Abstract Details

About 30% of patients with epilepsy are drug resistant (DRE) and require surgical intervention. Many DRE patients undergo intracranial EEG monitoring to localize the seizure onset zone (SOZ) and electrical stimulation mapping (ESM) to map eloquent cortices. However, ESM is associated with certain risks and limitations, therefore high-gamma modulation (HGM) has emerged as an alternative modality particularly for language mapping. Because epilepsy surgery decisions are currently based on ESM, and have been shown to insufficiently protect against post-surgical deficits, we evaluated HGM during a visual naming task in relation to cortical resections for determination of objectively measured neuropsychological outcomes.

We developed a mathematical model to determine long-term neuropsychological outcomes after epilepsy surgery based on high-gamma modulations during a visual naming task and anatomy of neurosurgical resection.

24 DRE patients with HGM language mapping and pre- and 1-year post-surgery NPE were included. Analysis of covariance models were fitted to analyze post-surgery neuropsychological evaluation (NPE) scores as a function of high-gamma power change in individual intracranial channels and whether that channel was resected or not, adjusted for respective pre-surgery NPE scores.

Resection of electrode contacts showing significant high-gamma power change was associated with a significant post-surgery decline in calculation (regression coefficient=-1.52, 95% CI -2.73 to -0.32, p=0.013) and passage comprehension scores (regression coefficient=-1.49, 95% CI -2.11 to -0.87, p<0.001), however, a paradoxical increase in post-surgery story memory immediate and verbal learning recognition scores was seen.

Current practice epilepsy surgery excludes ESM naming sites, however, our study shows that HGM mapping when analyzed with planned resection is helpful in identifying additional neuropsychological domains. We plan to add other clinically relevant variables, particularly seizure-freedom, to further strengthen our models. We also intend to evaluate model performance to obtain HGM thresholds that can identify specific NPE deficits.