some_dicom_files

Output

All output is standard freesurfer output written to OUTPUT_DIRECTORY

• output of
• • recon-all (with flags -3T -all)
• • segmentHA_t1.sh