N1 CCEP peak latency increased with distance in 6 of 8 SEEG patients, but distance did not predict latency in held-out patients

Question: does the N1 peak latency of cortico-cortical evoked potentials increase with distance from the stimulated SEEG pair, and does distance predict latency in a new patient? Data: OpenNeuro ds005448 v1.0.0, patients STREEF06-13, expert-checked N1 peaks from the dataset derivatives (input subset and hashes from AutoSci c8ca046e). Method: Theil-Sen slope and Spearman rho per patient, exact sign-flip tests over patients, distance-adjusted covariates, leave-one-patient-out prediction. Results: median slope 0.25 ms/mm (6 of 8 patients with 95% CI above 0, p = 0.031); within-site rho median 0.34 (7 of 8, p = 0.016); far (40 mm or more) N1s are 13.4 ms later than near (less than 20 mm) N1s. Apparent velocity is 2.1-4.9 m/s (lower bound, Euclidean distance). Distance explains little variance (median R2 0.034). Patient offsets are large (intercept 10-35 ms), so distance does not improve held-out MAE (16.8 ms vs 15.3 ms for a constant). Removal of window-edge peaks does not change the result. Limits: peak latency, not onset; 10-100 ms detection window; Euclidean distance; 1 center; epilepsy patients; no re-measurement on raw signals. BCI use: read-out windows must be calibrated in each patient.

cortico-cortical evoked potentials · SEEG · N1 latency · distance · epilepsy · brain-computer interface