In 8 SEEG patients, stimulation-pair orientation did not predict other-shaft N1 occurrence or amplitude beyond distance

We tested if the axis of a bipolar SEEG stimulation pair changes which contacts on other shafts show an expert-checked N1 CCEP when distance is known. A dipole model predicts up to 2 times more field on-axis than at 90 degrees. We used 8 patients (STREEF06–13) from OpenNeuro ds005448 v1.0.0, with 334 pairs, 30,815 pair-contact rows, and 2,316 N1s. The plan was fixed and hashed before outcomes were computed. In each patient, we fit y ~ log d + |cos θ| and compared the angle coefficient with a null that exchanges pair axes between shafts (1,000 permutations). The mean coefficient was +0.06 (4 of 8 positive; exact sign-flip p = 0.88; 95% interval of odds ratio 0.46–2.44). Two patients were significant alone, in opposite directions. The angle changed held-out AUC by −0.001 to +0.007 (distance-only AUC 0.72–0.95). A fixed dipole-field term was worse than distance in 6 of 8 patients. The angle did not change N1 amplitude (p = 0.41). The permutation-null mean was not zero (−0.29 to +1.05), so a test without a geometry-aware null can show false orientation effects. This result agrees with a previous study using a different contact set and null. For bipolar single-pulse SEEG stimulation, pair orientation was not a useful control for targeting responses at other-shaft sites. Effects smaller than an odds ratio of about 2 are not excluded. The package includes the numpy code, hashed plan, input files with provenance, per-patient results, permutation nulls, and two figures.

SEEG · CCEP · N1 · brain stimulation · stimulation orientation · electrophysiology · epilepsy · permutation test · reproduction