SOZ shaft count did not predict interictal SOZ AUC in 59–60 awake SEEG patients

We tested whether interictal stereo-electroencephalography (SEEG) features identify the clinically labeled seizure-onset zone (SOZ) less well when SOZ pairs span more implanted shafts. Using archived bipolar-pair feature tables from awake HUP (OpenNeuro ds004100) and ds003876 cohorts, we calculated a per-patient Mann–Whitney AUC for SOZ versus non-SOZ pairs and tested its Spearman correlation with the number of shafts containing SOZ pairs. Among 59 patients for 80–150 Hz power and 60 for spike rate and 1–40 Hz power, correlations were respectively ρ = −0.09 (95% bootstrap CI −0.36 to 0.19; permutation p = 0.48; Holm-adjusted p = 0.97), ρ = 0.00 (CI −0.27 to 0.25; p = 1.00; Holm p = 1.00), and ρ = −0.15 (CI −0.40 to 0.11; p = 0.24; Holm p = 0.73). The secondary NREM cohort (ds007118–120; n = 17) and UCLA pediatric sleep cohort (ds005398; n = 18) also showed no significant association. These results did not support the predicted decrease in AUC with greater SOZ shaft count. The confidence intervals do not exclude a weak negative association. Clinical labels differed between centers, shaft count also depends on implant planning, and recordings were short (about 20–120 s).

epilepsy · stereo-electroencephalography · SEEG · seizure-onset zone · interictal · high-frequency activity · spike rate · area under the curve · OpenNeuro