Commentary
Article
NeuroVoices: Alexander C. Whiting, MD, on Advancing Epilepsy Surgery and Brain Function Research
Author(s):
The director of epilepsy surgery at Allegheny Health Network discussed how clinicians can successfully pinpoint the source of epilepsy in the brain through innovative procedures like stereo electroencephalography.
Alexander C. Whiting, MD
(Credit: Allegheny Health Network)
Drug-resistant epilepsy (DRE) remains a major clinical challenge, as many patients continue to experience seizures despite optimal antiepileptic drug therapy. This refractory condition is often linked with progressive cognitive decline, further exacerbating the impact on patients' quality of life. Responsive neurostimulation (RNS) has emerged as a promising neuromodulatory treatment for patients’ ineligible for surgical resection or ablation, with increasing evidence supporting its role in seizure reduction. However, the effects of RNS on neuropsychiatric outcomes and its potential to mitigate cognitive decline in refractory epilepsy remain areas of active investigation.
A recent retrospective study evaluated the neuropsychiatric outcomes of 11 patients with DRE who underwent RNS implantation after a comprehensive diagnostic workup, including stereoelectroencephalography (SEEG) to localize the seizure onset zone. Pre- and post-operative neuropsychiatric assessments revealed no statistically significant changes in most cognitive domains, including memory, attention, and executive function. However, a modest yet statistically significant decline was observed in the Verbal Comprehension Index (P = 0.028). Overall, these findings suggest that RNS implantation may stabilize key neuropsychiatric functions in DRE patients, though further large-scale studies are warranted to assess the long-term cognitive implications of this treatment.
These results were presented at the 2024 American Epilepsy Society Annual Meeting, held December 6-10, in Los Angeles, by senior author Alexander C. Whiting, MD, and colleagues. In a new iteration of NeuroVoices, Whiting, director of epilepsy surgery at Allegheny Health Network, further discussed how the integration of machine learning and data analysis from SEEG can potentially help in developing personalized treatment plans for patients with epilepsy. Whiting also talked about the ways of how understanding of the brain's risk-reward circuitry can be applied to the treatment of other neuropsychiatric disorders. Moreover, he talked about the potential long-term benefits of using real-time data from patients with epilepsy for broader brain function research.
NeurologyLive: Can you provide an overview of using the SEEG in your latest research for epilepsy?
Alexander Whiting, MD: Epilepsy in general is a disease where, when it gets severe and the medications are no longer working, we start talking about surgery. Figuring that out can be very difficult because the brain is obviously very complicated, and somewhere inside this complex organ, something is causing all the problems. When we do MRIs for many of these patients with severe epilepsy, their brains often look completely normal. So, there’s something in there, like a needle in a haystack, and that’s what we’ve gotten pretty good at identifying with surgery: finding exactly where that needle in the haystack is causing all the problems. For some patients with particularly complicated epilepsy, we also need to determine which parts of the brain are involved in their epilepsy network.
We have a surgery called stereo electroencephalography, or SEEG, which I perform here at Allegheny Health Network. It’s a mouthful, so we just call it SEEG. In this procedure, I insert electrodes directly into the brain. Then, we bring the patients into the hospital, have them stay for a while, and they experience a few seizures. When they have a seizure with these electrodes in place, we can see exactly where the seizure starts, where it spreads, and what the epilepsy network looks like.
There are a lot of reasons I like this surgery. Patients appreciate it because it’s minimally invasive. In the past, epilepsy surgery was very scary, involving large incisions and being highly invasive. Now, we place these electrodes through holes so small that patients barely need a stitch to close them up. We essentially just poke them in. Many patients don’t even need to shave their hair, and they leave surgery looking and feeling much like they did when they arrived.
I like it because it’s safe. In all the major studies, including our own data from this program, the risk of a permanent complication, like a major surgical risk, is less than 1%. It really doesn’t get much better than that in neurosurgery. The epileptologists and neurologists working to identify the source of the seizure activity also love it because it provides incredible data. When we place the electrodes, sometimes we use 16, 17, or even 18 electrodes, each with 10 to 16 contacts. So, we’re essentially listening to 200 different parts of the brain. It’s an overwhelming amount of information, but it gives us remarkable granularity on exactly where a patient’s epilepsy originates and how it spreads. Once we know that, we can sit down with the patient and discuss all their options. There are many exciting things we can do now, like minimally invasive surgeries and devices that can prevent seizures before they start.
What research is being presented at AES 2024 involving the SEEG for patients with epilepsy?
We have a few things we’re presenting at AES this year, primarily focusing on the outcomes of patients who have gone through this program. What we’ve found is that the risk of SEEG is incredibly low and the yield is incredibly high. We’re really good at pinpointing where people’s epilepsy starts and understanding their networks. The outcomes after surgery are very positive—some patients experience seizure freedom rates as high as 70%. That’s remarkable, especially considering there aren’t many other fields in neurosurgery or surgery in general with success rates that high. This is a big part of what we’re presenting at AES.
What we’ve been doing more recently is exploring how the data we gather from these patients could be used to treat other diseases. These patients are often in the hospital for over a week, waiting to have a seizure. During this time, they’re interacting with the electrodes in their brains. They’re doing all sorts of things—getting into fights with siblings, paying their taxes, gambling on their phones, watching TikTok, and so on. As they do all these things, I’m watching their brain activity on the screen.
This raises an interesting question: Could we use this data to treat other diseases or deepen our understanding of how the brain works? A lot of what we’ve learned about brain function has come from indirect methods, like functional MRIs, animal studies, or observing patients after they’ve had a stroke. There hasn’t been much opportunity to have a direct, extended look at the brain of people doing various tasks. That’s where we’ve partnered with Carnegie Mellon University (CMU), which has some of the best data processors, signal processing experts, and machine learning specialists for analyzing big data. Pulkit Grover, PhD, professor of electrical and computer engineering at Carnegie Mellon University, is leading this work.
We’ve been having patients perform tasks or tests that activate certain parts of their brain, and we observe how different parts of the brain communicate with each other. What we’re focusing on now is the risk-reward circuitry in the brain. This is a fundamental part of every decision you make. Whether you're running a yellow light or choosing a salad over a steak, these decisions are influenced by the brain's risk-reward circuitry. It may seem simple, but there’s a lot of complexity in how decisions are made, and many mental health disorders—like obsessive-compulsive disorder (OCD), treatment-resistant depression, and addiction—are believed to involve dysfunction in this circuitry.
Our hope is that by analyzing this data from patients who have graciously agreed to participate, we can advance the field and ultimately address some of these difficult problems.
Looking ahead, what are your next steps for research with SEEG in this patient population?
I think what’s surprised us is just the volume of data and how good the data has been. A lot of that, I think, is because of the patients here who have been so willing to participate. They’re usually excited to do something because they’re lying in bed, bored out of their minds, waiting to have a seizure. But they’re also giving us their time, playing these games and doing these tests. I’m also very thankful to Grover at CMU because what he does is take all this data, and I always compare it to the movie The Matrix (1999), where you see that screen filled with numbers. That’s what we’re looking at—miles and miles of data coming out of these patients’ brains. Grover uses machine learning and signal processing to analyze it, and we can see which parts of the brain are communicating with each other while people are doing these tasks.
What we’ve been able to do—and we’re doing this now, not in the distant future—is look down to the microsecond when someone is making a decision. For example, when they’re deciding whether to run a yellow light, we can pinpoint exactly which parts of the brain are involved. We can also test this upstairs in the epilepsy monitoring unit. We’ve seen divergent streams in the brain, meaning different parts are activated. What we’re finding is that there’s a part of the brain that makes “good decisions” and a part that makes “bad decisions.” There’s also the control part of the brain, the one that says, “Let’s push through this.”
What we’re thinking is that we can probably apply this insight to various diseases, like addiction, treatment-resistant depression, and OCD, where decision-making can go off track and become uncontrollable. Can we intervene? We already do this with epilepsy. For patients with seizures, we can implant a device called a responsive neurostimulator. We place an electrode in a specific part of the brain, and it listens constantly. When it detects the beginning of a seizure, it can stop it before it starts. Why couldn’t we do something similar for other diseases?
For addiction, for instance, before a person engages in addictive behavior or spirals into a harmful decision, could we detect it in the brain and prevent it before it starts? For patients with OCD and compulsions, could we intervene to stop the compulsion and regain control? That’s where I see this going. And I don’t think it’s a distant, long-term future. I believe we’re going to be doing this in the next few years. Right now, we only use SEEG to implant electrodes in the brain for epilepsy, but I think in five or ten years, we’ll be doing it for all kinds of diseases. We’ll be able to say, “Something’s going wrong in this person’s brain—let’s find where it is, find the needle in the haystack, and fix it.”
Transcript edited for clarity. Click here for more AES 2024 coverage.
