How Brain Mapping And AI Improve Seizure Localisation And Surgery - Christos Papadelis, Ph.D., Cook Children's Medical Center, TX, USA
Can’t find where seizures start in drug-resistant epilepsy? Dr. Christos Papadelis, biomedical engineer and researcher, explains how ultra-high-resolution brain mapping and AI are improving seizure localisation and surgical decision-making in children. This episode explores how combining advanced neurophysiology with AI can identify seizure origins and networks, predict surgical outcomes, and move care beyond traditional limitations.
📺 Watch the Interview
🎧 Listen to the Podcast
👉 Listen on Spotify | Apple Podcasts | YouTube
This episode is sponsored by EASEE® by Precisis who’s had no influence over the editorial content or discussion. Learn more about EASEE® here.
Episode Highlights
Why seizure origins are still missed in many children
How 550-sensor brain mapping improves localisation
Identifying seizure networks and critical hubs
Predicting surgical outcomes using AI
About Christos Papadelis
Christos is the Assistant Vice President of Neuroscience Research at Cook Children's Health Care System, a Professor of Research in Bioengineering and Director of the Pediatric Brain Health and Neurosciences Center at the University of Texas at Arlington, and a Professor of Pediatrics in the Burnett School of Medicine at Texas Christian University. He has more than 20 years of experience in magnetoencephalography (MEG) and electroencephalography (EEG) technology with both adults and children.
Full profile: Christos-Papadelis
Topics mentioned
epileptogenic zone
drug-resistant epilepsy
paediatric epilepsy surgery
magnetoencephalography (meg)
high-density eeg
electromagnetic source imaging
intracranial eeg
machine learning
functional connectivity
-
00:00 Christos Papadelis
The task that we have now is like not to identify where the seizures come from, but where they are located, what we call “pathological hubs” of these networks in order possibly to resect these networks, these hubs, and the patient to become seizure free. To give it an example is like being in a busy area or a big metropolitan area and you have a traffic jam in the whole area. And maybe sometimes… this is because there's some constructions or some kind of an accident in a specific critical intersection. And if this will be fixed, then the whole network will be fixed. So, in a similar way, we believe that in the epilepsy, there's critical hubs that if we are able to resect or intervene there, then the brain will not seize anymore.
00:48 Torie Robinson
Welcome to Epilepsy Sparks Insights! I’m your host, Torie Robinson, and today I’m joined by Dr. Christos Papadelis, who is going to be sharing with us how ultra-high-resolution brain mapping and AI are transforming how we identify seizure onset and improve surgical outcomes in children with drug-resistant epilepsy!
We get into what “precision” really looks like in practice – using a combination of more than 550 ultra-high-density EEG and MEG sensors to localise seizure activity - with far greater accuracy than traditional approaches.
If you’re new here, please subscribe so you don’t miss future conversations - and let’s get into today’s episode - presented in partnership with EASEE®, by Precisis GmbH.
01:28 Christos Papadelis
We work on two main areas of epilepsy. One focus on children with drug resistant epilepsy. As you know, like approximately 30% of the children with epilepsy, they are unable to control their seizure with drugs, and one of the best available options for these children is epilepsy surgery. And this is a critical epilepsy surgery to be performed early in life because the child's brain has high levels of neuroplasticity. So, there's plenty of time for the child's brain to rewire and reorganise. With my team, we develop innovative tools and methods that help us to identify, with high precision, the area where the seizures they come from (what's called the epileptogenic zone). We use different kinds of methods, we use non-invasive techniques, and we are one of the places in the world that have this unique setup that allows the simultaneous recording of the magnetic and electrical activity generated by the human brain by using more than 550 sensors covering the child's head. We combine two different techniques, one technique called MEG or Magnetoencephalography (it measures the magnetic activity of the brain). And then we combine this with a high-density EEG - it's a technology with a high number of sensors that measure the electrical activity of the brain. And by bringing these two techniques together, we have a really high precision to localise where this is come in a child with drug-resistant epilepsy.
03:01 Torie Robinson
Could you just tell us how many - because I know you said 256 electrodes in the ultra-high-density EEG - could you say that because people will be like “Oh my God!”
03:01 Christos Papadelis
For the high-density EEG system, we use 256 electrodes. We have a kind of system that is very child friendly. You don't have to put individual leads for the EEG (as in the conventional EEG), we use nets and these nets; they are made out of electrodes that have attached to them sponges, we soak the whole net in a bucket of water, then all these sponges they become conductive, we place the net on the child's head in the first 10 minutes; we start performing EEG recordings with so high number of electrodes.
03:44 Torie Robinson
This is so cool because I think most people who don't know much about these advanced EEGs; they just think of the scratchy stuff that they have put onto their scalp and then have a few electrodes just plonked onto their scalp and then…
03:53 Christos Papadelis
Yes!
03:53 Torie Robinson
…electrodes plonked on there, and it's really annoying. But to have something just plonked into some water (it doesn't electrocute anybody!) and then have that put on your head, especially as a child or especially as somebody, I'm generalising, but say for instance, somebody with autistic traits or something, that must be an amazing, positive alternative.
04:13 Christos Papadelis
Yes, it's very helpful for the children and then most of the recordings there are quite straightforward. Also, MEG technology is very child friendly because if you perform MEG by itself without the high-density EEG there is no need to place any kind of electrodes on the child's head. It has a helmet and then the child puts its head inside of this helmet and lays down on a bed and in a very quiet environment with no noise or any kind of interference like [they] sleep for an hour, and we are able to record this activity from the brain and localise in several cases with high precision where the seizure came from.
04:53 Torie Robinson
So cool. And then you can, I believe you combine the data collected from the MEG with from the data collected from the EEG to make something beautiful and have good insight into the activity and identify the epileptogenesis. Is that right?
05:06 Christos Papadelis
We use a new technique called electromagnetic source imaging; so, we combine findings from these two methodologies into a single solution and then this helps us to improve the accuracy that these two techniques they offer. Because these two techniques, have different kind of like sensitivity profiles, what we call. They can localise generators in the brain with different kind of like sensitivity. So, we combine findings in one solution and then this gives better localisation.
05:36 Torie Robinson
Could you give us an example of = a child that you've worked with and what the outcome of this may have been?
05:41 Christos Papadelis
Oh yeah! We had a fascinating case where my team helped a child with drug resistant epilepsy and who struggled with this kind of disorder for more than 16 years. And the patient had visited different hospitals in the United States with no success, the different clinical teams, they had not be able to identify where this is they were coming from. And eventually she came here to Cook Children's and we helped the clinical team to solve this difficult case. My research team was involved heavily into this case. We performed MEG and high-density EEG recordings, we identified that the seizures that were coming from insula (it's one area in the brain that is a little bit like deep and it's more difficult to capture activity from there). We presented this evidence to the clinical team, the clinical team was convinced that like we have a case there, and then they placed intracranial electrodes invasively to confirm that's the case. They found out that indeed seizures they were coming from this location. The child was ablated with laser (and that's also another technique that Cook’s does, we are really innovative) and the child is seizure free now for five years.
06:54 Torie Robinson
Just to give people perspective, you know, often people with an epilepsy have symptoms other than seizures. How are they doing otherwise?
07:01 Christos Papadelis
It's not only seizures. And patients with epilepsy they don't deal only with the seizures. That's the main component, what is the most well-known and described, but they have several other comorbidities. Several of the people with epilepsy, they have like intellectual disabilities, they struggle at school, they have depression, they have anxiety, and there's a whole group of comorbidities of epilepsy that also needs to be studied. Just to put in the perspective and then the case I described before; the child that we managed to help with seizure freedom, eventually developed depression and anxiety after the surgery. So, this is a comorbidity that we need to examine more and find out how we can deal with that.
07:55 Torie Robinson
Well, I understand you're working with somebody who's actually got a project about those morbidities?
08:01 Christos Papadelis
Yes, in my team, we have a principal investigator, Crystal Coopers is working with children with epilepsy comorbidities. And she's trying to understand the different kind of concepts of emotional processing in the brain of the children with epilepsy. We try to develop, now, more projects for these comorbidities. Also, we work closely with the clinical team here, Dr. Scott Perry, he's our clinical partner and we work closely together with the clinical team to help children both in the pre-surgical evaluation but also in the epilepsy comorbidities.
08:01 Torie Robinson
Fabulous, so you're looking at the overall quality of life of the person. You're not just looking at trying to stop somebody's seizures from occurring.
08:48 Christos Papadelis
To add on that, now we have new also ideas and understanding how the brain works in patients with epilepsy; we found out that in several of the cases where we don't know where the seizures they come from, it's really hard to identify the epileptogenic, what we call epileptogenic zone. Then now we believe that there's a whole network in the brain that is responsible for the generation of the seizures. The task that we have now is like not to identify where the seizures come from, but where they are located, what we call “pathological hubs” of these networks in order possibly to resect these networks, these hubs and the patient to become seizure free. To give it an example is like being in a busy area or a big metropolitan area and you have a traffic jam in the whole area and maybe sometimes this is because there's some constructions or some kind of an accident in a specific, critical intersection. And if this will be fixed, then the whole network will be fixed. So, in a similar way, we believe that in the epilepsy, there's critical hubs that if we are able to resect or intervene there, then the brain will not seize anymore.
10:03 Torie Robinson
Are you saying that there's the potential, at least in the future, to consider surgery for more than one hub?
10:10 Christos Papadelis
There are some cases that we know, for example, that the seizures come from one location or from multiple locations. But sometimes, these areas, they overlap with what's called eloquent cortex, which are very critical for the functions of the child, and that we should not resect. For example, if we resect or we intervene with the motor cortex, then there are potential motor deficits for the patient. But what we can do, we may be like, resect or intervene in the neighbourhood area that is not so critical for that but it's critical for transferring or bringing the seizures like in the whole network.
10:52 Torie Robinson
Before we move on - with thanks to EASEE®, by Precisis GmbH.
10:57 Torie Robinson
You have a paper-related “Machine learning on interictal intracranial EEG predicts surgical outcome in drug resistant epilepsy”. Tell us about that.
11:06 Christos Papadelis
This is a paper we published recently with one of my postdoctoral fellows here, Hmayag Partamian. So, this paper, it's about using artificial intelligence technology for helping the analysing the large amount of data we collect from patients with drug- resistant epilepsy. Usually, in several of the patients, we put intracranial recordings, we put electrodes inside of the brain in order to identify where the seizures they come from. And these recordings, they may last like four days or even like weeks, we're actually waiting for a seizure to take place. But this kind of invasive (because they also have some risks for infection, for other kind of injuries in the brain). So with my team, what we did, we analysed data between seizures, what's called interictal activity. And by using artificial intelligence tools, we try to use this kind of recordings in a fully automated way (without using the clinical team or the epileptologist spend time reading all this complicated data) to identify where the epileptogenic zone is located. And we found out that the section of this area can predict surgical outcome of the patient with drug-resistant epilepsy.
12:23 Torie Robinson
I was speaking to another person with an epilepsy, recently, who the moment I mentioned AI was like “Oh my God, I don't trust AI, I only trust a clinician/physician”. And so, could you tell us why we should trust that AI, and also, would the data or assessment of the data that you've just explained, would that be shared with the person's neurologist or epileptologist or neurophysiologist as well?
12:48 Christos Papadelis
I don't see the AI as a tool that will replace the clinicians or the epileptologist, but the AI tool can help us also guide us in a specific direction and be a really useful tool in the busy schedule like these people they have in their everyday life in the clinical care of the patients. When we have intracranial recordings from patients for hours or even days, usually we have a technician who is going through all these recordings, trying to identify where the seizures occur, trying to make clips of these recordings, and then provide this information to the epileptologist. By using AI technology, we're trying to help all this process, make it much more efficient, and also guide the epileptologist, in a sense, to see where the seizures they come in the brain. And AI as a tool is used in other kind of fields of medicine with really high success. There are some articles, for example, that, how AI, can use radiologists in identifying the cancer findings in lung X-rays, for example. In a similar way, I believe that what we can do with AI technology, we can help the epileptologist to identify patterns on the, either in the scalp EEG or in the recordings, where the seizures they come from, which locations they are in the brain, or these kind of networks that are responsible for the generation of the seizures.
14:28 Torie Robinson
It just can help us and find out things that humans don't see for instance. And then we assess all the data and the suggestions after the AI analysis has been made by AI.
14:40 Christos Papadelis
And also the AI, if I can mention that too, the AI is a tool and this is a project we're working now. It's not only important for identifying where the seizures come from, but also, as I mentioned before, also to identify these functional connections in the brain in order to be able to find these critical hubs. We have a paper published recently with one of my post-doctorate fellows, Ludovica Corona, on functional connectivity in the brain. Actually, by using MEG and high-density EEG, we're able to identify these critical hubs in the brains of the children. And we saw that resection of these hubs can predict surgical outcome and make the patient seizure free.
15:28 Torie Robinson
I'm thinking of my own personal experience and I wish I'd had access to this as a child. It's amazing that some kids can have this these days. Obviously, it's not available to everybody, but I think it gives us a bit of hope for the future, for the children of the future who will then become adults.
15:45 Christos Papadelis
I strongly believe that we have so advanced technology nowadays for helping patients and children with drug-resistant epilepsy. Neuroimaging techniques; they have advanced significantly the last two decades. We have new software, new tools that can help the clinicians and the patients for helping this population. But what we're lacking now, we're lacking funding on how to this kind of translate this knowledge that we have from the technological aspects into the clinical practice. There's a gap there. And then we need help on how to take this knowledge that already exists at the technical level and help by using this knowledge how to help the children with drug-resistant epilepsy become seizure-free and also like not having all these comorbidities they struggle with epilepsy.
16:37 Torie Robinson
Getting it to the people that need it, to the patient, you know, it's, it's otherwise, what's the point?
16:44 Christos Papadelis
Yeah. Something I've learned, like, I'm a biomedical engineer (that's my background) and something I have learned in my career so far is that biomedical engineers; they have really great knowledge about the tools and the methods, clinicians; they know very well how to deal with the patients but the way how these two groups interact is somehow not efficient. We need more work to… how they talk to each other, what is the importance of the different problems, and how they can eventually, all these tools, they can be translated to have clinical impact to the patients.
17:28 Torie Robinson
A huge thank you to Christos for sharing his expertise with us! One of the most striking takeaways here is that we’re moving beyond simply asking “where do seizures start?” to understanding the networks and critical hubs that sustain them – and that shift has real implications for surgical strategy and outcomes.
Christos has also highlighted for us a major gap – not in technology, but in translation. The tools exist but the challenge now is getting them into clinical pathways so that more children can benefit earlier - when their neuroplasticity offers the greatest opportunity.
Thank you also to EASEE®, by Precisis GmbH, for supporting this episode.
If you found this episode helpful, please give it a like and subscribe, and hit the bell so you’re notified when new episodes drop. I’d also love to hear your thoughts or experiences in the comments below - because I do read them! Thank you for joining us, and see you next time.