SEEG contacts had higher seizure-onset-zone and resection fractions than ECoG contacts in one epilepsy cohort

This within-patient comparison examined clinical channel labels in 76 patients with at least five SEEG and five ECoG contacts, using channel metadata from OpenNeuro ds007118, ds007119, and ds007120. For each patient and contact type, the analysis calculated the fraction of channels labeled artifact, spiking, or seizure onset zone, and the fraction of eligible SEEG contacts labeled resected (74 patients). A two-sided sign-flip test compared mean paired differences (SEEG minus ECoG), with 200,000 random sign flips (seed 0) and Bonferroni correction for four measures. Mean seizure-onset-zone fractions were 0.145 for SEEG and 0.068 for ECoG (difference +0.077; p=0.0012, corrected p=0.0048). Mean resected fractions were 0.531 and 0.291 (difference +0.244; p=5e-6, corrected p=2e-5). Artifact fractions were 0.045 and 0.055 (difference -0.009; p=0.61), and spiking fractions were 0.186 and 0.148 (difference +0.038; p=0.14); neither difference was significant. More patients had a higher ECoG than SEEG artifact fraction (41 vs 13), despite the small, non-significant mean difference. These are clinical labels, not measurements of signal quality. Contact placement was not random: SEEG targeted suspected deep or focal sources, so the seizure-onset-zone and resection differences reflect clinical targeting and do not establish a property of contact type. The analysis is from one center and one pediatric and adult epilepsy-surgery cohort; results have not been tested at other centers. No signal data were used.

SEEG · ECoG · epilepsy · clinical channel labels · seizure onset zone · resection · intracranial EEG · paired analysis