11.5.5. Preproc: FreeSurfer (+ @SUMA_Make_Spec_FS)

Re. recon-all (FS)

This stage describes preprocessing the (T1w) anatomical volume primarily using FreeSurfer (FS). This provides information such as:

  • whole brain segmentation+parcellation (two versions: ‘2000’ and ‘2009’; user’s choice on what to use)

  • tissue maps

  • skull-stripping

  • surface mesh estimation (with parc+seg labels attached)

... and probably some other useful data that I am forgetting. For DTI-related applications, we mainly make use of the parc+seg maps and the surface meshes, created using their recon-all function with their defaults. This documentation was created using the current version of FS, v6.0.

In preparation for running recon-all, at present it seems like there are some important properties for the volume to have, in terms of spatial resolution and matrix dimensions. Please see “fat_proc_align_anat_pair: Align T1w -> T2w” for a description, even if you decide not to use that function to accomplish some of those things.


Disclaimer: while we like using FS and some functions therein, we are not experts in it– all FS-related questions about options or problems should be addressed to the FS gurus themselves. Any feedback on things to do differently would be welcomed and gladly discussed on our end.

The FS folks have provided some useful feedback on questions that have come up related to this work, so thanks to them for that.

Finally, we note that there are some differences between FS and AFNI in grouping FS parc+seg ROI maps into tissue maps, which are described in the next section.

Re. @SUMA_Make_Spec_FS (AFNI-SUMA)

After running recon-all, then the SUMA function @SUMA_Make_Spec_FS in AFNI is used to translate all the volumetric and surface data into formats usable in AFNI (i.e., to NIFTI, GIFTI, etc.). It also translates the FS-generated surface meshes into “standard meshes.” Additionally, as a final step it executes the @SUMA_renumber_FS script that makes tissue-based maps from the FS parc+seg labelling for each of the 2000- and 2009-map versions:

  • *_REN_gm.nii.gz :gray matter

  • *_REN_wmat.nii.gz :white matter

  • *_REN_csf.nii.gz :cerebrospinal fluid

  • *_REN_vent.nii.gz :ventricles and choroid plexus

  • *_REN_othr.nii.gz :optic chiasm, non-WM-hypointens, etc.

  • *_REN_unkn.nii.gz :FS-defined “unknown”, with voxel value >0

The lists for renumbering and grouping can be viewed here for the 2000 and here for the 2009 FS parc+seg maps. The main difference between these and the FS tissue segmentations is that, at present, the ROIs classified as ‘othr’ (Left-vessel, Right-vessel, non-WM-hypointensities, and Optic-Chiasm) are not excluded from the FS gray matter map using mri_binarize --gm.

recon-all and @SUMA_Make_Spec_FS

Proc: Running FS’s recon-all takes a fair amount of time (of order several hours and upward, depending on machine), but the default implementation is pretty straightforward, mainly requiring a T1w volume as input. After running that, the SUMA function @SUMA_Make_Spec_FS in AFNI is also quite direct to run, mainly just needing to know where the FS-generated set of directories are for a subject (though, not the inclusion of the -NIFTI flag, which is recommended). Therefore, the following can be run in succession:

# I/O path, same as above, following earlier steps
set path_P_ss = data_proc/SUBJ_001

# ID of individual subj from "tail" of path, above
set ss_id   = $path_P_ss:t

# FS function
recon-all                                     \
    -all                                      \
    -sd       $path_P_ss                      \
    -subjid   anat_02                         \
    -i        $path_P_ss/anat_01/t1w.nii.gz

# AFNI-SUMA function: convert FS output
@SUMA_Make_Spec_FS                            \
    -NIFTI                                    \
    -fspath $path_P_ss/anat_02                \
    -sid    $ss_id

-> produces a set of subdirectories in the new directory ‘data_proc/SUBJ_001/anat_02/’, as well as ‘data_proc/SUBJ_001/fsaverage/’; the AFNI- and SUMA-compatible files are contained in the subdirectory ‘data_proc/SUBJ_001/anat_02/SUMA/’.

Directory substructure for example data set


Output subdirectories made by FS’s recon-all and SUMA’s @SUMA_Make_Spec_FS: the latter function makes the listed directory “SUMA”, and the former function makes the others.

Directory substructure for example data set


Output files in the SUMA directory made by @SUMA_Make_Spec_FS.