Distance-adjusted N1 response rates showed no consistent drift across 31–74 minute SEEG sessions

We asked whether the probability that a stimulation pair evoked an expert-checked N1 at distant contacts changed over a clinical single-pulse electrical stimulation session. We used OpenNeuro ds005448 v1.0.0, expert-checked N1 tables, events.tsv and electrodes.tsv for 8 SEEG patients (patients 06–13). Across 350 stimulation pairs and 22,513 pair-contact units at least 10 mm from the pair, 4,040 units had an N1 (17.9%). For each patient, a logistic model of N1 on log distance estimated expected N1 probability; each pair’s residual was the mean observed minus expected N1. Pair time was the median onset of its pulses plus the run start (run label DDHHMM; first-day runs only). We calculated each patient’s Spearman rho between time and residual and a control correlation within shafts. Mean rho was 0.12 (95% CI −0.08 to 0.33; 6 of 8 positive; sign-flip p = 0.20). Within shafts, mean rho was −0.09 (95% CI −0.38 to 0.20; p = 0.53). The second half minus the first half was +3.4 percentage points (95% CI −2.9 to 9.7; p = 0.23). The log-amplitude residual also did not change (rho 0.11, p = 0.31). No consistent drift was found; the small positive trend is better explained by shaft order of stimulation. Limits include 8 patients, sessions of 31–74 minutes, one stimulation per pair, and 63 of 357 pairs using 1 mA for some or all pulses. The data do not exclude an increase of up to 9.7 percentage points. The package includes code, input files with SHA-256 hashes, tables, figure, and reproduction steps.

stereo-EEG · SEEG · single-pulse electrical stimulation · SPES · N1 · cortico-cortical evoked potentials · within-session stability · response drift · OpenNeuro · STReEF