Vagus nerve activity based prediction of epileptic seizures in rats

Continuous open loop vagus nerve stimulation (VNS) is an option for the treatment of refractory epilepsy. Animal studies have shown that VNS is able to shorten or even abolish ongoing seizures and that stimulation prior to a pending seizure may prevent the seizure from occurring. VNS therapy might therefore be improved by close loop stimulation, but seizures must then be predicted or detected. The current work investigated the possibility of predicting seizures based on vagus nerve (VN) activity. The envelope of the VN ENG power was synchronized averaged to the R peaks of the ECG and a feature was extracted from this average. Using this feature it was possible to predict onset of pentylenetetrazole (PTZ) induced tonic seizures in 4 out of 5 rats, 97 ± 59 s prior to the onset of tonic seizures. The current work indicates that it is possible to predict seizures based on VN recordings. However, further work is needed to evaluate the specificity and the sensitivity to noise of this method.