Abstracts

Rationale: Vagus Nerve Stimulation (VNS) as a means for seizure control has shown encouraging results although its mechanism of action remains largely unknown. What is clear is that the efficacy of therapy largely depends upon the values of VNS stimulus parameters, which are typically modified based on physicians feedback. Thus, patients might benefit from VNS if there was an objective measure of the efficacy of the therapy over time. Along these lines, we undertook a clinical study to investigate changes in the dynamics of scalp EEG over time in response to VNS therapy. In particular, we tested the hypothesis that effective VNS therapy causes a significant change in interictal resetting, that is, disentrainment (desynchronization) of pathologically entrained (synchronized) brain sites with respect to their rate of information exchange. Methods: Ten human patients with epilepsy and implanted VNS devices were enrolled in this longitudinal study (2 years). Each patient had a baseline EEG recorded after which VNS therapy was initiated alongside his/her preexisting AED regimen. Patients typically had monthly hospital visits to have their EEG recorded and therapy clinically assessed. Patients with >50% reduction in seizure frequency or severity were labeled as respondents to VNS. Up to 2 hour scalp EEGs were recorded per patient visit and analyzed to estimate how fast (resetting time) and how intensely (resetting level) entrained short-term maximum Lyapunov exponents (STLmax) of critical brain sites became disentrained. Probability distributions of resetting (PDRs) were constructed for each EEG record per patient s visit by employing Monte Carlo simulations. The divergence of time and level resetting were calculated for the respective PDR from each EEG record by measuring the Kullback-Leibler divergence from the corresponding PDR of the baseline EEG. ANOVA tests were used to correlate the divergence of time and level of resetting with a patient s clinical assessment.Results: ANOVA tests showed that respondents had negative divergences from baseline in both time and level resetting, while non-respondents had positive divergences. This was a very strong statistical finding using either time (p<0.0029) or level (p<0.0067) of resetting. The same result held when correlating divergences from time (p<0.0389) or level (p<0.0421) of resetting with clinical outcomes determined monthly per patient visit.Conclusions: The results of this study have shown a strong relationship between a measure of brain dynamics (resetting) and the clinical outcome in 10 patients with implanted VNS devices in the course of 2 years after implantation. It was shown that, when VNS therapy was effective, interictal resetting decreased, and when VNS therapy was ineffective, interictal resetting did not. Thus, this study indicates a possibly powerful mechanism of action of VNS on the epileptic brain. (Supported by NIH grant NS061310-01) S. Sabesan, N. Chakravarthy, K. Tsakalis, P. Pardalos & L.D. Iasemidis, J. Comb. Optim., vol. 17, pp. 74-97, 2009.