Public SEEG datasets contain few bilateral implants; one patient showed no significant cross-hemisphere N1 penalty

We asked if a single SEEG pulse gives fewer early CCEP (N1) responses on contacts in the opposite hemisphere than distance predicts. We used archived N1 tables from ds005448, ds004696 and ds004080 (28 patients, 20029 site-contact rows at 10-80 mm) and the hemisphere labels in the OpenNeuro electrodes.tsv files. A census of all 95 subject files showed that only 2 subjects (ds004696 sub-02 and sub-07) have depth contacts in the two hemispheres. 27 of 28 archived patients had no opposite-hemisphere rows, so the planned test across patients was not possible. In ds004696 sub-07 (27 sites, 111 opposite-hemisphere rows), the opposite-hemisphere N1 rate was 0.153. A log-distance model fit on same-hemisphere rows expected 0.101. At the site level, 3 of 6 sites were above and 3 below the expected rate (mean +0.066, 95% CI -0.070 to +0.212, exact sign-flip p = 0.44). Hemisphere did not improve the leave-one-site-out AUC (0.710 vs 0.713 for distance alone). Thus, in this one patient, distance alone was sufficient to estimate cross-hemisphere N1 reach. A test across patients needs other bilateral SEEG stimulation data. The package has the code (numpy), the census table, the rows used, input SHA-256 values and a figure.

SEEG · cortico-cortical evoked potentials · N1 response · hemisphere · epilepsy · electrical stimulation