MNE BIDS MEG Anonymization
Introduction to Import/Upload Process
Datasets must be organized prior to anonymization.
Several organizational options will be provided to assist in recognizing and linking the datasets.
There are a variety of options ranging from extensive renaming of files to manipulating search expressions (choose the option most appropriate for your lab) Prior to anonymization - a "master_list" will be created linking the appropriate input information.
#This installs mne_docker / mne-bids / mne-bids-pipeline git clone git@tako:mne_singularity cd mne_singularity sudo singularity build singularity_build.sif singularity_build.def
Link freesurfer or build from singularity:
ln -s PATH_TO_FREESURFER ./freesurfer
Create Head Surface to view anonymization
Data Curating before anonymization
This data curation will be important to generate a Master_List.csv file that will be used to perform the anonymization.
The Master_List.csv will link the subject MRI, subject MEG Rest Dataset, MEG EmptyRoom Dataset, coreg transform.
The Master_List.csv IS NOT ANONYMIZED, so do not share with the NIH or other outside groups.
After the Master_List.csv is generated, Step 2 will validate the Master_List and return errors/warnings when data cannot be found.
If a few errors are generated, these can be fixed directly on the Master_List.csv. If most/all of the datasets throw errors, the data curation may need to be changed.
This option requires the most space/manual labor, but is easy to troubleshoot
Copy all data into a single folder using the following format:
All MRIs must be of .nii or .mgz
All MEG datasets must be single files or directories: .ds, .fif, .sqd, 4D-dataset (?)
MRIs and MEG datasets must have the same prefix. EG - Subj1.ds Subj1.nii OR LastName_FirstName.fif LastName_FirstName.nii
Create a link file (csv format) that links the MEG subjectID to the MRI subjectID
Provide the top level MEG path
Provide the top level MRI path
Manually generate the Master_List.csv by hand using the template csv.
Troubleshooting can be tedious due to typing errors in the path.
This option requires the most knowledge to troubleshoot string parsing, but is the most flexible
Using a dictionary enter the search strings to extract the meg_path, mri_path, subjid etc: