Five OpenNeuro Naming Datasets Were ECoG-Only, With Four Sharing Most Patients

A BIDS channel-type census of five large OpenNeuro naming datasets found 0 SEEG sessions across 1,798 sessions and 564 subject folders; every subject was classified ECOG_only. Ages ranged from 4 to 49 years (median 14). Four datasets reused mostly the same patients: pairwise age-and-sex matches among shared IDs ranged from 102/102 to 119/119. In the fifth dataset, 6–7 IDs were shared with each of the other four but none matched on age and sex, so overlap is unknown. A sixth candidate dataset, ds006254, could not be assessed because all 34 channel tables and its participant table returned S3 AccessDenied. Thus these five naming datasets do not provide SEEG naming samples; combining the four overlapping ECoG datasets risks repeated measures from one patient pool. Channel types reflect curator-entered BIDS labels; signals, electrode coordinates, and implant images were not inspected. Analyses used the live OpenNeuro S3 mirror on 2026-10-11; recorded snapshot tags are references, not verified exact releases. File listings and SHA-256 hashes preserve the sampled state.

intracranial EEG · ECoG · SEEG · OpenNeuro · BIDS · naming · channel types · subject overlap · language mapping