Consider "study" BIDS dataset type for overall project

[yarikoptic]

Looked at https://github.com/s-ccs/LSLAutoBIDS/blob/main/docs/data_organization.md and thought of hopefully soon to be merged

• [ENH] Add "study" DatasetType to organize a collection of source and derivative datasets bids-standard/bids-specification#1972

and also see bids-standard/bids-specification#1861 (comment) for examples . So thought that it would be awesome if you considered making an entire "project" to become a BIDS "study" dataset! ;)

[behinger]

hey! thanks for the suggestion.

I thought I understood that sourcedata is for the raw - potentially non-bids things; and /sub-001 is a proper place to place the files (there was a slight inconsistency in the documentation which I just fixed).

Thus where would you put non-bids "sourcedata", and where the bidsified "raw" data? is that already a derivative for you?

[yarikoptic]

quick one:

• sourcedata/raw - could be your BIDS raw dataset under your "study" dataset (like in dirty example here https://github.com/OpenNeuroStudies/OpenNeuroStudies/tree/main/study-ds000001) or under your derivative dataset e.g. https://github.com/OpenNeuroDerivatives/ds000001-mriqc/tree/c96e4ee137036cda42497c0928cb252f3e1009d8
• the rest of sourcedata/ is generally for any source data which could be non-bids or even bids if it is used as source for this dataset (e,g, entire study if study)

overall, you are not wrong in above. My point was to potentially rework "custom" layout you have/promote:

data
├── bids                  # Converted BIDS data
  ├── projectname1
  ├── projectname2                
├── project_stimulus      # Experimental/Behavioral files
  ├── projectname1
  ├── projectname2          
├── projects 
  ├── projectname1        # Recorded Raw data
  ├── projectname2 

as a proper "bids study" layout. Somewhat relates to similar ones in

• Add DatasetType="project" and rework existing "layout" example into a proper BIDS dataset bids-standard/bids-specification#1861 (comment)
• ongoing current work/discussion in [ENH]: Add BIDS study layout nipoppy/nipoppy#687

[behinger]

okay I understand better. Note that the outcome is very different

└── bids
  └──projectname/
    └── code
        └── log files
    
    └── sub-<label-sub>
        └── ses-<label-ses>
            └── datatype (eg: eeg)
                └── converted BIDS files
                    ├── sub-<label-sub>_ses-<label-ses>_task-Duration_run-001_eeg.vhdr
                    ├── sub-001_ses-001_task-Duration_run-001_eeg.vmrk
                    ├── sub-001_ses-001_task-Duration_run-001_eeg.eeg
                    .........
            └── beh
                └──behavioral files
            └── misc
                └── experimental files (This needs to stored in zip format)
    └── sourcedata
        └── raw xdf files
    └── dataset_description.json
    └── participants.tsv
    └── participants.json
    └── README.md

which should be a proper BIDS dataset.

The point of aboves folder structure is that in our case, projects_stimulus is a netdrive on a different computer with the same structure as the projects folder of the current computer. In order to ensure a proper datastructure, we opted to have a bids folder, which could (by configuraiton) be anywhere else on the computer as well