Neuroscientist at the Yale School of Medicine, Dr. Hal Blumenfeld, MD, PhD talks about a new way to determine if people with epilepsy are suitable to be behind the wheel.
Interview conducted by Ivanhoe Broadcast News in 2022.
For starters, when someone has epilepsy, most of us have heard that term and have a basic idea of what’s happening, but can you explain for our viewers what’s going on in the brain. What are some of the concerns or considerations that a person with epilepsy then has to consider?
BLUMENFELD: When people have epilepsy, they have seizures. These are episodes that can happen unpredictably at any time and interfere with people’s ability to do things like respond normally to people around them, participate in school or work and safely drive a car. And because these episodes can happen unpredictably at any time, people can’t plan for them and it can interfere significantly with their quality of life.
Now, to my understanding, there are different levels of epilepsy, some of which can be controlled with medication and some can’t, is that part of what you and your colleagues looked at in your current search?
BLUMENFELD: Yes. Seizures can be controlled with medications about three-quarters of the time successfully, which is wonderful, and there’s other treatments too when the medications don’t work. And people sometimes outgrow seizures, especially certain types of seizures called absence seizures, where young people, children, or teenagers, brief staring spells, they stop and stare and can’t respond for a few seconds. People often will outgrow those or can be successfully treated with medicines that completely control the episodes as far as people can tell. The challenge comes when we do other tests- brainwave tests called EEG, where we record the electrical activity in the brain to see if there’s any hints or clues that subtle seizures might still be happening even if people or their family members or friends don’t know they’re having these absence or staring attacks. There could be subtle clues in the brainwave that might say they could be still going on. And we have a real challenge when people don’t think they’re having these episodes anymore, their family members don’t think they’re having episodes anymore, but the brainwave EEG signals suggest maybe they are. And the problem is right now, doctors don’t know for sure if those people should still be treated with medicines or whether it’s safe for those people, for example, to drive a car if they are not having any obvious seizures but their brainwaves show these worrisome signs.
When you’re talking about a task like driving a car, why is it important to really have a handle on whether or not these momentary seizures or a couple of seconds seizure are still happening?
BLUMENFELD: As we know from driving safety rules about texting and driving and other distractions, alcoholic course in driving, even a momentary lapse for a few seconds in attention can be very dangerous when you’re on the road. And if you were to have a brief absence seizure attack where you temporarily can’t respond even for a few seconds, that will be very hazardous for driving. And fortunately, like I said, a lot of times people outgrow their seizures and other times medications work to control seizures, but if people still have brainwaves signals called spike in wave discharges, which is what we see on the brainwave test, that signal possible subtle brief episodes, it’s hard to tell if people can respond or not unless you do special behavioral testing during those episodes. We have ways in a research setting of testing people’s behavior with split-second timing, showing them something on a screen and having them react quickly during the actual spike in wave discharge to see if it is interfering with their behavior. That’s the best way to test it. The problem is that that specialized research, behavioral testing can’t be done most of the time. It’s really hard to get the split-second timing. And most medical centers and most doctors offices don’t have that ability to do that, it would be very hard. So we’re stuck in a way to make decisions right now. In most the situations, doctors have to decide just based on what people tell them about their symptoms and based on judgment looking at the EEG, looking at the brainwave activity, and they have to make their best guess about driving or not for these people. And either way, it’s a problem. If someone is told not to drive and maybe they really could be safe for them, that has a big impact on their quality of life. Driving is really important for people functioning in the world. On the other hand, of course, if someone is given the ability to drive, and God forbid, it’s not safe for them, that is a life or death situation and has an enormous impact, not just on the person driving, but on other people out there in the world. So it’s an important decision, and right now with available technology and available medical care, it’s a decision that we don’t have complete information for a lot of times and it’s left to clinicians discretion to decide.
So can you tell me what you and your colleagues are looking at that in your mind may be a more accurate way of determining whether or not a person can go back to driving or continue driving?
BLUMENFELD: It’s very exciting we’ve used the power of machine learning or artificial intelligence to gather clues from the brainwave EEG activity and decide objectively if there are safe or unsafe patterns on the EEG. And so what we did is, we took data from a large group of patients at three different sites, so we gathered a big amount of data and we took that special behavioral testing where we do know for sure if people can or can’t respond during the spike-wave episodes. Where we did behavioral testing and we tested people during the spike-wave and said, look, can this person respond or not during their spike-wave discharges? And some people can and some people can’t, and sometimes even the same person. There are some other spike wave discharges where they can respond and somewhere they can’t. And we fed all that information into the machine-learning algorithm. The teacher basically, the teacher to tell the difference between brainwave activity that’s safe and brainwave activity that’s not safe for driving and for responding. And that was the main strategy here was to teach the artificial intelligence algorithm to know the difference based on their brainwave alone. Is it safe or not safe for this person to drive? Go ahead.
How accurate is the machine-learning in your study?
BLUMENFELD: So we want it to be very safe and never tell anyone that they can drive if they might possibly be unsafe for driving. So we adjusted the algorithm and made it so that in fact 100 percent of the time, if the algorithm says it’s safe for someone to drive, it was safe for them to drive. We never had a case of a situation where the algorithms said, this person is safe, they’re not impaired and in fact, when we looked at the data to see what the behavioral testing they were impaired. It was very good at doing that. But because we made it so conservative, it means that there are some people where it said, sorry, you can’t drive. And it’s possible maybe they would be able to drive if we did the behavioral testing. So what this means is that right now, if you took, we looked at a group of 23 people that a fresh group of people that weren’t part of the original teaching the algorithm approach, a fresh group from another site and we said, let’s try this out on 23 people and see how well this works. For those 23 people 100 percent of the time, it was correct in saying when it’s safe to drive. But it only identified about one-third of the people who could have been safe to drive to drive. In other words, it gave a third of people a chance to drive who wouldn’t be able to otherwise. But still less two-thirds of people who were still stuck and without this, without the extra special research testing which isn’t available that they still can’t drive. So we’re taking steps in the right direction, but there’s still more to go.
What needs to be done so that this is a tool that can be used pretty readily for evaluation?
BLUMENFELD: A few things. We need to do this in bigger groups of data with more people and really confirm that first of all, when we tell people it’s safe to drive, it really is safe. So far it’s very encouraging. We’ve got 100 percent safety record, which is good. But we want to also increase our ability to know that it is safe for those people that right now we can’t give an answer to. So like I said, it’s only about a third of the time we can say to people, yeah, you can safely drive. We need to get a bigger group of data and more people to test out these algorithms and fine-tune it and make it really practical. The other thing of course, like anything of this kind it’s a research full right now, it’s not FDA approved yet, it’s not clinically available. It needs to go through the whole validation process to make it user-friendly, to make it something that could be not just used in research setting, but by using any doctor’s office ultimately so that it’ll give real answers to real people out there in the world.
How important is it in your mind to be able to extend, to be able to get the approvals and get this into the doctor’s offices so that it can be used nationwide, maybe worldwide?
BLUMENFELD: Yeah, that’s our hope, is that this kind of research always starts out. First steps. I’m going to research setting with smaller groups and it looks very promising and then we have to bring it to a bigger group of people, test it out, test safety, and take next steps, partner with industry perhaps, partner with others who can support this work and bring it into the public sphere. And ultimately this is for public safety, so it needs to be approved by the FDA and needs to go through the process. We need partnership with industry to make this commercially available. We’re academic researchers, that’s not my forte. This is not about anything commercial, it’s about helping people, but to make things really work in the real world that’s going to be necessary.
Is there anything that you would want to make sure our viewers know about your research?
BLUMENFELD: I think it’s important to know that epilepsy is treatable. People with seizures can get their seizures under control and live hopefully good, normal quality of life. Driving safety is a very important issue right now for people with epilepsy. Right now, the guidelines are based on our best guesses about who is or isn’t safe to drive. But there’s more work to be done to help people drive safely when they can and cannot drive when they should not be driving. There’s important work and we’re on the forefront of that work right now as I think these exciting studies.
How far away in your best guess would it be before this technology would be ready for that FDA approval process and then distribution?
BLUMENFELD: I think five years is a realistic horizon. We’ve been doing this for many years already, developing the methods, understanding this type of epilepsy we’ve been working on for years. We’ve learned a lot about what does or doesn’t impair consciousness during seizures. We’ve learned a lot about the brain networks that are effective when people lose consciousness, their seizures. We’ve learned a lot about the signals in the brain and how they differ when someone does or does not lose consciousness, their seizures. We found that there are differences in the brainwave activity that we can pick up and that’s what this algorithm is built on, is all that knowledge that we’ve been really building for decades now. I think we’re really at an exciting point after working on this for many years where we’ve found through this study the first real practical application and then a small group, but looks very promising and having done other type of work recently where we translated basic research into the human realm, I think five years is a realistic timeline to bring this to the next level, which like I said, involves more people and getting the regulatory approvals and making this practically available.
END OF INTERVIEW
This information is intended for additional research purposes only. It is not to be used as a prescription or advice from Ivanhoe Broadcast News, Inc. or any medical professional interviewed. Ivanhoe Broadcast News, Inc. assumes no responsibility for the depth or accuracy of physician statements. Procedures or medicines apply to different people and medical factors; always consult your physician on medical matters.
If you would like more information, please contact:
Colleen Moriarty
(203) 376-4237
Sign up for a free weekly e-mail on Medical Breakthroughs called First to Know by clicking here
