In ds004080, surface N1 CCEP probability falls with distance with no age effect, while stored detections exclude all 202 depth contacts

We used the stored N1 detections in OpenNeuro ds004080 v1.2.4 (derivatives/av_ccep/ccepData_V2.mat; one center: UMCU, 74 patients, age 4-51 years; clinical single-pulse electrical stimulation). For each recording contact and stimulated pair, we calculated native-space distance to the nearer stimulated contact and the fraction of recordings with a stored N1 detection. For grid and strip (surface) recording contacts (337,139 recordings), median per-patient N1 probability was 0.65 at 0-10 mm, 0.55 at 10-20 mm, 0.18 at 20-30 mm, 0.07 at 30-40 mm, 0.03 at 40-60 mm, and 0.012 beyond 60 mm. This curve has the same shape and almost the same values as the SEEG curve from a different center (ds005448: 0.60 at 0-10 mm and less than 0.02 beyond 40 mm). Age did not predict N1 probability at 10-30 mm (Spearman rho = -0.04, permutation p = 0.73; 74 patients). Twenty-five patients had 202 depth contacts, but no depth contact was in `good_channels` in any run; therefore, the derivative file has no N1 detections for depth recording contacts. Seventeen patients had 101 stimulated pairs with a depth contact. These pairs had 0 stored N1 detections in 6,056 surface recordings, while grid/strip pairs in the same patients had probability 0.53 at 10-20 mm. The zero is a processing exclusion, not a physiological result. Do not use these derivatives to compare depth and surface stimulation or recording; detect N1 again from the raw BrainVision files. Limits are one center, surface (ECoG) recording contacts, the dataset authors’ detector, and a distance definition that can differ from the ds005448 study.

cortico-cortical evoked potentials · N1 · electrical stimulation · electrocorticography · stereo-EEG · response probability · distance · OpenNeuro