BIDS_coding
Converting a lab’s raw recordings into the BIDS standard is mostly the same job every time — rename and reorganize files, strip identifying information, validate the result — but the specifics (which fields count as PHI, what the source layout looks like) differ dataset to dataset. This toolkit handles that with a JSON-driven configuration layer instead of a new one-off script per dataset: point it at a config describing the source layout and what needs to change, and the same converter handles a new dataset without new code.
What it does
- Converts intracranial EEG datasets (with multi-echo MRI support) into the current BIDS specification for intracranial electroencephalography.
- Removes PHI as part of the conversion — not a separate pass — so there’s no window where an intermediate file is BIDS-formatted but still identifiable, which matters for NWB/DANDI compliance downstream.
- Runs per-file or iterates a whole dataset from the command line, with a shell-script wrapper for repeated/batch runs across subjects.
- Builds on the same conversion approach as Data2Bids, extended for the lab’s intracranial-specific needs rather than the general fMRI case that project targets.