Abstracts

Rationale: Brain responsive neurostimulation is a safe and effective adjunctive treatment for medically intractable partial onset seizures, does not have negative neurocognitive effects and may improve some aspects of neuropsychological function (Loring et al., Epilepsia 2015: 56:1836–44). In a randomized controlled trial (RCT) of the RNS® System, patients with seizures from neocortical foci had statistically significant improvements in verbal fluency, whereas patients with mesial temporal lobe epilepsy (MTLE) had improvements in verbal memory, with the maximal benefit in those with poorer baseline memory scores. However, improvements were not correlated with changes in clinically reported seizures. Since interictal discharges can transiently impact cognitive performance (Binnie and Marston Epilepsia 1992: 33(s6):11-7), quantitative electrocorticographic (ECoG) data obtained by the RNS System were used to assess the relationship between memory performance and interictal spikes. Methods: Neuropsychological assessments were performed in subjects participating in a RCT of the RNS System (NeuroPace, Inc., Mountain View, CA) during a pre-implant baseline, at the end of a blinded period, and at 1 and 2 years post-implant (Y1, Y2). Baseline memory scores were characterized as “high” or “low” using a median split for the Rey Auditory Learning Test (RAVLT). The average interictal spike rate (ISR) was calculated from ECoGs stored routinely by the neurostimulator based on time-of-day. At least 10 stored ECoGs were required over the 3 months prior to each post-implant assessment to be included in the analysis. Pearson correlations were calculated between the average ISR and the RAVLT scores, as well as between the change in RAVLT score and the change in average ISR for Y1 compared to the blinded period, and for Y2 compared to Y1. Results: Table 1 shows the n, average number of ECoGs per patient, and mean ± SD for the RAVLT score and ISR for the MTLE and neocortical patients at the different assessments. MTLE patients with low baseline RAVLT scores had significant negative correlations between ISR and RAVLT scores at Y1 (r=-0.59; p=0.0007) and Y2 (r=-0.42; 0.047), as well as between the change in the ISR and change in RAVLT at Y1 (r=-0.47; p=0.01). Correlations were not significant between ISRs and RAVLT scores for MTLE patients with high baseline scores, or for changes in these measures. In addition, there were no significant correlations for patients with neocortical onsets, whether baseline RAVLT scores were high or low. Conclusions: There is a statistically significant negative correlation between the rate of interictal spikes and verbal memory performance in patients with MTLE and poor baseline memory performance, but not for patients with MTLE and higher baseline memory performance or for patients with neocortical onsets. These results suggest that interictal spikes could contribute to poor memory performance in patients with MTLE, and that reduction in interictal spikes could improve memory performance in those with baseline memory deficits. Funding: None.