Dr. Guger shares BCI's evolution
Host:
Mouser continues its exploration behind the technology of brain computer interfaces or BCIs with Dr. Christoph Guger, Founder & CEO at g.tec medical engineering. He joins In Between the Tech today to share BCI's evolution and his company's role in the advancements of neurotechnology.
Thank you for joining us Dr. Guger on In Between the Tech. Please tell us about g.tec medical engineering and the work you’re doing?
Christoph Guger:
It started actually when I was at Johns Hopkins University in the USA studying. And one professor over there told me about Gert Pfurtscheller from Austria who was actually a professor of my home university in Graz. And he developed brain compute interface technology 25 years ago already - longer than 25 years ago. So I returned to Austria and started my master thesis and also PhD in his lab. And I built the first real-time brain compute interface. So this means we could control in real time a cursor on the screen. Right after the finish of my PhD, I established g.tec medical engineering with Günter Edlinger and we started to sell BCI technology for researchers and universities. So they used the technology to develop the BCI technology to design new applications or to do basic neuroscience research. This is what we are still doing. So, we are selling to many different research centers or technology. So, this includes also the big ones like Meta and so on, but also companies in aviation like BMW or Airbus, Boeing, Air Forces from different countries. So they're all using it for making aviation better. And of course many people are working on medical applications to treat patients who are to have better assessment tools.
And about 10 years ago we started with recoveriX. So, this is our neuro rehabilitation system. At this time we used it only for upper limb treatment in stroke and we figured out that the stroke patients, even in a laconic state, did get nicely better if they just train for a couple of hours and sometimes it's life changing. So people cannot really move the hand, they train for 18 hours and then they can use it again. So one of our first patients for example, was a hairdresser. She was pretty young, like 30 years old, had a stroke, her right hand was paralyzed, and she had to shut down her hairdressing institute and all the doctors told her this will be, for her, until the rest of her life. So, she was unemployed, the social security system paid her salary and then she came into a clinical trial with recoveriX, she trained for 18 hours and already after three, four sessions the hand functions were coming back and after the recoveriX session she was opening her hair dressing institute again and she started to work again. This is exactly what we want to achieve to bring people back to work or to life. And one outstanding result was that this works even for patients who had a stroke 31 years ago. Because normally everybody here after one year there's no progress anymore. But we actually figured out there's no impact on the years after the stroke. So it's independent how long ago the stroke happened, and this is very positive of course for all the stroke patients. And then we started to treat lower limb functions because many of these patients have also problems when they're walking. This worked even better for the simple reason that for walking, you need always both legs to be working together. So you see the effect much better than for hand movement because for hands you can always compensate with the healthy one.
And then one neurologist, Tim von Oertzen was asking us, why are you not trying it for multiple sclerosis? So we tried it also for MS and we figured out it's working even better. The reason is that the lesion in the cortex is smaller and sometimes it's really difficult even with magnetic resonance imagining to figure out the lesion. And interesting is that recoveriX is also working for these patients because in this case the myelin of the axons is destroyed by MS and with recoveriX we are producing neuroplasticity. This means all the damaged neurons are replaced by healthy ones. So they're reconnecting and forming better connections so that the fine gross-motoric skills and so on improve. And, very interesting for multiple sclerosis was also that the fatigue gets much better. So, these people are exhausted after a few minutes of activity. In Vienna for example, we had a language therapist and before recoveriX she could treat children for one hour and then she had to go to sleep because she was so tired. And after recoveriX she could treat six children after in a row and afterwards she went even shopping because it was possible again. So the life balance is just much better after doing it.
And at the moment we are running a clinical study for Parkinson's. This is also very interesting because multiple sclerosis or Parkinson's patients are traditionally declining so they get worse and worse and doctors give them medication to slow down the decline, but there is nothing which makes them better. And with recoveriX we found actually something that makes both better, multiple sclerosis and Parkinson patients. So it's the first tool that makes them really better. This gives a very nice option to neurologists to help patients. We had for example, a Parkinson patient, she was still pretty good but walking a little bit slower, but when she was playing dart, the tremor was a handicap, so she didn't play so precisely and after recoveriX, now, she hits the 60 almost all the time and is beating her husband all the time. And the husband, of course, nowadays is not happy anymore, but is exactly what we want to achieve.
Host:
How has BCI evolved since your company was founded in 1999?
Christoph Guger:
Yeah, in 1999 we were basically moving a cursor to the left and to the right hand side. And people found this already fascinating. A few years later we came out with intendiX. So this was the first P300 spelling device. This sounds very technical, but it was a brain compute interface that allowed locked in patients, so people who were completely paralyzed, to select letters on the screen and to form sentences and to communicate again. So at this time such a system was costing around 13,000 euros - pretty expensive. A few years later we came out with the Unicorn brain interface. So this is a system that we are still selling. It's costing around 1000 euros, it includes the same P300 speller and this is now already very affordable solution for these locked in patients. So the technology became much smaller and cheaper thanks to companies like Mouser who provide us also with very nice and better, more cost effective and smaller electronic components so that we can build technology like that.
Host:
We can immediately think of medical applications for BCI, but could you give us other examples of where this technology might be used?
Christoph Guger:
Gaming is, for example, an important application. Our Unicorn brain interface that I just mentioned before is also sending data into Unity. And in Unity you can very nice build 3D environments and computer games. And we are also developing this technology for the maker scene or for engineers so that they can take the brain compute interface and plug it into games. So we are taking care of all the hardware, so all the signal acquisition of the brainwaves, all the signal processing, and then they can just place BCI controlled elements into a game and they can also combine it with keyboard control or a joy stick control and in multiplayer gaming. And, of course, there are big game companies who are also people specialized on games so they can fully utilize nowadays BCI control, which otherwise would be really difficult to do everything yourself. So, a game developer doesn't know how to analyze brainwaves and vice versa.
So, with such a toolbox it will spread out much more quickly than for medical applications because everybody likes to play computer games, of course. In this case, it's just very important that the brain compute interface gives you one, unique control option for the game because keyboards or joy sticks are highly optimized for information transfer rate, but they are blind of what's going on in your brain. With the BCI technology, you can actually extract that. So we can find for example, an engagement index which shows how engaged your brain is and then in your game for example, you can only compete or reach the next level if you're actually cool enough and not too nervous. So it gives completely new options for game developers.
Host:
We’ve discussed various applications and uses for BCI technology. What breakthroughs have you witnessed while working within this technology?
Christoph Guger:
Yeah, costs and size of equipment is of course a very important aspect. When I started '95 at Johns Hopkins University, I got EEG data on a music tape, which I found pretty funny and then I had to digitize it to get it into MATLAB. So it's not so long ago that people were using music tapes. So a few years later we could read the data directly into computers in digital form. So this means all the ADCs were included in the bio signal amplifiers, and you could actually put it on the head. In 1999, bio signal amplifiers were pretty big boxes and when we went to the first university conference, we put actually a bio signal amplifier for a BCI system into the floppy disc of a laptop computer. And this was pretty innovative at this time because it was just much slimmer and portable, lightweight. So this means you could use a computer at the patient's home, for example, do allow them to spell with the brain compute interface.
Then a big step was wearable technology. So, the technology became smaller again. And nowadays, EEG headsets are fixed on the head then you get small or short electrode cables to the electrodes itself, and this made the EEG recordings much more stable and better because the whole system is used moving with your body. And for that reason nothing is pulling on the electrodes like electrode wires, which were 1.5 meters before. So, the data quality improved dramatically.
And one big step was also that we put the amplify directly into the electrode. So, the amplification was as close as possible to the brain and this was also a big step forward in brain compute interface accuracy. So it made everything much, much better than with the long wires and amplifiers far away from the subjects. And nowadays it's getting even slimmer and more lightweight.
So just this week we released a new tool that's called Unicorn BCI Core-8, and the whole brain compute interface is now 24 grams. So, it's a very small box sitting on your head. And with this unit we can also do fascinating work for example with dog or horses because it's so small that you can also put it on animals and then you can also study what's going in their brain. And people are also working with dolphins because they have a very unique acoustic system. So they want to figure out how the brain is doing all this acoustic processing. So there are many innovative applications coming out.
Host:
Please tell us about g.tec’s recoveriX system for the neurorehabilitation of stroke patients and patients with Multiple Sclerosis.
Christoph Guger:
The recoveriX system has electrodes, exactly where the sensory motor cortex. So this is the cortical regions where motor movements are planned and executed and we're asking people to imagine a left hand movement, a right hand movement or a foot movement, and the brain computer interface is able to pick it up when people are imagining this type of movement. And then we trigger in real time an electrical stimulation of the corresponding hand to a foot. So this means we are pairing cognitive processes with motor behavior again, and this is what people really love. If they were paralyzed for a couple of years and then they imagine the movement and the hand moves again, this leads to a lot of motivation. We are doing 240 movement imaginations in one therapy session and people are coming back 25 times. So in total they have to do 6,000 movement imaginations. And this is as often as a child needs to learn to walk and this is essential for neuroplasticity. So you need many, many repetitions of the same task to make people better. And after these 25 sessions, we can say that like everybody improved, this can be a lot, but it can also be just a little bit. But on average we get quite good improvements. There are only a few patients that don't improve, but there must be another neurological event like people get another stroke and otherwise people get better.
Host:
What do you see as the biggest hurdle to the adoption of BCI systems? Where is the resistance?
Christoph Guger:
The resistance is basically that people don't know about it. And if you have patients and you see a new medication or a new treatment opportunity on the Web, nobody has done it before, then you doubt it because why should it help you? So we have to create awareness and we did it actually by establishing our own rehabilitation centers where we are treating patients with our own physiotherapist or occupational therapist and then people were telling other patients that they improved. This is actually the most important marketing that we can do - satisfied patients who recommended it. And in this way it's also spreading out to hospitals, rehabilitation centers, and doctors, physicians that the patients are getting better, but you need a lot of these patients to convince hospitals and other people that this really helps.
We established also a franchise system. So this means a businessman or just a physiotherapist can get all this neurotechnology. The recoveriX system including a training in the standard operating procedure takes a couple of days to train them because it's a medical product and then they're allowed to use it with patients and this is nicely spreading out. So we started here in Schiedlberg, that's a very small village in Austria and we treated already more than 30,000 recoveriX sessions, so it's a lot. And then the next center was in Linz, this is like 20 kilometers from here and the next center was in Vienna and now we are in Munich and Passau. So it's spreading all this out from where we are starting because people see the improvements and then they come to one of the centers and then other physiotherapists get interested and it spreads out. So we started also a couple of years ago in Thailand in Bangkok and this was pretty successful. So they had nice improvements of patients and now they have a big investor who wants to establish 168 recoveriX centers in Thailand so that every patient can actually get the nice treatment in a short distance.
Host:
What role do you see AI playing in BCI?
Christoph Guger:
For BCIs and brain compute interfacing, we are using - since the beginning - signal processing techniques like machine learning or also deep learning, AI to train the brain computer interface on the EEG data of every single person. And this is important because brains are pretty similar but in detailed movement imaginations and other sorts at different regions. So we have to identify how these EEG patterns look like and where they're located and for that, of course, signal processing techniques like AI, deep learning, machine learning are super important because you cannot just see it with your eye when you look at the data. So you need signal processing to be able to do it. For recoveriX, of course, we are training the brain compute interface on every single person. Same also for the system that we use for coma patient, but we are also uploading all the data that we are acquiring to a server. So we have now 30,000 data sets just from our own recoveriX center on the server and we deep learning, we can analyze the data, then we can figure out for example, who is improving more and we can also quantify what's ongoing in the brain, so how much changes we can produce and also which changes in the brain are important so that fine and gross-motor skills are coming back. And with these techniques of course we can figure out important information to make the therapy better. Our long-term goal is also to have a marker in the data so that we can produce maybe in future warnings to patients if there's a risk, a neurological risk for them so that they can see a doctor as quick as possible or if there's a new treatment option occurring so that we can inform these people that they should do it.
Host:
What is the one application for BCI systems that you would most like to see become a reality?
Christoph Guger:
There are so many because, at the moment, in most of the hospitals, doctors are just looking at EEG data. So they have the waves on the computer screen, but there's not so much what you can extract by just looking at the data and there are so many important applications which are all good for neurological patients. So we just mentioned a few of them.
We had also a patient with Cauda Syndrome in Italy and this man was paralyzed within four months. So, he was diagnosed at the first problems. Four months later he was paralyzed. So that's autoimmune illness affecting also the myelin of the axons and there's no treatment option actually available. And then our neurologist, Rossella Spataro, did the recovery training with the patient and he recovered within a couple of hours again so he could move the hands again, he could talk again and otherwise he would be paralyzed at the moment. So, it has a dramatic effect for rare disease, which makes it, of course, super important to have this technology available. But this is all in the medical space.
For non-medical applications, I think the gaming is a very important aspect. Then you can play computer games and beside that you can extract information of the EEG, send it to a server, and you can also analyze it. And while you are watching, for example, TV or are playing a computer game, the deep learning algorithm is analyzing EEG data and you get the warning that maybe this illness that I mentioned before is developing in your brain and you should see a neurologist as soon as possible. So these medical assessments can become a side product of entertainment.
Host:
What has been the biggest ethical concern regarding the development of BCI systems?
Christoph Guger:
Of course there are very strict guidelines in the medical industry, so you have to model, follow the normatives from medical equipment, and these are the strictest normatives around the world. So even for aviation for airplane, the normatives are not so strict as for medical products, for the simple reason that you don't want to bring a patient at risk who is already ill. So, medical equipment is very well controlled and designed. This is also important and there are already a lot of very good normatives, just people have to follow them. There's also one for cybersecurity for example, so that nobody can access the data that you're recording. But of course, a big question is what happens with your EEG data if it's transmitted to a server. Maybe your data is sold to companies and they're extracting whatever out of the data. So this is something that must be handled carefully.
Host:
How does pairing EEGs with fNIRs open up new avenues for BCIs and neural research?
Christoph Guger:
So EEG recordings are measuring electrical activity of neurons. So it's just a voltage measurement that we are doing and these signals are very fast. So, if you imagine something that change within a few milliseconds, and this is a big advantage if you want to control something. The fNIR is functional near infrared spectroscopy. So in this case we are shining light into the brain and the blood where the hemoglobin of the blood is reflecting the light, and then we can estimate how much blood flow is in a certain brain region. It's similar to the principle of a magnetic resonance imaging machine. It's just portable. So you get little LEDs or lasers on your scalp and then you can estimate the blood flow.
This technology is long lasting, so the reaction time is not very fast, so it's more like every 10 seconds, you can see a change. So, this makes this effect why it's interesting to combine it with EEG, because then you get short lasting brain activity and long lasting brain activity into your brain compute interface and you can analyze both at the same time. So, a finger movement for example, takes in your brain, a few hundred milliseconds and then everything shuts down. In the fNIR signal, there are biomarkers for example, for pain that you can detect with fNIRs, but you wouldn't see it with EEG data because it's a long lasting change. That's one of the examples why the combination of EEG and fNIRs is a good one.
Host:
Can you give us an overview of your cortiQ system and how it's used in brain surgery?
Christoph Guger:
So our cortiQ system is a so-called high gamma mapping system. High gamma refers to the frequency range that we are extracting of the EEG data. So, traditionally people are looking at alpha, beta ranges, so this means between eight and 30 hertz or lower frequency in the EEG because this basically what you can see with your eye. The amplitude is pretty big and when you look at the traces you can see something. The high gamma is between eight to 200 hertz or even the one kilohertz and it's so tiny that you wouldn't see it with your eye. This again, the reason why we need signal processing techniques to extract it, but a high gamma has one very important fact. So, if I move the finger, it occurs immediately, and I can very nicely detect it.
And this is what we use in cortiQ. We instruct the patient to move for example, the thumb and a few milliseconds later we can already see where the region is on the cortex and then we can mark it, and then we instruct the patient to move the pinky finger and then we see exactly where the pinky finger is located. And then we read for example, an English text via the loud speaker of the computer, the person is just listening and a few seconds later we know where the auditory cortex and Wernicke's area is located. Wernicke's area is responsible for understanding language or if we want to figure out where POCUS area is located, we show an image to the patient and the patient has to name the image - car, flower - and whenever the person imagines the word the brain, the POCUS area is lighting up. So, this is expressive language area and then after a few seconds we have located it. So we can create a map within a few minutes of the most important centers of a patient. And this is very important information for the neurosurgeon. If he cuts out, for example, the finger region, the patient cannot move the fingers anymore. And this is of course something that the neurosurgeon wants to avoid. This is, at the moment, done with electrical cortical stimulation. This means neurosurgeons are implanting 256 electrodes all over the cortex and then they're applying electrical currents between all pairs. So, they're stimulating electrode, one versus two, one versus three, and so on. And whenever the patient shows a reaction, then they write down this is the finger region, this is the language region.
This mapping procedure takes normally one or two days and it's very demanding for the patient and for the neurosurgeon, with cortiQ, we can do the same, but much faster, so we need like four minutes for getting the result and then the mapping procedure is over and the patient can relax again. There's also another very important fact on the temporal base - this is more or less the bottom of the brain. You cannot really do electrical cortical stimulation because it's difficult to access because you have all the bones around this region. But with cortiQ, we can also extract information from there. And the face area for example, is located on the temporal base. And some years ago neurosurgeons were resecting the face, the temporal base, so where the face region is located and this has the consequence that people cannot recognize other faces anymore. So you don't recognize your mother, your father anymore because the brain region is missing. With cortiQ, nowadays, we can find where the region is located and they can preserve it, which has also a very important impact for these patients, of course.
Host:
For those experiencing memory loss or the inability to recognize people, such as patients with dementia or Alzheimer’s, is BCI technology being used to treat those diseases?
Christoph Guger:
For sure. There are a couple of research groups around the world working on dementia. For example, Tomasz Rutkowski from RIKEN institute in Japan in Tokyo is doing that. They're using our Unicorn Hybrid Black and they have an assessment tool developed with BCI technology to detect early dementia. This is super important because if you detect it early enough then you can still do something against it. So activities are very important point to avoid dementia so they can do more sports for example. But this early prediction is super important and with systems like recoveriX, I believe we could also help dementia patients. We just didn't do a clinical study yet because we were busy with all the other illnesses. But we see also in our patients - stroke, multiple sclerosis, and Parkinson's, that concentration, performance, and memory is improving when we are doing hand movements.
Another thing that we see is that language was swallowing is improving, and this is my favorite side effect that we are doing. So, people are supposed to imagine hand movement and suddenly they speak better because POCUS area gets activated again. The scientific explanation is that we are activating the sensory motor cortex with hand movements. And if you look at the language network in your cortex, then you need the auditory cortex to hear something. Then you need Wernicke's area to understand what you're hearing. Then you need to POCUS the area to formulate the answer. And finally, you need the motor mouth region and the tongue region to say the word. With recoveriX, we are activating exactly these motor regions and so we have an effect on the whole language network and people swallow better and speak better.
The same is the case for pain. So in Finland for example, recoveriX is used for many years for chronic pain patients where medication doesn't work and they're just doing motor movement imaginations and suddenly they get pain free. And also here the training of the sensory motor cortex has an effect on the brain pain network and we can improve it. For multiple sclerosis, very often bladder control is a problem, so they have to wake up like five, six times during the night and visit the toilet after recoveriX. They don't wake up anymore, they can sleep for a whole night or at least they have to visit the toilet much less frequently. That's also a very important aspect for quality of life.
Host:
Please tell our listeners about your BCI Spring School.
Christoph Guger:
We have top speakers of all the elite universities around the world and it's just very nice to network. So many people are finding their future PhD supervisor in Spring School and you get just a lot of inspiration of what's out in the field. And it took me less than a week to get all the speakers on board for Spring School. So they are eager to be in for a keynote lecture. So, it's all for free, nobody's paid, but we are just getting the best brains in neurotechnology and it's worth the 10 days to listen what's out, and this can direct you into a certain area where you want to spend the rest of your life working on this part of technology.
Host:
For more information or to register for BCI Neurotechnology Spring School, please visit: www.gtec.at/bci-neurotech-spring-school-2025.
If you're looking for more content on brain computer interfaces, be sure to visit Mouser's Empowering Innovation Together page for articles, videos, and more on the subject. Up next in Mouser’s exploration of the newest technologies is renewable design. Be sure to tune in to learn about the innovations driving this field. Visit, mouser.com/empowering-innovation.