Annual Meeting Abstracts: View
(Abst. 2.049), 2018
Interictal ECoG Features Correlate With Cognitive Outcomes in Individuals With Epilepsy
Authors: David W. Loring, Emory University School of Medicine; Kimford J. Meador, Stanford University; Beata Jarosiewicz, NeuroPace, Inc.; Tara L. Skarpaas, NeuroPace, Inc.; and Martha J. Morrell, Stanford University / NeuroPace, Inc.
Content: Rationale: Brain-responsive neurostimulation is a safe and effective adjunctive treatment for adults with medically intractable partial onset seizures, does not have negative neurocognitive effects, and may improve neuropsychological function (Loring et al., Epilepsia 2015: 56:1836–44). In a randomized controlled trial of the RNS® System, patients with neocortical (NC) seizure foci had statistically significant improvements in verbal fluency, and patients with mesial temporal (MTL) seizure foci had improvements in verbal memory. These improvements were not correlated with changes in clinically reported seizures, but higher verbal memory scores were associated with lower spike rates in the left MTL (Loring et al., AES, 2017). Here, we retrospectively tested whether cognitive performance is also related to any other electrocorticographic (ECoG) features. Methods: Neuropsychological assessments were administered during a pre-implant baseline period (0 weeks), and at 20, 55, and 104 weeks post-implant. The assessments included the Rey Auditory Learning Test (AVLT), the Boston Naming Test (BNT), and the Brief Visuospatial Memory Test (BVMT). The average total spectral power (1-125 Hz), and power in delta (0-4), theta (4-8), alpha (8-12), beta (12-25), low gamma (25-50) and high gamma (50-125) were computed from 1-4 minute long segments of ECoG activity recorded roughly twice per day over the 3 month period preceding each time point (except for baseline). ECoG segments containing ictal activity were excluded from the analysis. Relationships between cognitive outcomes and spectral power were assessed within-patient using Pearson correlation (r) of the 2 measures across time, producing 1 r value per patient (if unilateral) or per lead (if bilateral) for which both measures were available at 2 or more timepoints. Spectral power was weighted by the total number of recorded ECoGs. Distributions of r values were compared to 0 using 2-tailed weighted t-tests with a = 0.01. Results: For patients with seizures arising from the neocortex, changes over time in delayed recall (AVLT and BVMT) scores were inversely correlated with changes in alpha power and directly correlated with changes in beta power (Table 1). In contrast, for patients with MTLE, changes over time in BVMT scores were directly correlated with changes in theta power and inversely correlated with low and high gamma power (Table 2). Conclusions: These analyses reveal differential changes in ECoG power associated with changes in cognitive outcomes in patients with MTLE compared to patients with seizures arising from the neocortex. Specifically, decreases in low frequency activity and increases in high frequency activity were associated with cognitive improvement for patients with neocortical foci, whereas the inverse was true for patients with MTLE. Future studies could examine whether these spectral power changes have any implications relative to coupling across low and high frequencies (e.g. theta gamma coupling) and cognitive performance in patients treated with brain-responsive neurostimulation. Funding: None