Grey-matter enrichment coexists with within-grey-matter SOZ signals in 18 pediatric sleep SEEG patients

We tested whether grey-matter (GM) or white-matter (WM) position explains interictal seizure-onset-zone (SOZ) signals in archived 80–250 Hz pair features from an OpenNeuro pediatric sleep cohort (ds005398). We analyzed 1,147 GM or WM pairs from 18 UCLA patients, using dataset channel labels and a single 120-second sleep window. The plan was written and hashed before analysis, with no deviations. Patient was the unit; analyses used per-patient AUCs, exact sign-flip tests, patient bootstrap 95% confidence intervals, and Holm correction across five primary tests. SOZ pairs were more often GM than WM (17.7% versus 8.0%; GM-indicator AUC 0.584, Holm p=0.0008). Within GM, both tonic level and 3× burst rate separated SOZ from non-SOZ pairs (AUC 0.644, Holm p=0.025; and AUC 0.682, Holm p=0.0058, respectively). Tissue matching reduced tonic AUC by 0.036 (95% CI −0.062 to −0.009; Holm p=0.038), a small effect that was not significant in the sensitivity analysis counting deep nuclei as GM (−0.026; Holm p=0.20). Matching did not significantly reduce burst-rate AUC (−0.014, 95% CI −0.040 to 0.010; Holm p=0.33). The matched AUC remained above chance, and the within-GM signals remained, so the preregistered criteria for tissue fully explaining either SOZ signal were not met. In WM, only eight patients had SOZ pairs and neither feature was significant. Limitations include dataset labels not checked against electrode coordinates, clinical SOZ labels, one window without sleep staging, 18 patients, low power for WM analyses, and a channel-relative burst threshold. The exact OpenNeuro release used is not recorded. The original plan, code, input table and hash, and results are preserved in the contribution archive.

stereo-EEG · seizure onset zone · grey matter · white matter · high-frequency activity · burst rate · pediatric epilepsy · sleep · AUC · tissue matching