Mayo ECoG CCEP N1 probability decreases with distance, but exponential fit superiority is not supported

We tested the exponential distance rule for expert-labelled N1 responses to cortico-cortical evoked potentials (CCEPs) in five Mayo Clinic patients from OpenNeuro ds004774 v1.0.0. The analyzed recording contacts are ECoG contacts, despite an earlier description of this cohort as SEEG. For each patient, we fitted exponential and power-law response probabilities to 10,530 (stimulation pair, recording contact) cells at Euclidean distances of 10–100 mm; 836 cells had an N1 label. The fitted exponential decay constant was positive in all five patients (mean 0.094/mm; patient-bootstrap 95% CI 0.068–0.121/mm), and the CI included the Utrecht value of 0.070/mm. The exact one-sided test gave p = 0.031, but the Holm-adjusted p = 0.0625; with five patients this test cannot reach 0.05 after correction. Exponential fits were better in four of five patients, but the mean log-likelihood difference was +0.004 per cell (95% CI −0.002 to 0.009; p = 0.31), so these data do not show that the exponential law fits better than a power law. Results include the prespecified distance, same-shaft, second-rater, and binned analyses. Limitations are five patients at one centre, mostly ECoG grids with clinically surface-projected positions, Euclidean distance, and one rater for three patients. The source files, hashes, results, and numpy-only reproduction code are included; raw recordings are not included.

cortico-cortical evoked potentials · CCEP · N1 response · exponential distance rule · ECoG · distance decay · epilepsy · OpenNeuro