Seizure-Predicting Headband: AI That Warns You Hours Before It Strikes | NeuroHelp

 What if you could know a seizure was coming hours before you felt it? NeuroHelp co-founder Orin Shriki, PhD, built an AI headband that reads your brainwaves while you sleep and predicts seizures — no hospital visit required. It's already used by Israeli Air Force pilots and truck drivers for fatigue monitoring. The goal: a brain thermometer in every home. 

_______________________________________________

Your life is better because of Israelis you've never met. 

⭐ Loved this? A 5-star review takes 5 seconds and is how new listeners find us. 

🎧 All episodes + every platform: pod.link/1585604285

📺 YouTube: @israelitrailblazers | 📧 Substack: findinginspiration.substack.com 

Transcript

GUEST OREN 0:00

We can now predict seizures a few hours in advance just based on the EG data.

HOST JENNIFER 0:11

For decades, seizures felt like lightning strikes. Terrifying because they seemed random, but it turns out they're not. They could be predictable, and that changes everything. I am Jennifer Weissman, and this is the Israeli Trailblazers Show, where Tiny Israel solves massive problems for mankind while the news shares the clickbait stories. Today we're speaking with Professor Orin Shrikey, who has developed this incredible headband called Epiness that sees seizures before they happen. Because the best crisis is the one you prevent. Welcome, Professor. How are you?

GUEST OREN 0:57

Hi, Jennifer. Doing great.

HOST JENNIFER 0:59

What is the coolest thing you've seen someone do with their brain?

GUEST OREN 1:03

You can draw just by thought, draw paintings, control robotic devices. This is really cool. We do it in my lab, students controlling drones, controlling robots, doing very cool stuff just with your thoughts, just with the brain. That's incredible. In my lab, we use a non-invasive means of recording brain activity and reading them out in real time. That's really cool.

HOST JENNIFER 1:26

You have developed the Epiness headband that is essentially a way for the brain to predict a seizure. How does that work?

GUEST OREN 1:38

We established a startup company called NeuroHelp.

HOST JENNIFER 1:41

Now put that on again. This is a very Richard Simmons-looking headband. Remember Richard Simmons?

GUEST OREN 1:46

This is the headband that we developed. We're working on more advanced versions and just show how it looks on the inside. You have electrodes that can penetrate the hair and these flat electrodes on the forehead. And so this headband really measures what we call EEG, electroencehalography, the electrical activity of the brain. Non-invasively, it's very easy to put. It doesn't stimulate the brain, it's just passive monitoring of brain activity. We have algorithms running in the background that we developed in the company. They are based on our research at Ben Gurion University. We can now predict seizures a few hours in advance, just based on the EG data. The hope would be to notify people living with epilepsy and their family and caregivers in advance so that they can take countermeasures. They can take medications that can prevent decisions in the first place. And of course, it's important for safety. Just don't go to risky activities like hiking, biking, swimming, driving. Driving, of course, people with epilepsy they cannot drive for at least a year from the last seizure. Our goal is really to reduce the uncertainty that accompanies the seizures. You never know when it's going to get you, like in the mall, in some activity with friends. The goal is to really predict the seizures in advance, but also to even prevent them. And it's also very important to detect the seizures. Sometimes it's hard to predict. Not all seizures are predicted. It can vary from one person to another. So we can also detect them in real time. That sounds more basic, but it's still a very tough task to detect seizures and discriminate them from other types of activity in the brain. And for parents, the children with epilepsy, this is critical. They effectively don't sleep. They sleep with the children. They are always at the fear that their child is going to experience a seizure soon. So for parents, it's very important to know about seizures as they happen and, of course, know about them in advance.

HOST JENNIFER 3:40

So you wear this headband, you go to sleep. What is it that your algorithm can detect? A change in the brainwave? Or what exactly is it that identifies, oh, this person is about to have a seizure?

GUEST OREN 3:56

It's not one thing. What we get is raw data, just indeed brain waves. That's the input. Just brain waves going up and down from different channels of this headband.

HOST JENNIFER 4:06

Yeah.

GUEST OREN 4:06

So then we extract meaningful measures that quantify the underlying dynamics of the brain, many measures of brain dynamics from the different channels. And then we use AI to process all these measures. So it's a very high-dimensional picture. If you look at the seizure as the seizure is going on, you will be able to see it in the EG. It has quite a clear signature. It's not a trivial task. Even for neurologists, epileptologists who are well trained, they do not always agree about a seizure. But if you ask them prior to a seizure, even a few seconds or a minute before seizure, they wouldn't see anything. So in order to predict seizures, you really need this kind of high-dimensional seizures. And it's only through AI tools that you can predict seizures a few hours in advance.

HOST JENNIFER 4:49

You would start with the indication of seizures. However, there are other uses of this headband, it would seem, like pilots who are gonna fly and maybe they're too tired to fly, or truck drivers who are on long hauls and it's too much for them. What are some other indications one could use for this eppiness headband?

GUEST OREN 5:12

There are many applications. Our grand vision is what we call remote neurology, to have this device at every home, like you have a thermometer, and be able to monitor your brain from time to time to identify deviations that may mean that some neurological disorder is developing. One of the applications is for sleep deprivation, fatigue. It's very easy to put this kind of a headband and you can monitor the truck drivers, doctors. We also had a project in my lab with the Israeli Air Force in recent years trying to monitor pilot brains under different circumstances, for example, to prevent loss of consciousness. When fighter pilots do high G maneuvers, they can lose consciousness. It can end very tragically. And if you can detect this spatial disorientation and, of course, fatigue, cognitive workload, and other things. In terms of neurology, we are working on early diagnosis of dementia, like Alzheimer's. You can detect Alzheimer's through brain activity in advance of being diagnosed with Alzheimer's? Several studies in recent years suggest that there are indicators in brain activity that of an upcoming Alzheimer's years before it really starts to manifest. So it's still a very difficult task. And most of the studies they use cumbersome technologies like MRI, which means that you have to come to the hospital and do a full scan in an MRI machine. Our goal is to really allow these kinds of things to take place at home with this headband. So it's very easy. You just sleep with it, say once or twice a month, or maybe put it once in a few days for a few minutes. So you could record your baseline activity and then identify deviations. And of course, it will all be processed by a trained neurologist that produces these early markers of different disorders. The goal is to have this kind of personalized diagnosis. So we really know your brain, how the baseline of your brain activity looks like, so we could really identify substantial deviations, and then eventually we can close the loop. So we talk about also personalized medicine, personalized treatment. That's the goal of this, what we call remote neurology, to take everything outside of the hospitals, do as much as we can at home in terms of early diagnosis and closing the loop by treating in a personalized manner so we can really bring it to everyone. And our device really produces clinical grade EG data and it's affordable. And that's the goal to do everything as much as we can at home.

HOST JENNIFER 7:40

Remote neurology. That's the theme. That means you take control of your own health care, you monitor yourself at home, you don't go to the hospital, you don't involve your doctor until you see a deviation or a change in the marker. I assume you have an app on your phone.

GUEST OREN 7:59

That's correct.

HOST JENNIFER 7:59

This is amazing. How did you come up with this idea?

GUEST OREN 8:02

My research centers around transitions in brain activity. For example, what does it mean to experience a seizure? It's a sharp transition in brain activity. And some networks in the brain, they switch into oscillations, into synchronized activity. It's a very sharp transition. Another example of transition is hallucination. Like when people hallucinate, it could be in schizophrenia, when people hear voices in their head, tinnitus, like you have this ringing in the nose, or due to psychedelic. What really happens is again some kind of a very sharp switch or transition in brain activity. And I was working on this using measures from physics, actually, from the physics of transitions between phases of nature, like between liquid and solid, magnetic and non-magnetic phases. The thought was that this would be the lowest-hanging fruit, epilepsy. That we can identify useful, efficient markers of epilepsy using these ideas that we bring from physics. And that's how we started with epilepsy. But then gradually the vision grew up, and our major goal is really remote neurology.

HOST JENNIFER 9:08

So remote neurology as a whole, but the out of the box, this product, epiness, is a headband to predict a seizure.

GUEST OREN 9:17

Yeah, we developed a platform. We have the platform in terms of the algorithm. So any EG data that we get could be from Parkinson's, Alzheimer's, migraines, we can train our algorithms to detect the relevant unique patterns and then generate relevant markers, early diagnosis.

HOST JENNIFER 9:34

Incredible. And do physicians welcome this new technique, this new home-focused personal headband that can assess all these problems?

GUEST OREN 9:46

We are working closely with epileptologists, a neurologist at hospital. So our RD team meets with them on a weekly basis to go over the data, to understand the decisions of our machine learning algorithms, to see if they had false alarms or misses, why did it happen, and can we fix that? So we are working very closely with them and are really waiting for this kind of technology because they know it would really help their families and it could also reduce the burden.

HOST JENNIFER 10:15

I would imagine the upstream cost to someone actually having a seizure, what it takes to manage that patient after the seizure, aside from the family burden, having nurses and doctors and hospital stays and other sorts of tests, where if you could alleviate all of that, that would be a tremendous savings.

GUEST OREN 10:36

You can improve adherence to medication, reduced hospitalizations, save leave days. So we can really be very helpful to the system, but it wouldn't be easy.

HOST JENNIFER 10:46

Yeah. Remote neurology. I love it.

GUEST OREN 10:49

It's now under clinical trials in Israel, and we hope to soon start full clinical trials for FDA clearance in the US.

HOST JENNIFER 10:58

You mentioned it should be like a thermometer, that this headband could be in everyone's home, detecting maybe sleep apnea, Parkinson's, Alzheimer's, a whole host of sad, debilitating problems that we all may end up with.

GUEST OREN 11:14

This is really the next challenge of mankind, really understanding brain-related disorders. We don't understand all of them. It's about the combination of understanding, deep understanding in neuroscience, and understanding in machine learning, AI, that really allows us to create these tools.

HOST JENNIFER 11:31

Epilepsy, it haunts entire families. But what if seizures become predictable? This future isn't theoretical. The epiness headband is being used and tested in Israel right now to help mankind. I'm Jennifer Weisman, and this is the Israeli Trailblazers show, where a nation smaller than New Jersey tries to solve big problems facing all of humanity. Unfortunately, the media only pushes conflict. That is why we exist. We are sharing Israeli contribution. Please hit subscribe and please post this on all of your socials. Until next time.