In 31 HUP patients, distance correction did not change SOZ power AUC; a power–distance association was not detected

A preregistered analysis tested whether bipolar interictal log power increases with inter-contact distance in an independent center and whether distance correction changes seizure-onset-zone (SOZ) separation by 1–40 Hz power. We analyzed archived features from 31 HUP patients (3,229 adjacent bipolar pairs; 30 patients had SOZ and non-SOZ pairs) and fsaverage electrode coordinates from OpenNeuro ds004100 version 1.1.3. The median pair distance was 5.55 mm (5th–95th percentile: 4.17–8.20 mm). The mean within-patient Spearman correlation between 1–40 Hz power and distance was +0.039 (95% CI −0.029 to +0.106; 20/31 positive; Holm-adjusted p=0.56). This was a non-detection with low power: the interval includes zero and the RAM estimates (+0.05 to +0.09). After per-patient regression of power on log2 distance, mean SOZ AUC was 0.672 (AUC−0.5=+0.172; 95% CI +0.103 to +0.236; Holm p=0.00045). The corrected-minus-raw AUC change was −0.0005 (95% CI −0.012 to +0.013; p=0.95). SOZ pairs did not differ in distance (AUC=0.474). Thus, in this dataset, distance correction did not change the interictal SOZ power marker. In an uncorrected secondary analysis, the 30–45 Hz aperiodic exponent increased with distance (mean rho +0.085; p=0.0038), which needs confirmation. Distances came from fsaverage template-warped coordinates and are not physical contact spacing; the range was narrow, one 120-s interictal clip and clinical SOZ labels were used per patient, upstream features were not recomputed from raw EDF, and the sample came from one center. The H1 result does not establish absence of an effect.

stereo-EEG · interictal · seizure onset zone · bipolar power · electrode distance · epilepsy · AUC · HUP