Abstracts

Rationale: We developed a closed-loop responsive Vagus Nerve Stimulation (VNS) platform to investigate the just-in-time effect of VNS stimulation on seizure frequency in the lithium-pilocarpine animal model of epilepsy. Methods: Eight male Sprague-Dawley rats were induced to chronic epilepsy by the lithium-pilocarpine model of epilepsy. Each rat was subsequently implanted with 8 Tungsten microwires (two in centromedial thalamus, two in hippocampus and four on cortex). Rats were then implanted with two helical leads made of platinum/iridium coated with silastic (Cyberonics Inc., Houston, TX) wrapped around the left vagus nerve in the ventral cervical region for VNS electrical stimulation by programmable stimulators (A-M Systems Inc., Sequim, WA). The experiment involved two phases per rat: stimulation (S) and no stimulation (NS), 4 to 5 days each in duration. During (S), VNS was administered upon warnings issued when analyzed EEG dynamics from critical pairs of brain sites were found to be significantly (p<0.05, t-test) entrained over the preceding 10 min interval. VNS stimulation was delivered until the critical pairs of sites disentrained or up to a maximum of 1 hour interval after a warning. By design, the maximum rate that just-in-time (JIT) VNS stimulation could occur was once every 3 hours; in reality the mean rate of JIT VNS was a fraction of this value, depending on each rat's EEG dynamics. This closed-loop procedure was performed automatically and in real time by our online, JIT seizure prediction software [1]. During (NS), no stimulation was administered at the issued warnings (control phase). Results: Seizures occurred in both phases of the experiment and were detected by our in-house developed seizure detection algorithms. The frequency of seizures in each phase was then estimated and compared with each other, per rat and across rats. Six of the eight rats showed a reduction in seizure frequency during the (S) phase compared with the (NS) phase. Cumulatively, across all rats, a statistically significant (p<0.05, t-test) reduction of the mean seizure frequency, from 8 seizures a day without stimulation to 2 seizures a day with stimulation, was observed over the experiment's duration (see Figure 1). Conclusions: Responsive, just-in-time, VNS was found to be effective in significantly reducing the seizure frequency in the lithium-pilocarpine animal model of epilepsy. These results provide an initial supporting evidence for the use of closed-loop just-in-time VNS stimulation in the treatment of epilepsy. Acknowledgement: This study was funded by Cyberonics Inc., Houston, TX. [1] Good L., Sabesan S., Marsh S., Tsakalis K., Treiman D. & Iasemidis L.D., Int. J. Neural Systems, 19(3), 173-196, 2009.