Interictal attenuation in pediatric electrocorticography can be reliably detected by EEG readers.

Overview

Intraoperative electrocorticography (ECoG) helps to demarcate epileptogenic cortex, but a commonly observed feature, interictal attenuation, has received little attention. This may limit its use in the determination of the resection margin. In order to test how reliably EEGers can discern attenuation, we assessed how well EEGers agree with each other and with an objective, quantified measure of attenuation. ECoG segments (n=34) were evaluated for attenuation by two EEGers independently and in consensus, and by an amplitude spectral analysis-based quantitative method. A third EEGer divided the 34 ECoG segments into 3 subgroups-physiologic field present, physiologic field uncertain, and physiologic field absent-based on the clustering patterns of the attenuated electrodes. Inter-rater agreement between two independent EEGers (kappa=0.56) was moderate, and between consensus EEGers and the quantitative method (kappa=0.71) was substantial. These agreements were especially good among the physiologic field present subgroup where the attenuation clearly involved contiguous electrodes, and thus, more likely pathologic (kappa=0.64 for two independent EEGers and kappa=0.78 for consensus EEGers and quantitative method). Our results suggest that interictal attenuation, especially when involving contiguous electrodes, is an ECoG marker that can be consistently and reliably discerned by trained EEGers.