Location-adjusted SEEG peak frequency correlated between hemispheres in 166 patients (Spearman rho = 0.53)

We tested whether 4–13 Hz intracranial SEEG peak frequency is shared across distant regions within a patient. We analyzed archived bipolar spectra for 256 patients in the RAM free-recall datasets (27,974 pairs; 25,545 pairs with a peak). After applying the unchanged peak method and removing the MNI-location effect with a leave-one-subject-out cubic model, median adjusted peak frequencies in the left and right hemispheres correlated across 166 patients (Spearman rho = 0.528; 95% CI 0.384–0.653; two-sided permutation p = 0.0001). Anterior–posterior and inferior–superior comparisons had rho = 0.320 (n = 194) and 0.432 (n = 200), respectively; both had Holm-adjusted p = 0.0005. Adjustment for site, sampling rate, age, sex, and dataset left the primary association similar (rho = 0.532). Aperiodic exponent and peak height also correlated between hemispheres (rho = 0.734 and 0.665). The findings are consistent with a shared between-patient component in peak frequency across distant regions, with a local component also remaining. The sample comprised epilepsy patients aged 18–64 years, using one 60-second task segment. Medication, vigilance, seizure-onset contacts, and recording factors shared across contacts were not controlled and may contribute to the shared component. Location was modeled from MNI coordinates without anatomical labels or matched regions.

SEEG · intracranial EEG · peak frequency · theta · alpha · hemispheric correlation · MNI coordinates · spectral analysis · epilepsy · RAM free recall