FreeSurfer Group-level Analysis

The scripts in this repository are to run Group level analysis on structural MRI data using FreeSurfer (<https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/GroupAnalysis>). The repository contains the following three scripts: runMrisPreproc.sh, runGLMs.sh, and runClustSims.sh

These scripts were created by graduate student, Gabriela Suarez, and former MiND lab data manager, Jared Burton.

STEP 1: Creating a group file with mris_preproc

• runMrisPreproc.sh
  • Combine all of individual structural images into a single dataset before running group analysis.
  • The data are resampled to the fsaverage template, which is in MNI space.
  • Uses a single command mris_preproc - requires the following arguments
    • --fsgd: An FSGD file
    • --target: A template to resample to (fsaverage)
    • --hemi: Indicate which hemisphere to resample (lh, rh, or both)
    • --cache-in: Specify which smoothed images we want to use in the analysis
  • The runMrisPreproc.sh script allows the mris_prepoc command to be executed flexibly across studies depending on the needs of the user

STEP 2: Fitting the general linear model with mri_glmfit

• runGLMs.sh
  • After running mris_preproc the subjects are concatenated into a single dataset. Now we can fit a general linear model with FreeSurfer’s mri_glmfit command
  • mri_glmfit requires the following arguments
    • --y: The concatenated dataset containing all of the subjects’ structural maps
    • --fsgd: An FSGD file
    • --C: A list of contrasts
    • --surf: The hemisphere of the template to analyze
    • --cortex: A mask to restrict our analysis only to the cortex
    • --glmdir: An output label for the directory containing the results
  • The runGLMs.sh script allows the mri_glmfit command to be executed flexibly across studies depending on the needs of the user

STEP 3: Cluster Correction with mri_glmfit-sim

• runClustSims.sh
  • After running the general linear model with the mri_glmfit command and creating group-level contrast maps, it is necessary to correct for the number of tests that have been run.
  • Cluster correction is conducted using the mri_glmfit-sim command
  • mri_glmfit-sim requires the following arguments
    • --glm-dir: Directory containing output files for glm
    • --cache: Vertex-wise threshold & direction of test (options: pos, neg, abs)
    • --cwp: Cluster-wise p-threshold (always set to .05)
    • --2spaces: Correction for analyzing both hemispheres
  • The runClustSims.sh script allows the mri_glmfit-sim command to be executed flexibly across studies depending on the needs of the user