Freesurfer
FreeSurfer is a software package for the analysis and visualization of structural neuroimaging data.
It is developed by the Laboratory for Computational Neuroimaging at the Martinos Center for Biomedical Imaging.
FreeSurfer provides full processing streams for structural and functional MRI and includes tools for linear and nonlinear registration, cortical and subcortical segmentation, cortical surface reconstruction, statistical analysis of group morphometry, diffusion MRI, PET analysis, and much more. It is also the structural MRI analysis software of choice for the Human Connectome Project.
For expansive documentation on using and understanding FreeSurfer tools, please visit the FS Wiki.
Running the recon-all command
Recon-all (recon = reconstruction) is a command that takes an anatomical dataset (T1-weighted image) and performs many common preprocessing steps on it, with the goal of converting the 3D brain image (.nii or .dcm) into a 2D surface. It is helpful to imagine the reconstruction as taking a crumpled balloon (T1w) and blowing it up (inflated surface). This inflated surface is particularly helpful when analyzing regions of the cortex where some voxels may contain signal from two separate gyri. It is also helpful when analyzing signal found in the sulci.
To run recon-all on Oscar:
The -i flag points to the T1-weighted anatomical file. If you are running recon-all from the directory containing that file, you can simply put the file name. Otherwise, you must include the full path.
-s describes the subject name, which will be attached to the output files. This can be anything.
-sd points to the subjects directory, where all output will be stored. This is a particularly important option to include when using Oscar, since the default subjects directory is within the Freesurfer module and lacks write privileges. With this flag, you can create your own output directory and it will be named what you specified with the -s flag.
-alltells recon-all to perform all preprocessing steps. For the full list of steps, please refer to Freesurfer's documentation found here: https://surfer.nmr.mgh.harvard.edu/fswiki/recon-all.
This processing takes 6-8 hours to run.
Recon-all Output
These are the output directories you will see, located in the specified -sd path:
label contains text files which hold spatial information on different regions of the brain, along with those regions' atlas annotations.
mri contains many different brain volumes with various levels of preprocessing. Importantly, this is where volumes such as the skull-stripped brain, the subcortical segmentations of the brain, and the brain mask are saved.
scripts contains the log files for the recon-all command. Notably, recon-all.log is where a full history of the workflow can be found.
stats contains files with information on the thickness and volume for each segmentation and cortical parcellation.
surf contains the recon-all generated brain surfaces. This is where both the inflated and non-inflated surfaces are stored.
touch contains files which are created each time a step of recon-all step is executed. These are used by Freesurfer to determine where to begin if the script is paused and restarted.
The temporary directory (tmp) and trash should both be empty at the end of processing.
Output from recon-all can be viewed with Freesurfer's image viewer (freeview). To view a surface from the surf directory, use the -f flag followed by the file name. To view a volume from the mri directory, use the -v flag. Volume files have the .mgh or .mgz extension, which are unique to Freesurfer and stand for Massachusetts General Hospital, and Massachusetts General Zipped, respectively.
Segmentation and Parcellation
Included in recon-all's workflow is segmentation of the subcortical white and grey matter structures (hippocampus, amygdala, caudate, putamen, thalamus, etc) and parcellation of the cortex. Parcellation is conducted with respect to two different atlases, the Desikan-Killiany atlas (/mri/aparc.DKTatlas+aseg.mgz) and the Destrieux atlas (/mri/aparc.a2009s+aseg.mgz). The main difference between the two is that the Destrieux atlas contains more parcellations and is used in more fine-tuned analyses.
Volumes/surfaces can also be opened within the freeview window via the file tab.
The tutorial below is helpful in becoming more familiar with the output of recon-all:
Last updated
