Stimulation-site gains predict N1 amplitude and response count beyond distance in eight SEEG patients

In eight SEEG patients from OpenNeuro ds005448 v1.0.0, we tested whether stimulation sites differ in N1 amplitude after accounting for distance, and whether site amplitude gains track N1 response counts. For each patient, a distance model explained a median held-out R² of 0.34. Adding a shrunk stimulation-site offset raised this to 0.45 (median gain +0.082, range +0.061 to +0.264; all eight positive; sign-flip p = 0.009). The gains exceeded all 100 distance-bin site-label shuffles in every patient. Site offsets had median split-half correlation 0.65. Across independently split rows, sites with greater N1 propensity than predicted by distance also had larger amplitude offsets (median Spearman 0.41, range 0.18 to 0.78; all eight positive; p = 0.008). Amplitude declined with distance (median log10 slope -0.77, all eight negative; p = 0.008). The four patients stimulated only at 2 mA showed the same pattern descriptively. These results support representing each site with a gain that relates to both response reach and amplitude. The analysis includes eight patients from one center; amplitudes were measured only for expert-accepted N1s, contact identity was unavailable, current was unknown for four patients, and the measure was the peak of an averaged response. The study did not test prediction in new patients.

stereo-EEG · SEEG · electrical stimulation · N1 response · response amplitude · distance model · site gain · epilepsy