Abstracts

Rationale: To explore the tolerability and efficacy of low-frequency stimulation (LFS) of white matter tracts (WMT) in patients with intractable mesial temporal lobe epilepsy (MTLE). Deep brain stimulation has had limited success in epilepsy. However, WMTs have not been stimulated, and low frequencies have not been adequately explored for treatment of epilepsy. Methods: Eleven adult patients with MTLE underwent implantation of depth electrodes in WMTs: the corpus of the fornix (n=2) and lateral to the splenium of the corpus callosum where fibers of the dorsal hippocampal commissure travel with the crus of the fornix (n=9). All patients had hippocampal electrodes. 1-Hz stimulation of the WMT in all subjects elicited hippocampal evoked potentials (EPs), confirming connectivity. The EP duration was approximately 200 msec suggesting that 5-Hz stimulation would elicit continuous interference of hippocampal activity. LFS-WMT occurred in 4-hour sessions (1-3 sessions/patient at 5Hz, current 8mA, pulse width 0.2msec) while in the video-EEG unit. Results: There were no complications, and hourly mini-mental status examination scores showed a mean increase of 0.9 during LFS-WMT (n=11; P<0.001, Generalized estimating equations [GEE]-Identity Link Function [ILF]). Patients 8-10 proved to have an extrahippocampal seizure onset, so they were excluded from the spike and seizure analysis, but were still included in the MMSE analysis. LFS-WMT resulted in a significant reduction of hippocampal spikes (n=8; p=0.001, GEE-ILF), which persisted for the subsequent 4 hours (p=0.05). Seizure odds were reduced by 91% in 1-2 days following each 4-hr LFS-WMT session (n=8; p = 0.001, GEE- Logit Link Function), without medication changes. Conclusions: Our data show that LFS-WMT activates the hippocampus and other areas of Papez circuit, and reduces epileptiform discharges and seizures, without affecting memory. Stimulation of WMTs appears to be attractive because it interferes with spontaneous electrical activity of a relatively extensive cortical area, and requires less current injection, implying more safety and a longer battery life. Our data illustrate the need for an exploratory trial to confirm safety and tolerability of LFS-WMT using different parameters using a chronic implantable pulse generator.