Each participant has their own folder inside bids/. The pipeline begins by converting the original input files into BIDS-compliant format before any cleaning is applied. This ensures future steps follow the strict BIDS standard.

​
Example Layout

sub-451559/
├── sub-451559_scans.tsv
└── eeg/
    ├── sub-451559_task-tutorial_eeg.json
    ├── sub-451559_task-tutorial_eeg.vhdr
    ├── sub-451559_task-tutorial_eeg.vmrk
    ├── sub-451559_task-tutorial_eeg.eeg
    ├── sub-451559_task-tutorial_channels.tsv
    ├── sub-451559_task-tutorial_events.json
    ├── sub-451559_task-tutorial_events.tsv
    ├── sub-451559_space-XYZ_coordsystem.json
    └── sub-451559_space-XYZ_electrodes.tsv
These files are a direct conversion from the original inputs.
If you need to save disk space, you can safely delete them after processing, since they can always be regenerated from the raw recordings.

​
Key Files

FilePurpose
*_scans.tsvRecords acquisition date/time for each data file.
*_eeg.vhdr/.vmrk/.eegThe BrainVision EEG file trio (header, markers, raw data). This is the converted raw EEG recording in BIDS format.
*_eeg.jsonMetadata describing acquisition parameters (sampling rate, reference, filters).
*_channels.tsvChannel list with type, name, and status (e.g., marked bad).
*_events.tsv / *_events.jsonEvent timing and event metadata.
*_coordsystem.jsonCoordinate system definition for electrode positions.
*_electrodes.tsvElectrode names and 3D positions.

​
Why BIDS First?

  • Standardization: Once in BIDS, every participant has the same predictable structure.
  • Interoperability: Outputs can be validated with the BIDS validator and read by tools like MNE-BIDS or EEGLAB-BIDS.
  • Traceability: Original input → standardized BIDS copy → cleaned derivatives.
The subject folder ID (e.g. sub-451559) may not match the original filename.
The mapping between recording filenames and subject IDs is tracked in participants.tsv.