BIDSonym - a BIDSapp for the pseudo-anonymization of neuroimaging datasets

The amount of neuroimaging data being shared increased exponentially in recent years. While thisdevelopment introduces prominent advantages concerning open, reproducible and sustainable neu-roimaging, the process of data sharing must ensure the privacy of participant data. A requirement fromboth, Ethics Review Boards and data sharing resources, datasets need to be (pseudo-) anonymized priorto sharing in order to limit participant re-identification. Depending on the dataset at hand, this processcan however become cumbersome and prone to errors. Here we introduce BIDSonym, a tool for auto-mated pseudo-anonymization of neuroimaging datasets. BIDSonym supports multiple de-identificationprocedures and operates on neuroimaging, as well as metadata files. In addition, all metadata infor-mation present in the respective files is gathered and evaluated. Its outputs furthermore allow usersto conduct a more in-depth assessment of potentially sensitive information present in a given dataset.Through its workflow and utilization of the Brain Imaging Data Structure (BIDS), BIDSonym’s appli-cation is reproducible, requires no manual intervention and is agnostic to idiosyncrasies of small andlarge scale datasets.