Brain networks involved in epilepsy and techniques that affect those networks are receiving attention because of the great potential for intervention on the network level. Advancements stem from an increased sophistication of neuroimaging techniques and improvements in data processing. Neurostimulation techniques are evolving, along with our understanding of epilepsy and of network abnormalities. Ultimately, our goal is to have the largest impact on epilepsy and seizure control with the smallest intervention. 

The two types of neurostimulation are open-loop and closed-loop. Open-loop stimulation, historically used most widely, is not dependent upon seizures or feedback to trigger the stimulation. Open-loop techniques include the original vagus nerve stimulation system and the more recently approved deep brain stimulation system for epilepsy, and they have been found to be effective, perhaps owing to happenstance stimulation timing or to the effect of the stimulation on brain networks over time. Closed-loop techniques include responsive neurostimulation (RNS), which provides stimulation based on electrographic changes identified by a computer within the device. The closed loop refers to the feedback that triggers the stimulation. In conjunction with pharmacotherapy, RNS has been shown to be effective for refractory epilepsy, and this has generated increased enthusiasm for the future of brain-computer interfaces in epilepsy. There is much to be excited about with RNS; however, challenges to using this technique intelligently lie in identifying both the ideal electrode placement and the most impactful epileptic abnormality. A learning curve still exists, but I am enthusiastic about the technology’s development. 

Another interesting aspect from a research perspective is that long-term studies of patients with epilepsy who received RNS show that their seizure control generally improves over time. Potential confounds exist in these studies, and differentiating responses after 5 or 6 years across the devices and medications is difficult. A critically important question is: What was the natural history of the epilepsy in those patients who entered the trial? In other words, what would have happened if there had been no RNS intervention? A separate consideration relates to the heterogeneous patient groups regardless of all of the participants having epilepsy. Some forms of epilepsy may respond to the treatment, while others may not. The challenge in identifying the efficacy of each of these techniques is, in part, the challenge to better understand the natural history of epilepsy. Having a marker to track improvement or worsening of the epilepsy, in addition to seizure diaries, would be extremely valuable.