Closed-loop seizure prediction and prevention in rats with kainate-induced seizures

Many studies have shown that continuous or intermittent electrical stimulation of the brain can reduce or prevent the occurrence of epileptic seizures in humans and animal models. However, there have been relatively few studies that assess the effects of stimulation delivered just prior to seizure onset. Here we use a kainate-induced seizure model in the rat to test a closed-loop seizure prediction and prevention system. An algorithm was created that extracts a measure from the activity of populations of single neurons, and predicts the probability of a seizure in real time. Once a seizure is predicted, high frequency current pulses are applied to the hippocampus to attempt to inhibit the network and prevent the seizure from occurring. Results show that although not every seizure could be prevented, the majority of stimulation trials delayed or prevented a pending seizure. These results suggest that a closed-loop seizure prediction algorithm based on neuronal activity coupled with intracranial stimulation may be more effective than random stimulation at preventing the onset of seizures.