Dependencies

• Freesurfer
• MATLAB

Directory Structure

There is a strict expected directory structure. It is depicted below.

subjects/: The folders in subjects/ are created by recon-all. You should not create them manually, or recon-all will give an error that it was previously run.

studies/: You create all the folders in studies/.

1. There should be one folder for each of your studies (experiments)
2. Within study_1/:

• analysis/ contains all your codes for running the analysis. This repository is my analysis directory.
• rawdata/ is a record of the files that come directly from the scanner.
• The run_01_unpack.sh script converts and reorganizes the data from the rawdata/ folder to unpackdata/ and into either 3danat/ or bold/ depending on whether it is EPI or MPRAGE data.

Pre-process Functions

run_01_unpackdata.sh

• Calls dcm2niix
• Must be preformed locally because the rawdata has face and MARCC is not HIPAA compliant
• The script may need to be modified depending on the run orders

run_02_deface.sh

• Calls mri_deface
• Removes the face from the MPRAGE
• Inputs are the subject name and folder name of the anatomical run
• Must be done locally for same reason as above
• FS mri_deface has been deprecated, so if there is an error, there is not good support to solve the problem. pydeface may be a good alternative although FSL is a dependency. See run_02_pydeface.sh for an example with
• Example call: source run_02_deface.sh s001 002

run_03_reconall.sh

• Calls recon-all
• Performs the cortical reconstruction on the defaced MRPAGE
• Inputs are again the subject name and the folder name of the anatomical run
• Must be done on MARCC
• Takes approximately 16 hours
• Example call: sbatch run_03_reconall.sh s001 002

run_04_preproc.sh

• Calls preproc-sess
• Runs the processing for all functional runs
• Input is the subject name
• Performed on MARCC
• Takes about an hour
• Example call: sbatch run_04_preproc.sh s001

Single Subject Analysis

run_00_mksess_loc.sh and run_00_mksess_main.sh

• Calls mkanalysis-sess and mkcontrast-sess
• These call freesurfer functions that define how the analysis will be conducted. They only have to be one one for the entire project.
• Example call: source run_00_mkses_main.sh s001 lh
• Should be done on MARCC

run_05_glm.sh

• Calls selxavg3-sess
• Finds the beta values for the specified analysis
• Must be called separately for all analyses (and hemispheres)
• Inputs are the subject name, analysis name, and -run-wise or nothing
• run-wise is called for the maintask only
• Done on MARCC
• Takes about 15-30 minutes
• Example call: sbatch run_05_glm.sh s001 maintask-surface-lh-sm0 -run-wise
• Example call: sbatch run_05_glm.sh s001 psts-surface-lh-sm5

run_06_MNImain.sh and swurm_06_MNImain.sh

• Calls mri_surf2surf
• Converts the maintask beta values from native space to MNI space
• The swurm scrip submits many batch calls of run_06_MNImain.sh
• Inputs are the subject name and the hemisphere
• Must be called for both the left and right hemispheres
• Done on MARCC
• Takes about 1 minute
• Example call: source swurm_06_MNImain.sh s001 lh

run_07_MNIloc.sh and swurm_06_MNIloc.sh

• Calls mri_surf2surf
• Same as above but with the localizer name
• Must be done for all localizers and hemispheres
• Done on MARCC
• Takes about 1 minute
• Example call: source swurm_07_MNIloc.sh s001 lh FBO

Prior to the next to steps:

• On the MARCC login node: interact -X -t 2:00:00
• module load matlab
• matlab -- A MATLAB GUI will open

run_08_probROIs.m

• Run in MATLAB and uses MRIread and MRIwrite
• Defines the ROIs using parcellations
• Input is the subject name
• Call in the MATLAB command line
• Example call: run_08_probROIs('s001')

run_09_reorgdata.m

• Run in MATLAB and uses MRIread
• Selects the data within each ROI and saves it in a format for further analysis
• Input is the subject name
• Call in the MATLAB command line
• Example call: run_09_reorgdata('s001')

Checking Functions

check_surf.sh

• Uses freeview
• Input is the subject number
• Done after run_03_reconall.sh
• Done locally
• Example call: source check_surf.sh s001

check_registration.sh

• Uses tkregisterfv
• Input is the subject number
• Done after run_04_preproc.sh
• Done locally
• Example call: source check_registration.sh s001

check_motion.sh

• Uses plot-twf-sess
• Input is the subject number
• Done after run_04_preproc.sh
• Done on MARCC in an interactive session, but the png file that is created in unpackdata/${sid}/bold/ will not open
• Example call: source check_motion.sh s001

check_ROIs.sh

• Uses tksurferfv
• Inputs are the subject number, hemisphere, ROI, and contrast
• Done after run_08_probROIs.sh and after the data has been transferred from MARCC to your local machine
• Done locally
• Example call: source check_ROIs.sh s001 lh biomotion motion_translation

Licensed with MIT License