Abstracts

Rationale: The role of the thalamus in the generation and propagation of paroxysmal events has been postulated since Penfield and Jasper developed the centrencephalic theory of epilepsy in 1954 (Meeren et al., 2005). While some seizures may arise directly from thalamocortical oscillations (Steriade and Amzica, 2003), the extensive reciprocal connectivity of the thalamic nuclei with the cortex and other subcortical areas may also facilitate propagation of seizure activity to a larger epileptic network (Timofeev and Steriade, 2004; Archer et al., 2014). Accordingly, electrical stimulation of the thalamus has been demonstrated to reduce seizure frequency and severity (Salanova et al., 2015; Dalic et al., 2022).

Methods: In this single-center study, responsive neurostimulation of the thalamus with the NeuroPace RNS System was used to treat patients with epilepsy (N=41). The electrodes were implanted in the centromedian, pulvinar, or anterior nucleus depending on the epileptic etiology and semiology. The RNS System configured for two 90 sec. scheduled (interictal) local field potential (LFP) recordings to be stored each day. Here, we investigate changes in the average within-recording power of pre-defined frequency bands (delta, theta, alpha, beta, low gamma, and high gamma) of the thalamic LFPs over time.

Results: Within subjects, we observe prominent clusters in mean frequency in some bands of the LFP channels over time. Further, we note several interesting characteristics of these clusters. First, the clusters are often demarcated by the time of day when they were collected, and therefore may represent wake/sleep differences. Second, the time trend of the putative day/night clusters can be different across frequency bands, can have different slope and direction, and can diverge over time. This divergence implies that the long-term neuromodulatory effects of thalamic responsive neurostimulation on background wake/sleep activity may differ. Finally, the clusters are more prominent on some channels than others, even between adjacent contacts on the same leads. This may reflect a different position with respect to the targeted thalamic nucleus and may help to localize the electrodes after implantation.

Conclusions: These results suggest a possible wake/sleep specific neuromodulatory effect of thalamic responsive neurostimulation.  They also underscore the importance of scheduling LFP recordings during times of probable wake and sleep. Furthermore, given the central role of the thalamus in sleep oscillations and regulation of consciousness, as well as the fundamental interaction between sleep and epilepsy, consideration of sleep/wake cycles may be an important component of thalamic stimulation therapy.

Funding: Not applicable