Surface Reconstruction (recon-surf)

The surface pipeline (recon-surf): it reconstructs the cortical surfaces from the whole-brain segmentation, maps the cortical parcellation onto them, and computes cortical thickness and other measures, as FreeSurfer’s recon-all does, but much faster.

What it computes

• The white matter and pial surfaces of both hemispheres, and inflated and spherical surfaces.

• Cortical thickness, area, volume and curvature for every surface point.

• The cortical parcellation mapped onto the surfaces, with statistics per region.

• A registration to FreeSurfer’s fsaverage (sphere.reg), for comparing thickness maps across subjects.

• Segmentations and statistics refined with the surfaces, which replace those of the segmentation modules.

• All files that FreeSurfer’s downstream tools expect, under FreeSurfer’s names (see FreeSurfer downstream modules).

The output files are listed in the output files overview.

What it needs

• The outputs of the whole-brain segmentation and of the corpus callosum module, which the same run can compute.

• A FreeSurfer license, because it uses FreeSurfer tools.

• Voxel sizes between 0.7 mm and 1 mm are supported (smaller voxels are experimental); below 1 mm, the pipeline runs in its high-resolution mode.

The surface reconstruction takes about 20 to 40 minutes, because it processes both hemispheres in parallel by default. With --threads 1, it processes them one after the other and takes considerably longer.

Options

The surface reconstruction runs by default, after the segmentation. The most important options of run_fastsurfer.sh for it:

• --seg_only: skip the surface reconstruction.

• --surf_only: run only the surface reconstruction, on an existing segmentation.

• --3T: use the 3T atlas for the Talairach registration, for better eTIV estimates of 3T images.

• --threads_surf <n>: the number of threads for this part.

• --no_surfreg: skip the registration to fsaverage (saves time, but is needed for cross-subject analyses of thickness maps; recon-surfreg.sh adds it later).

• --fsaparc: also compute FreeSurfer’s own cortical parcellation, in addition to the mapped one.

Usage:

run_fastsurfer.sh --surf_only --sd <subjects_dir> --sid <subject_id> \
    --fs_license <freesurfer_license_path>

All options are described in the run_fastsurfer.sh reference. To run recon-surf.sh directly, see recon-surf in the command reference.

References

If you use the surface reconstruction in your research, please cite:

• Henschel L, Conjeti S, Estrada S, Diers K, Fischl B, Reuter M. FastSurfer - A fast and accurate deep learning based neuroimaging pipeline. NeuroImage 219 (2020), 117012. doi:10.1016/j.neuroimage.2020.117012

The surface reconstruction is largely based on FreeSurfer, see how to cite FreeSurfer.