​
The derivatives/ Folder

The derivatives/ folder follows the BIDS derivatives standard, where processed data are kept separate from raw-like files. Each pipeline version creates its own subfolder, ensuring that outputs remain traceable to the exact software and parameters used.

​
Example Layout

bids/
└── derivatives/
    └── autoclean-v2.2.1/
        ├── dataset_description.json
        ├── metadata/          (pipeline-specific metadata)
        ├── intermediate/      (temporary / intermediate outputs)
        ├── logs/              (detailed run logs)
        └── sub-451559/        (subject-level cleaned outputs)

​
Key Components

ItemPurpose
autoclean-v2.2.1/Pipeline-specific derivatives folder, versioned for traceability.
dataset_description.jsonRequired by BIDS derivatives. Documents the pipeline, version, and references.
metadata/Stores additional pipeline metadata (e.g., preprocessing settings, QC parameters).
intermediate/Temporary or step-wise files (not always needed for downstream analyses, but useful for debugging).
logs/Processing logs at the derivative level, providing more detail than the top-level workspace logs.
sub-<ID>/Subject-level cleaned outputs, in BIDS derivative format (e.g., artifact-corrected EEG, ICA solutions, epoch sets).

​
Why Derivatives Matter

  • Traceability: Every derivative is tied to a pipeline version (autoclean-v2.2.1), making analyses reproducible.
  • Separation of concerns: Raw-like BIDS inputs are preserved, while processed data live here.
  • Compatibility: Other BIDS-aware tools know to look in derivatives/ for cleaned or secondary outputs.
  • Debugging: metadata/, logs/, and intermediate/ help you track down how results were produced.
If disk space is limited, you can archive or delete the intermediate/ folder after successful runs. The final cleaned outputs will remain in sub-<ID>/.