Abstracts

Rationale: Vagus Nerve Stimulation (VNS) is an adjunctive therapy for patients with medically refractory epilepsy. The primary metric used to assess response to any treatment for epilepsy is seizure frequency reduction. Seizure severity reduction is also clinically meaningful to patients and can be measured objectively. Previous studies (e.g., Neuromodulation 2006; 9(3); 214-220 and Neurosci. Lett. 2013; 536; 14-18) have shown that seizures are accompanied by spatial synchronization of the brain signals as measured by electroencephalographic (EEG) signals and increase in heart rates as measured by electrocardiaographic (EKG) signals. In this analysis, we utilized quantitative features obtained from a combination of EEG-EKG data around ictal events to identify if automated delivery of VNS at the time of seizure onset reduces the severity of seizures as assessed by reduction of EEG spatial synchronization as well as the duration and magnitude of heart rate increase.Methods: 51 patients were enrolled across two clinical trials (NCT 01325623, NCT 01846741) wherein patients were implanted with the AspireSR® VNS Therapy System. This system automatically delivers stimulation when an increase in heart rate, which may be associated with a seizure, is detected. EEG and EKG data were collected from patients prior to and after implant. A total of 124 pre-VNS implant seizures and 156 post-VNS implant seizures were available for analysis. To quantify the severity of each seizure, we extracted three features 1) heart rate change (in %) during a seizure, 2) duration of the heart rate change (in seconds), and 3) measure of spatial synchronization of EEG during a seizure. The extracted features from all seizures were then provided to an unsupervised Fuzzy C-Means (FCM) clustering algorithm to determine whether seizures in this population can be grouped based on their severity.Results: Application of the FCM algorithm resulted in clustering of the seizures into two independent groups. Further investigation revealed that one group consisted of seizures from pre-VNS and the other group consisted seizures from post-VNS. The mean value and standard deviation of the features across seizures pre- and post-VNS are shown in Table 1 along with their corresponding t- and p-values. It appears that the mean value for each feature is significantly lower (p<0.05) for post-VNS treatment compared to pre-VNS treatment. In addition, Table 2 demonstrates the classification performance for determining pre- and post-treatment seizures is 83.57%. These results indicate that the combination of selected features show the effect of VNS Therapy in reducing ictal EEG synchronization and the magnitude and duration of heart rate increase.Conclusions: This study proposes a new method of measuring the effectiveness of VNS Therapy by combining EEG and EKG derived features to assess seizure severity. Using these features, we demonstrate that the closed-loop AspireSR VNS Therapy System may have the ability to reduce seizure severity.