Abstracts

Rationale: Patients with drug-resistant epilepsy treated with responsive neurostimulation (RNS) may require scalp EEG monitoring for various reasons, e.g. to assess therapeutic response, evaluate new ictal behaviors, gauge the degree of consciousness impairment during long episode detections, or reconsider surgical options. As such, it is critical to identify RNS stimulation artifacts produced by various stimulation settings on scalp EEG to avoid misinterpretation.

Methods: We identified eight adult patients with RNS who underwent scalp video-EEG monitoring in the epilepsy monitoring unit and obtained representative epochs of stimulation artifact amidst patients’ background or epileptiform activity. We aimed to collect stimulation examples that varied according to RNS lead location (mesial temporal, neocortical, or thalamic), stimulation pathway (bipolar, lead-to-lead, or cathodal), and stimulation frequency (high or low).

Results: We observed a distinct pattern associated with RNS stimulation artifact, characterized by high-amplitude, short-duration, spike-like activity with a field extending over the region closest to the stimulating electrodes (whether mesial temporal as shown in Figure 1 or neocortical/corticothalamic as shown in Figure 2) and over the generator (typically placed over the parietal bone on either side). RNS stimulation with more localized pathways (bipolar or lead-to-lead) would typically result in more focal artifact patterns compared to those with more regional cathodal pathways. The RNS burst duration (whether short at 100-200ms or 5000ms) and stimulation frequency (whether 200 Hz or 10 Hz) would be reflected on the scalp EEG, allowing the artifact to stand out from typical interictal background activities including post-craniotomy breach effect. In particular, three examples of RNS stimulation artifact during ictal epochs demonstrate that the artifact can be distinguished from ictal activity seen on scalp EEG due to the artifact’s morphology, sporadic or periodic pattern, and lack of evolution especially with respect to the visibly evolving ictal rhythm. Identifying these artifact signatures, especially when occurring in a sequence of five bursts indicating a long episode, improved temporal correlation of RNS electrocorticography recordings to scalp EEG patterns to determine whether RNS detections were near seizure onset.

Conclusions: RNS stimulation artifact can be distinguished from interictal and ictal activity by the presence of high-amplitude, short-duration, spike-like activity with a characteristic broad field and features correlated to programmed stimulation parameters. This is the first in-depth evaluation of RNS stimulation artifacts on scalp EEG across various implant locations and stimulation paradigms.

Funding: This study received no funding support. DSE has received clinical trial grant support and speaker bureau honoraria from NeuroPace.