sourcedata folder is optional. It contains all raw, unedited data.
It contains one folder for each subject, and (optionally) at least one session folder (all our clinical cases have at least one session folder). Per the specification –
'BIDS in its current form is designed to harmonize and describe raw (unprocessed or minimally processed due to file format conversion) data... source data... is defined as data before harmonization and/or file format conversion (for example E-Prime event logs or DICOM files).'
As the data here is custom, by lab. Our folders are set up as defined below. This structure is critical: when set up this way, our dcm2bids2.sh script allows the data to be automatically written into a BIDS analyzable form.
Contents: Lab-specific (examples for Benjamin lab)
|participants.tsv||[file] [tsv] [Optional]. This is a two-line, tab-separated participants.tsv file holding the ID of the participant along with relevant clinical, cognitive and demographic data important for clinical fMRI interpretation. This file is collected at participant testing and added to the main participants.tsv file later.|
|participants.json||[file] [json] [Optional]. Paired file for participants.tsv. This exists simply to explain the columns in participants.tsv|
|sub-[x]_ses-[y].dicom.tgz||[file] [tgz] [Optional]. Archive holding all collected dicom files.|
|logfiles||[folder] [Optional]. Benjamin lab convention: All log-files from the stimulus-presentation software at runtime.|
|notes||[folder] [Optional]. Benjamin lab convention: This folder contains multiple text files. One documents any notes on data acquisition, e.g., poor task compliance. Another is a readme written by the dcm2bids2 script, documenting the original scanner names of each sequence and their re-named BIDS conversion name (which runs were discarded is also noted).|
|posttest||[folder] [Optional]. This holds the logfiles from the tasks run after the MRI session.|
Note: Full details available at BIDS website; the full specification is definitive.