In OpenNeuro ds006107 (166 subjects), 0 of 340 contact groups have a collinear depth-shaft geometry and 0 channels have type SEEG, thus this dataset supports only ECoG work

We assessed whether OpenNeuro ds006107, an intracranial EEG dataset with seizure-onset-zone, resection, and ILAE outcome labels, contains electrode groups with a depth-shaft geometry or channels labeled SEEG. We used 166 electrodes.tsv files, 164 readable channels.tsv files (sub-27 and sub-163 had no readable channels.tsv), and participants.tsv; no EDF signal files were used. Contacts were grouped by the letters before the trailing number in their channel names. A group met the prespecified depth-shaft signature if it had at least 4 contacts, its first principal component explained at least 0.99 of coordinate variance, and its median nearest-contact distance was 2–7 mm. None of 340 groups met all three criteria. The maximum first-component fraction was 0.958 and the median was 0.662. Channel types were ECOG and n/a; no channel was labeled SEEG. The cohort included 140 Derivation and 26 Validation subjects. ILAE outcome classes 1–5 numbered 114, 3, 19, 16, and 14, respectively. Groups follow amplifier banks rather than necessarily individual electrodes, so a short depth shaft numbered within a grid bank could be missed by the geometry test. The channel-type audit provides an independent second check. Electrode coordinate space was not checked. The source was the public OpenNeuro S3 bucket listing observed on 2026-10-13; the exact dataset release is not established. These results support using ds006107 for ECoG sleep, SOZ, and outcome studies, not for SEEG studies.

intracranial EEG · electrode geometry · ECoG · SEEG · OpenNeuro · epilepsy · dataset audit