In 32 HUP SEEG patients, 60-Hz line-noise excess was lower on SOZ pairs and did not explain high-frequency separation

We tested a pre-registered question: does relative 60-Hz line-noise excess confound the separation of seizure-onset-zone (SOZ) and non-SOZ pairs by interictal 80–150 Hz high-frequency activity (HFA)? We analyzed the first 60 seconds of the presurgery interictal SEEG run in 32 patients from the HUP iEEG dataset (OpenNeuro ds004100 v1.1.3), using 3,325 within-shaft bipolar pairs and clinical SOZ labels. The primary line-noise excess index (LNX) was the log10 ratio of mean 59.5–60.5 Hz power to mean 55–58 and 62–65 Hz power, computed without notches. Its mean per-patient SOZ-versus-non-SOZ AUC was 0.408 (95% CI 0.351–0.471; Holm-adjusted p=0.0051); 9/32 patients had AUC above 0.5. Thus SOZ pairs had lower relative line-noise excess, opposite to the pre-registered confound criterion. Absolute 60-Hz power was higher on SOZ pairs (AUC 0.653), consistent with a higher broadband power floor and not a clean artifact index. LNX and HFA were negatively correlated in all 32 patients (mean Spearman rho −0.42; 95% CI −0.47 to −0.37). HFA AUC was 0.664, and after within-patient linear adjustment for LNX it was 0.650 (95% CI 0.583–0.710; Holm-adjusted p=0.0002); the difference was 0.014 (95% CI −0.010 to 0.034; p=0.25). The HFA separation therefore remained after adjustment in this dataset and analysis. This result is limited to one 60-second clip per patient and HUP SEEG. SOZ labels were clinical, impedance data were unavailable, and LNX is an indirect proxy; the 60-Hz and neighboring bands may contain brain activity, and the adjustment was linear.

stereo-EEG · seizure onset zone · line noise · high-frequency activity · interictal · power spectral density · artifact · HUP iEEG