BIDS — Brain Imaging Data Structure — started in 2016 as a way to standardize how MRI data is named, organized, and described. Since then it has expanded to EEG, MEG, intracranial EEG, and now fNIRS (BIDS Extension Proposal 30, 'BEP-030', currently in draft).
The pitch is simple: any researcher who can read your dataset's directory tree can run any analysis pipeline. No more 'what does column 4 in this CSV mean?'
BIDS-fNIRS specifies four things: directory naming (sub-001/ses-01/nirs/), required JSON sidecars (the metadata file we publish at /downloads), the SNIRF binary format for raw signals, and validator tooling.
EEGBase exports every session in BEP-030 format automatically. That's why our pre-print can include a multi-clinic registry that didn't have to manually clean up file naming across 412 sites — the platform did it for them at upload time.
If you're a clinician thinking 'this sounds like research-team plumbing, why does it matter to me?' — three reasons. (1) Insurance auditors can verify your protocols ran exactly as documented. (2) Your data is portable to any future tool that reads BIDS. (3) If you ever want to publish from your own data, you don't have to retroactively clean it up.