Skip to content

FreeSurfer

FreeSurfer is not currently supported natively on Ubuntu 24 (the operating system used on SPICE). To use FreeSurfer, run it through the FreeSurfer and MNE-Python container.

This container includes:

  • FreeSurfer
  • MNE-Python

FreeSurfer license

You need a freesurfer license. If you don't have one, follow this link. Upload it to SPICE and note the path to the license file for use in the next section.

Start the container

Run the following commands from a terminal.

PROJECT_DIR=/data/projects/<project_name>
FS_LICENSE=<path_to_freesurfer_license>
SUBJECTS_DIR=<path_to_SUBJECTS_DIR>

singularity shell \
    --bind $PROJECT_DIR \
    --cleanenv \
    --env FS_LICENSE=$FS_LICENSE \
    --env SUBJECTS_DIR=$SUBJECTS_DIR \
    /scratch/singularityContainers/mne_freesurfer.sif

This mounts your PROJECT_DIR into the container, making all files in that directory available inside the container. You can also specify paths to other directories if needed using the --bind option.

Now you are ready to run freesurfer! 

recon-all -subject <subject id> -i <path_to_t1> -all

FreeSurfer and MNE-Python

If you are performing MEG/EEG source localisation, you will also need MNE-Python to generate the Boundary Element Model (BEM) surfaces required for forward modelling. The container includes MNE-Python for this purpose.

Create the BEM surfaces using:

mne watershed_bem --subject <subject_id> --subjects-dir <path_to_SUBJECTS_DIR>

Optionally, you can create a dense scalp surface. This is useful for visual coregistration, alignment of digitised head points, and marking fiducials. 

mne make_scalp_surfaces \
  --subject <subject_id> \
  --subjects-dir <path_to_SUBJECTS_DIR> \
  --mri <path_to_MRI> \
  --no-decimate

The path supplied to --mri is relative to: 

$SUBJECTS_DIR/<subject_id>/
For example, if the MRI file is located at: 
$SUBJECTS_DIR/<subject_id>/orig/001.mgz

then use: 

--mri orig/001.mgz