Public OpenNeuro coordinates cover only 13 of 160 SEEG SOZ pairs, leaving normative HFA testing underpowered

We tested whether a leave-one-patient-out normative spatial baseline of interictal NREM high-frequency activity (HFA, 80–300 Hz) improves separation of seizure-onset-zone (SOZ) from other stereo-EEG (SEEG) bipolar pairs. We used 728 pair features from 43 patients (160 SOZ pairs) in an earlier internal publication and BIDS electrodes.tsv coordinates attributed to OpenNeuro ds007118, ds007119, and ds007120, version 1.0.0. Coordinates were available for both contacts in 122/728 pairs (17%) and 13/160 SOZ pairs (8%); only seven patients had both SOZ and non-SOZ pairs with coordinates. In this subset, the normative score did not improve mean per-patient AUC (0.398 vs 0.404 raw; difference −0.006, 95% CI −0.017 to 0.000; sign-flip p=1.0; Holm p=1.0) or pooled AUC (0.450 vs 0.487 patient-centered raw; difference −0.037, 95% CI −0.100 to 0.019; bootstrap p=0.214, Holm p=0.428). Sensitivity analyses gave the same conclusion. The test had very little power: most patients had only one or two SOZ pairs with coordinates, and raw HFA AUC in this subset was 0.40, compared with 0.68 across all 43 patients in the source publication. These results do not show that normative maps fail; they show that the available public coordinates cannot test spatial normalization for SEEG depth contacts. Coordinates were data-owner fsaverage projections and were not checked against MRI. The coordinate files came from the OpenNeuro S3 mirror; preserved hashes identify the analyzed files but do not establish that they match an immutable snapshot.

stereo-EEG · SEEG · epilepsy · seizure-onset zone · high-frequency activity · normative mapping · electrode coordinates · OpenNeuro · AUC