Five test pulses did not establish a 4–30 Hz SEEG covariance change between shafts in eight patients under independent pulses with stable means and second moments

We tested noise covariance between electrode shafts after electrical stimulation in eight patients with epilepsy. We used public SEEG data from OpenNeuro ds004696 version 1.0.1. Five early pulses per site set each readout scale. Five separate later pulses supplied the test. The primary sample had 83 sites and 881 readout-site rows. The test used 4–30 Hz Fourier coordinates from 100 to 500 ms after a pulse. Under independent later pulses with stable means and second moments, disjoint pulse pairs estimate squared covariance change without direct-square bias. Mean corrected Q was 0.519; its descriptive 95 percent patient interval was -0.037 to 1.178. The sign-change reference probability was 0.133. The covariance evidence remains uncertain. The direct squared estimate was 2.177. The separate variance ratio was 1.836, with interval 1.389 to 2.457. This exploratory study extends an existing method to a different cohort and spectral representation. The package includes compact data, proof, figures, source pins, fresh source checks, and replay records. It does not test online control.

SEEG · electrical stimulation · noise covariance · Fourier analysis · pulse response · epilepsy