Abstracts

Rationale: Clinicians rely mainly on patient self-reported seizures regarding seizure frequency. However, self-reported seizures may be inaccurate due to various reasons. A responsive neurostimulator (RNS), which is used as an adjunctive therapy for intractable epilepsy, also provides intracranial electrocorticographic recording (ECog), and therefore delivers valuable information on seizures. We directly compare seizure frequency by patient’s report with RNS recordings in this study. Methods: Case series at George Washington University Hospital (GWUH). We report both “self-reported” seizures and “RNS-detected” seizures. Self-reported seizures were obtained from patients or caregivers and documented in clinic notes. RNS magnetic-swipe events were also reviewed and then further confirmed by patients. Long episodes (30 seconds or above) were used as surrogates for “RNS-detected” seizures. Long episodes were also validated from ECog library where they were stored. We reported the seizure frequency of the most recent 9 months. Results: Total 5 patients (3 females; age range: 30-58) were identified, and the RNS devices have been implanted for 6 months to 4 years. Three patients (# 1, 2, 4) have a unilateral temporal implantation. Two patients (# 3 and 5) have RNS electrodes implanted in bilateral mesial temporal regions. In this cohort, patients were also placed on 1-3 anti-seizure drugs (AEDs) for seizure control. These AEDs include briveracetam, lamotrigine, lacosamide, levetiracetam, oxcarbazepine, phenobarbital, valproic acid and zonisamide.One patient (# 1) accurately reported his seizures, which were overall 1-3 seizures a month and were characterized by altered awareness and confusion. This self-reported seizure frequency was consistent with RNS ECog findings. In addition, most of the time when a long-episode was detected (> 80%), the patient reported a clinical seizure.Two patients (# 2 and 3) underreported their seizures. Among them, one patient’s (# 2) ECog recordings showed about 3-7 long episodes per month; however only ~ 50% were reported by the patient (loss of awareness). The other patient (# 3) had overall ~ 10 long episodes a day and the patient only reported 1-2 seizures per month. Two patients over-reported seizures (# 4 and 5). Both patients (# 4 and 5) reported frequent events for which the patients interpreted as seizures (overall 1-5 and 2-9 per month respectively). These self-reported events were characterized by dizziness, seizure-like-feeling from head (#4) and “altered-hearing” (hearing from long distance) (#5). Multiple magnet-swipes events were also captured in a patient (#4). There were no long episodes associated with these self-reported events or magnet-swipe episodes. Therefore, these self-reported events are not seizures, and instead are likely hypervigilant non-specific events. Interestingly, one patient (# 5) also reported occasional (about once every 2-4 months) focal seizures with altered awareness which were confirmed by the RNS ECog recording. Conclusions: Patients may either over-report or underreport seizures. Providers could underestimate seizure frequency by patients' self-report. It can also be challenging to differentiate focal seizures without awareness from non-epileptic hypervigilant behaviors by patients’ descriptions. RNS ECog recordings provide further objective electrophysiological evidence, and therefore could be helpful in the diagnosis and treatment. Funding: No funding