Shaft access raised expected SOZ balanced accuracy for a fixed HFO rule in 18 SEEG records with equal source counts

We tested whether source channels on the test electrode shaft change an SOZ label score. SOZ means seizure onset zone. We used two high-frequency oscillation detector counts from 18 public SEEG sleep records. SEEG means stereoelectroencephalography. The sample had 1,285 bipolar channels. A fixed rule copied the nearest source label in log count space. Both source procedures used equal source counts. Both main procedures excluded channels that shared either raw contact with the test channel. We averaged exactly over fair source subsets and distance ties. Mean expected balanced accuracy was 55.07 percent with access to the test shaft. It was 50.41 percent after whole shaft exclusion. The difference was 4.66 percentage points. Its descriptive 95 percent patient bootstrap interval was 1.77 to 8.23 points. The contrast after subtraction of a uniform source control was 3.47 points. All six planned checks had positive mean differences. Two patients supplied 54.6 percent of the main signed difference sum. Sensitivity was low under both procedures. This result concerns the fixed rule and released labels in this cohort. It does not establish clinical benefit or accuracy in new patients. The package preserves source bytes, retrieval pins, code, weights, results, figures, and reproduction instructions. Independent checks passed. Eight output files matched a full offline repeat.

seizure onset zone · high-frequency oscillations · stereoelectroencephalography · source selection · balanced accuracy