Counting each SEEG person once preserved the CCEP N1 latency-distance slope in 17 people from 2 institutions

A pooled analysis of cortico-cortical evoked potential (CCEP) N1 latency-distance slopes treated 20 stereoelectroencephalography (SEEG) patients from three OpenNeuro datasets as independent. An independent comparison of OpenNeuro S3 listings found matching signal objects for three ds004457 and ds004696 subject pairs. Reanalysis of the archived inputs with each person counted once included 17 people from two institutions. The pooled slope remained positive at 0.455 ms/mm (95% CI 0.311 to 0.599); 16 of 17 slopes were positive, and the sign-flip p value was 0.0102 (Holm-adjusted p = 0.020), compared with 0.0022 in the original pool. The center-permutation p value changed from 0.56 to 0.14 (0.18 when grouping by the two institutions), providing no evidence of a center difference at this sample size. For three people with the same recording, two feature pipelines gave slopes differing by as much as 0.81 ms/mm (z = 2.28). Per-patient standard errors therefore do not include pipeline variance, and pooled confidence intervals from mixed pipelines may be too narrow. The analysis used archived pair-level tables and did not re-detect N1 peaks from raw signals. The identity check used S3 ETags and sizes rather than full downloads and applies to the listing state on 2026-10-11.

CCEP · SEEG · N1 latency · latency-distance slope · duplicate records · reanalysis · OpenNeuro