NeuroPace: Controlling Epilepsy With a Brain Implant

A smarter technology reduces seizures in epileptic patients by identifying dangerous patterns of brain activity

Loading the podcast player...

Stephen Cass: Hi. You’re listening to IEEE Spectrum’s “Techwise Conversations,” and I’m your host, Stephen Cass.

Back in 1972, Michael Crichton published the technothriller The Terminal Man, later made into a movie of the same name. The plot centered on a man who was implanted with a microcomputer to predict and control his epileptic seizures. Unsurprisingly, this being a Crichton novel, things don’t go too well.

But now, 40 years later, the FDA looks close to finally approving a real-life implant to control epilepsy, and which, unlike the science-fiction version, has performed well in clinical trials, safely easing the symptoms of many patients.

My guest today is Frank Fischer, the CEO of NeuroPace, the maker of the new so-called RNS implant. He joins me by phone from his office in Mountain View, California.

Frank, welcome to the podcast.

Frank Fischer: Thank you, Stephen.

Stephen Cass: So many listeners will be somewhat familiar with the kind of deep-brain stimulation that is used to treat Parkinson’s disease. How is the RNS implant different?

Frank Fischer: We utilize what’s known as responsive neurostimulation, whereas the Parkinson implant either delivers stimulation continuously—well, for Parkinson’s, it typically is continuous stimulation—and what we do is, once the system is implanted, it is tuned to look for the patient’s specific abnormal electrical activity that may lead to a seizure, and once that pattern is detected, and it could be one or two patterns for a given patient, basically the device then delivers imperceptible levels of stimulation to disrupt that abnormal activity so it does not get picked up by essentially the rest of the brain, resulting in a seizure. So, whereas for, let’s say, a Parkinson’s implant, where there may be stimulation for 24 hours in a 24-hour period, for patients that we have that have high degrees of what is known as electroform activity, the device may deliver 5 minutes of stimulation in a 24-[hour] period. So that’s the basic conceptual difference.

Stephen Cass: So is the implant actually put into the brain? Where are the electrodes implanted? And I understand there’s a range of places you can place those electrodes.

Frank Fischer: Yeah, we have patients in our clinical trials that have epileptic foci in many different parts of the brain, and so the device itself is placed in the skull. It’s the same thickness of the skull and curved, so it replaces a part of the skull, and it is connected to two leads—insulated wires with electrodes on the end—and those electrodes are placed near the patient’s epileptic foci. We’re able to treat patients with one or two foci, and therefore place the leads near one or both foci.

Stephen Cass: And what have the clinical trial results been like so far?

Frank Fischer: Well, we’ve conducted a pivotal study, and in that pivotal study there’s what’s known as a blinded evaluation period, where two months after implant, half the patients have their devices turned on, and half the devices are turned off, and the primary effect of this end point was to look at the difference in seizure-frequency reduction in the treatment group as compared to the control group.

In the treatment group, the average, or percent, reduction, if you will, of seizures was just about 38 percent—was 37.9, specifically—and then the control group was 17.3 percent, and this is each of these patients compared to their baseline frequency, and that difference was statistically different with a p-value of .012. If we look out over the, essentially, the entire two-year pivotal trial, in epilepsy, typically a responder to a treatment is identified by a reduction of 50 percent or more in their seizure frequency compared to their baseline. And in our case, in the two-year period, 55 percent of the patients were responders by that definition of 50 percent reduction or greater as compared to their baselines.

Stephen Cass: Why doesn’t it help even more patients? Why is it only 55 percent who saw that decrease?

Frank Fischer: Well, it may well in the future. Basically, the patients that were in our clinical trial were very, very challenging patients. They had—they came into the study taking, on average, 2.8 antiepileptic drugs concurrently. A third of these patients had had prior epilepsy surgery, where a portion of their brain that was thought to generate seizure activity was removed. A third had had the only medical device that’s approved for treatment of epilepsy called the VNS, and that treatment was not successful for them. So, virtually every treatment that was available had been tried and had not worked. So, this was clearly the most difficult category of patients in which to do this clinical trial. So, I think between that and the fact that patients will be less severely impacted as we expand the population, plus the fact that we will continue to learn more and more about the most appropriate therapeutic settings for given types of patients and things of this nature, it’s certainly our hope that performance will continue to improve in the future.

Stephen Cass: So there’s been a lot of interest in things like heart pacemakers being vulnerable to malicious hackers. What’s the risk of someone hacking the RNS implant?

Frank Fischer: Basically, there’s a protocol that exists that requires the implant to interact with what we call the programmer. The programming mechanism is inductive, so it requires that a wand be placed adjacent to the implant and interactions to occur. It’s always possible that if someone really wanted to devote incredible resources to trying to find out a way to replicate a program or program activity and cause the device to operate in a manner that would be inappropriate, it would be possible to do that. However, if you really think about it, you’ve kind of got two options: One option is that the device doesn’t function, and in that case the patient is exactly where they were before the implant in terms of the inability to have the implant respond to seizure activity. Or you could have the device deliver a whole range of inappropriate stimulation that might lead to a seizure, which would be entirely theoretical. In our experience, we don’t know of any instances in our clinical trials where stimulation, you know, actually led to a seizure. So, I think it’s one of these things that is technically possible, highly unlikely, and most importantly, the real question would be, Why would anybody want to do that for this type of implant?

Stephen Cass: So what other neurological conditions might benefit from this technology?

Frank Fischer: Well, this is now theoretical, because we haven’t done actual work in this regard, but it’s totally logical to me that any neurological condition that results in changes in brain state, where the brain state, by monitoring the brain state that can be determined and then treated by virtue of the delivery of stimulation, may well be able to be treated by virtue of this technology. So if you think of some examples of that, I mean, depression, for example, is a state that the neurologic condition of the patient varies; bipolar is in that category. It’s possible that movement disorders are in that category in a different way. So, you can think of a whole host of different things that potentially this technology could be applied to, and that’s what makes the future so interesting.

Stephen Cass: Well, Frank, thanks so much for joining us today.

Frank Fischer: Thank you, Stephen.

Stephen Cass: We’ve been speaking with CEO Frank Fischer about his company’s neural implant to treat epilepsy. For IEEE Spectrum’s “Techwise Conversations,” I’m Stephen Cass.

This interview was recorded Tuesday, 5 November 2013.

Illustration: NeuroPace

Audio engineer: Francesco Ferorelli

Segment producer: Barbara Finkelstein

Read more “Techwise Conversations,” find us in iTunes, or follow us on Twitter.

NOTE: Transcripts are created for the convenience of our readers and listeners and may not perfectly match their associated interviews and narratives. The authoritative record of IEEE Spectrum’s audio programming is the audio version.