Linear contact number adjustment reduced the 4–8 Hz/gamma change contrast 10–30 s after marked seizure end in 30 SEEG patients

We tested spectral balance after marked seizure end in public SEEG data. We used OpenNeuro ds004100 release 1.1.3. The main sample had 31 patients and two recordings per patient. We compared pure seizure onset-zone pairs with pure unmarked pairs within contact name groups. The endpoint was the change in the 4–8 Hz/gamma power ratio from pre to post. The post window was 10–30 seconds after marked end. The baseline was 60–40 seconds before marked onset. The unadjusted contrast was 2.94 dB, with a 95% patient bootstrap interval of 1.26–4.69 dB. The planned sign test gave p = 0.001971. It assumes independent patients and sign symmetry under the zero null. Linear contact number adjustment used 30 patients and 69 groups. In those same groups, the estimate changed from 2.62 to −0.06 dB. The adjusted pointwise interval was −1.57 to 1.49 dB. Both absolute bands fell at onset-zone pairs. Only 10 of 70 groups had overlapping contact number ranges between classes. The adjusted estimate depends on a linear model and limited position overlap. It does not establish equivalence or an anatomical cause. The result limits the use of this unadjusted ratio as an onset-zone recovery marker. The package preserves all input spectra, metadata, source versions, code, numerical results, and reproduction instructions. New source retrieval and an independent decoder checked all four epochs in two runs from three fixed patients.

SEEG · seizure onset zone · spectral power · gamma power · postictal · contact number adjustment