What we talked about
Dr. Ewelina Kurtys is a scientist-turned-entrepreneur pushing the frontier of bio-inspired computing. In this conversation, she traces her path from a PhD in neuroscience and 20+ peer-reviewed papers to commercializing deep-tech and advising startups. Curiosity pulled her beyond academia into fast-moving environments where she could turn technical knowledge into real-world impact.
Show notes
Today we’re joined by Dr. Ewelina Kurtys, a scientist-turned-entrepreneur with a PhD in neuroscience and more than 20 peer-reviewed papers. After her academic career, she moved into technology commercialization, advising deep-tech companies on sales, partnerships, and strategy. She is currently a Strategic Advisor at FinalSpark, where she applies neuroscience expertise to build next-generation biocomputers from living neurons. She’s also founded ventures like Psync, a mental wellness app, and Ekai.io, a business development consultancy. With experience spanning research, entrepreneurship, and innovation, Ewelina brings a unique perspective at the intersection of biology, computing, and business.
Full transcript
Federico Ramallo (00:00) Welcome back to the PreVetted Podcast where we spotlight extraordinary people and remarkable talent reshaping our world. Today we are joined by Ewelina Kurtys. She’s a scientist turned entrepreneur. She has a PhD in neuroscience and more than 20 peer reviewed papers. After her academic career, she moved into technology commercialization, advising deep tech companies on sales partnerships and strategy.
She’s currently a strategic advisor at Final Spark where she applies neuroscience expertise to build next generation biocomputers from living neurons. She’s also founded ventures like Sync, a mental wellness app, and Ekai.io, a business development consultancy. She has experience spanning research, entrepreneurship, and innovation. So she’s going to bring an interesting perspective.
on the intersection of biology, computing, and business. Ewelina, welcome to the show.
Ewelina Kurtys (01:00) Thank you so much.
Federico Ramallo (01:01) So tell us what inspired your transition from neuroscience research into entrepreneurship and business development.
Ewelina Kurtys (01:10) So I was always curious when, okay, I dreamed about being scientist, but I knew that it will not be only this. And I always was curious about what is outside academia, what is in the real world. So actually when I was still a researcher, I was exploring and thinking what can I do? And after many trial errors and some networking, I figured out that I could work with the small companies, I could work with startups.
because they offer a lot of flexibility and it’s also interesting, it’s also usually a lot of learning because when you work with a small company you can do a lot of jobs at the same time.
So I went in that direction and started to work with startups. I get fascinated by this, how you can use technical knowledge, not only to generate scientific paper, but also how to change the world. And also I become fascinated with artificial intelligence because this is a very powerful way of changing the world. And I started to work on commercial applications of AI.
And yes, and this is how it started. I transitioned to the commercial side of engineering.
Federico Ramallo (02:13) interesting. Looking back at your PhD journey, what was the most important lessons that still guide you your work today?
Ewelina Kurtys (02:21) Well, I think what is very important is that you really have to enjoy what you are doing regardless the results. And it’s always good to have this kind of approach that no matter the outcome, the work which you do will make you a better person. So I think it’s a good strategy because in research, you know, things usually very often don’t work, but you become more skilled and you get wiser, hopefully. So this is important. And I try to keep this rule also.
academia that every work has to somehow improve me no matter the results because the results don’t always depend on you and you know many things can happen but what is important that you get that you get better so you are they learn something or you know that you you can become better person through your work so I think that’s something what I did during my PhD and I try to continue with that
Federico Ramallo (03:08) Amazing, yeah, I think that self-improvement is is core to entrepreneurship, right? And interesting that it’s also core on research as well
Ewelina Kurtys (03:18) Yes, definitely. In every actually risky project where you cannot guarantee the results because things often don’t work and you sometimes spend months on some experiments and they don’t work.
Federico Ramallo (03:28) Right, right. I mean you can do everything right and still fails, right? But that should not be considered a failure, rather not the expected outcome.
Ewelina Kurtys (03:34) So thank
No, because you can learn something
or try to so that you are better after doing the work than before. So that’s important. You yourself are better at least.
Federico Ramallo (03:45) Right. Right.
So how did you first encounter this idea of bio-inspired or wetware computing?
Ewelina Kurtys (03:53) So I was in London at that time and I was going to many conferences to learn about artificial intelligence. And I met Final Spark founders at one of the conferences in London. And I was very impressed because they are engineers. They are actually trained in signal processing. They have PhD in signal processing. So this is totally different. It’s not biology, it’s mathematics actually. So I was fascinated that they transition.
to totally new field. And also for me personally it’s perfect combination because it’s combination of neuroscience, what I know from my research and engineering, what I know from the commercial work. So this is very cool combination.
Federico Ramallo (04:29) interesting so at finer’s part you’re working on building computers powered by living neurons right can you
Ewelina Kurtys (04:36) Yes, absolutely. So we
want to build processors so that instead of GPUs or CPUs, we use living neurons as a processor in the new type of computer.
Federico Ramallo (04:46) Amazing. Can you explain us a little bit how that technology works?
Ewelina Kurtys (04:50) So at the moment, you know, we have prototypes, it’s still the real computer doesn’t work yet, you could say, but we have done a lot of work in prototyping. And basically what we do, we make around structures of neurons, which consists of around 10,000 neurons each. And these are called the Neurospheres, or you can also call them Organoids. And this round 3D structures we put on the electrodes.
usually on the 8th you can see this on our website finalspark.com there is a section live and you can see how it looks how this electrodes looks usually neurons you cannot really see with the naked eye because they are semi-transparent and what you can see are the signals also on the website so we measure constantly 24 hours per day 7 days a week we can measure signals from neurons and some of them are also visible on the website
And the whole idea is because neurons in our brain, they generate electric activity through some movement of chemical ions through the membrane. So they generate spikes, is so-called electrical activity. And they also can be affected by many chemicals, for example, neurotransmitters such as dopamine or serotonin.
And in the lab, we try to use this to program them. So we send them electrical signals through electrodes, or also we can release to the medium, to the liquid in which neurons are placed. We can release some chemicals, for example, dopamine, and this also can change the behavior of neurons. So we try to program them and the objective is so that they can process information.
Federico Ramallo (06:23) Right, right, I remember that at one point they were talking about mechanical computers, right, which is something that was the path for calculating ballistic trajectories and things like that, but basically were based on gears, right, and other mechanisms, right, which at that point was faster than binary computers, right.
Ewelina Kurtys (06:39) Yes.
Yes.
Federico Ramallo (06:45) now binary
computers went up in processing power but then there has been a renewed interest in these mechanical computers or basically they are designed for a very specific use case, right? ⁓ So I’m wondering
Ewelina Kurtys (06:57) Yes.
Federico Ramallo (06:59) I’m wondering if something similar is happening with neurons, right? Where you’re basically being able to build a specific type of computer for a specific use case that is going to be providing much more efficiency and performance compared to binary computers, right?
Ewelina Kurtys (07:18) No, actually not really because we are aiming very ambitiously on general computing. So we would like to make a computing device. And we think that generally the tasks which are today done on digital neurons would be great to be done in living computers. For example, generative AI, because it costs a lot of energy today and actually…
Federico Ramallo (07:24) Ohhhh
Ewelina Kurtys (07:40) generative AI, like solving some complex problems or generating ideas, could be a great task for living neurons.
Federico Ramallo (07:47) Ok, so the benefit of the neural computers would be for generative AI, right? ⁓
Ewelina Kurtys (07:54) Yes, that’s what
we have in mind, although it’s very speculative at this stage because we are very early stage. for now, for example, we managed to store one bit of information in our neurons. So, you know, just to give you idea that is very early stage. And for now we have some assumptions, but of course they have to be tested in practice.
Federico Ramallo (08:12) Right, right. Well, I think that’s the case when you’re on the edge of innovation, right? You’re taking additional risks and it’s hard to make any promises in the future, right? Because there’s so many unknowns, right? ⁓
Ewelina Kurtys (08:24) Absolutely. Yes, I mean,
we have of course a timeline, we have plans very specific for the next 10 years. But you know, it’s always a risk and for sure many things will change and I’m sure we also will have some unexpected results. So yes, so this is always it’s hard to say what exactly will happen.
Federico Ramallo (08:45) Right. So the expectations that you have is that the living neural computers are going to be able to be faster for generating AI, or what are the benefits that you’re planning to achieve with living neural computers?
Ewelina Kurtys (09:00) Energy efficiency, because neurons are one million times more energy efficient than digital. So that’s the main thing. They are also easy to scale, very easy to produce large amounts of neurons. And we know for sure they work because this is why we can talk today, because our neurons are processing information. So we definitely don’t expect them to be faster than digital because actually when we look at the human brain, we can have already some ideas about.
where neurons are good and where not. So we can say for sure that computers have much more memory, they are much faster for repetitive tasks, but for complex tasks they use a lot of energy and they not very efficient. And this is in contrary to the human brain. So we expect that neurons in our bioprocessor will be great for this kind of tasks for which human brain is good.
Federico Ramallo (09:50) interesting. would you be able to do parallelization with the neurons? I understand that’s something similar to what happens in the brain, right? Where the brain can have bifurcations on the neuron path and basically can solve complex issues in parallel compared to a single core computer, right?
Ewelina Kurtys (10:08) Yes, actually living neurons, they can do a lot of stuff at the same time and also they have a lot of, they have much more connections between each other. Also a lot of recurrent connections which are possible in digital but they are much more difficult. So brain is, we can say, much more connected than digital neural networks.
Federico Ramallo (10:29) Right. So what are the main challenges of developing this type of computers?
Ewelina Kurtys (10:36) So actually there are many challenges because neurons are processing information in totally different way than digital computers. So we have to figure out everything from the scratch again. And this is really difficult because nobody really knows how neurons encode information. We know a lot about processing. We know that they produce spikes.
electrical spikes. We know that they encode information in time and space, so it matters when they are spiking and where in the brain, but we don’t really know what does it mean. So we cannot read really human thinking today, but that’s kind of necessary. We have to be able to somehow translate.
these signals into some specific knowledge and this is very, very difficult. And this is why we have our lab fully automated. So basically running experiments in the lab at Final Spark means that you write a script in programming language, Python, and everything happens automatically because we need so many experiments, such a huge volume, and that’s much more efficient.
So it’s very, very difficult because nobody really knows how to program neurons and you actually need a new framework, some kind of mathematical framework to describe how neurons encode and process information and how we should code them, how we should program them. So like kind of new programming language, totally new.
So for now, you write code in Python. Of course, we still use a digital interface and then everything is translated into machinist behavior. So for example, you can code in Python that you will send some electrical spike or maybe some dopamine to the neuron and then it will happen in the hardware. So you programmatically can change the hardware and things will happen.
but we still don’t have perfect formula on how to interact with neurons to get some meaningful interactions.
Federico Ramallo (12:27) Right, I see. So when you’re saying that you don’t have meaningful interactions, are you also referring to the non-deterministic nature of the outcomes? Or that’s a different topic?
Ewelina Kurtys (12:38) I think that that’s not really clear in science, whether brain is deterministic or not deterministic. I think it’s for sure a lot of chaotic and we cannot predict. It’s not predictable, let’s say, but whether it’s deterministic, I’m not sure it’s a consensus in neuroscience. And you are up, but you are going in the right direction because brain is generally messy. So if we talk about this, it’s very messy and it’s not stable.
This is extremely difficult because computers are stable system. So you put some input, you get output and you know, things are working the same way every day. But brain is plastic. So also our brain and neurons in vitro, they are plastic. So that means they might change over a few days. So.
For example, today you do one experiment, you put some input, you get output, and tomorrow you can get a different output for the same input. So this is actually extremely difficult, that it’s unstable. And for us, for today, it’s not possible to predict always the answer. So when we send some stimulation, we don’t always know what we get back. So it’s hard to predict. So that’s why it’s so difficult also.
Federico Ramallo (13:46) Interesting.
Interesting, Yeah, as far as understand, we don’t have a really good understanding of how the innards of our brain actually works, right? We have some hypotheses, but we have not been able to understand the innards, right? ⁓
Ewelina Kurtys (13:59) Absolutely not.
Yes.
Yes, absolutely.
And especially the cortex, is a lot of signals and we don’t really… For us it’s just a mess in the signals.
Federico Ramallo (14:20) Right. And I understand that part of it is the complexity of the chemistry process, but also, and I think you mentioned this, the amount of neurons that our brain has, right? That makes it much more complicated to process and understand the data that we could even collect, right?
Ewelina Kurtys (14:27) Mm-hmm.
this.
Yes.
Federico Ramallo (14:45) interesting. So you mentioned that biocomputer could fit in the future as of AI and data processing, right? Are you thinking of other use cases for the future?
Ewelina Kurtys (14:56) No, we are actually focusing on AI because it’s using exponentially increasing amount of energy and we think that biocomputing can solve this problem because it’s so much more, one million times more energy efficient. So we try to make solution for the increasing energy consumption by AI.
Federico Ramallo (15:09) wow.
Right, I see. And does the neuron, maybe this is too early to say, does the neuron computers suffer degradation, like, know, a silicon chip, you know, should not degrade over the next 10, 20 years, whatever, they become obsolete eventually, But working with a living being, living neuron, then I think that…
That could be the case, right?
Ewelina Kurtys (15:46) Yes, this is a very important concern actually and we put lot of effort to keep neurons alive for a long time. In the laboratory conditions, especially when you put neurons on the electrodes, it’s very difficult. So far we managed to keep them alive on the electrodes for three months, which is really a lot for industry standards. We also have some patents for this, for the solutions related. And of course we want to still increase this time.
So in the future, we hope we can keep them alive for years. We know from nature, neurons can live even 100 years because in our heads, they usually don’t divide. When we are adults, we have most of the neurons the same. Only in some regions, there might be some neurons formation, but in most of the brain, we have the same neurons for our lifetime. So we know they can live long, but in the lab, it’s much harder.
because you have to actually try to mimic the conditions of the human body as close as possible and every slight deviation is actually problematic. So that’s why it’s so difficult to keep them alive for long time. But we definitely plan to extend this.
Federico Ramallo (16:56) Interesting, interesting. I thought that the neurons would also go through a regeneration process like the rest of our body, know, the cells that they regenerate, right?
Ewelina Kurtys (17:06) No,
actually that’s a big problem also for humans because when we have some brain injury it’s very difficult, know. That’s why brain actually has different solutions because it can rewire. So let’s say if we lose some part of the brain then let’s say other parts of the brain can take over the same task but of course this is to some limit and this is actually a big problem also in neurodegenerative diseases that usually when you have
you know, neurons, you lose them over your time. So that’s really something to worry and care about. Otherwise, yes, we can get slow degradation and cognitive decline, over time.
Federico Ramallo (17:49) interesting. knew about the cognitive decline, but I haven’t made that connection. Very interesting.
Ewelina Kurtys (18:00) Actually, there many reasons,
but usually over age, yes, neurons are dying. Sometimes you have imaging studies when you can look at the change in the volume of the brain. And especially if people are sick, like if they have neurodegenerative, but even just aging can just make the shrink the brain. That’s why it’s so difficult to keep cognitive abilities for long time.
Federico Ramallo (18:22) Right.
Right, right. We were able to, throughout the years, generations, you know, live longer and keep our bodies, you healthier. now we’re getting to the limit of, I mean, the body could be healthy, but our cognitive function is not there anymore, right?
Ewelina Kurtys (18:33) Yes.
Yes, this is a big problem. Although I’m sure there are solutions for that. for now, at least a lot in the lifestyle, nobody yet has a pill. But of course, researchers are working on longevity solutions. So maybe one day we just can drink some elixir and then we get younger.
Federico Ramallo (19:05) Wouldn’t that be nice?
Ewelina Kurtys (19:06) I know actually there are a lot of research in longevity to try to mimic some of the things related to healthy lifestyle so yeah there is a chance for the future but for now only the lifestyle
Federico Ramallo (19:16) the
Right, right. Also, I don’t remember what I read about this, but anyway, I’ll try to explain it the best I can. But, you know.
One of the things that gives humanity hope is that new generations, new babies are born blank basically. That means that there’s hope for change, there’s hope for new generations to take new cultures and…
understand new concepts, right? But the older we as humans we get, right? The more experience we get, the more we tend towards the bitterness and the, know, because we, maybe not bitterness, but we lose this sense of exploration and this…
awe for new things, right? A new baby can be surprised by a light that can turn on and off or a new toy, right? And we lose that feeling the older we get because we’ve already seen it, we’ve already done it, so there’s nothing more than surprises anymore, right? So we kind of lose this concept of joy, right? So if it was the other way around, if new kids were born with
Ewelina Kurtys (20:07) Mm-hmm.
Federico Ramallo (20:30) the collective memory of the parents, then we as humanity will be lost, right? Because it will tend towards that bitterness and lose of interest for anything, right?
Ewelina Kurtys (20:48) Mm-hmm.
Federico Ramallo (20:48) So your work on PhD, you produce more than 20 peer-reviewed papers, right? Which publications you feel represents your contributions to neuroscience the most?
Ewelina Kurtys (21:01) I think that my PhD research, because I had a very nice topic actually, I investigated the effect of nutrition on brain inflammation. And that actually was very helpful because I changed my view on my groceries at that time because it was about how nutrients can affect the brain. I read a lot of papers about this and I realized how important it is.
to eat unprocessed food and look try to make the diet anti-inflammatory because it can prevent some diseases and also can help healing. So healing the brain. So yes, I appreciated the importance of this and what I studied some combinations of nutrients in vitro on microglia. So microglia cells are
kind of like immune cells of the brain. So I was researching how nutrients can change the activity of microglia so that they’re not so inflamed. And I also did a really cool research in brain imaging so that I enjoyed a lot. That was my favorite. So I did PET imaging. So it’s positron emission tomography. It’s such a complicated name. But the idea is that you inject radioactive substance.
and the radioactivity is something that is radiating so you can detect this from the distance. So in this way when you put a specific substance with a radioactive label to your body you can basically image the body, some function of the bodies.
of the body. So for example activity of microglia, these immune cells of the brain, can image this non-invasively just using PET imaging. So that was, think, my favorite studies and I researched some effects of nutrients on how you can prevent or how you can help to heal the stroke effects. So that was interesting for me.
That was the nicest, what I remember.
Federico Ramallo (22:53) Interesting. What was your most surprising discovery? You you built as a hypothesis and then you found completely unexpected results that quite surprised you.
Ewelina Kurtys (23:06) Actually, to be honest, sometimes I think that science in some ways is following common sense, because for example, we know from four years that diet and exercise is helping you to keep healthy. And now we still have new studies being published to show the results that
how diet and exercise can improve your brain function and other things. Although you knew this already. So I think very often we test something we are kind of sure about. So surprisingly, there is not so much surprise in academic research, I would say. In many ways, you follow the intuition. actually, it’s still a…
know, interesting because there a lot of ancient stuff which maybe people believed sometimes and after years you prove them in the lab. Of course, not always, but I would say a lot of stuff are just confirming your common sense about healthy lifestyle.
Federico Ramallo (24:02) Right, right, the grandmother giving you advice that was passed on generation and generations that, you know, then…
Ewelina Kurtys (24:10) Yes.
After you prove this in the
lab, that is indeed true. I would say a lot of stuff about the brain. Of course, there are some details about how exactly which chemicals exactly are working in the brain. But I’m not sure it can be so much surprising because usually, in the science, think you make also some incremental little steps forward. So you usually based on some previous knowledge.
Federico Ramallo (24:18) Right.
Ewelina Kurtys (24:38) so you make little step forward. I think it’s not that surprising. Of course, sometimes you have such a surprise like when, I don’t know, penicillin was discovered. Sometimes you have some serendipity, maybe, but in many ways it’s just, it’s often also you get results you expected. So maybe what can be, you know, surprising is when you plan some study and it doesn’t work.
Federico Ramallo (24:55) Right.
Ewelina Kurtys (24:59) and you don’t get any results, that can be surprised. That sometimes can be surprised. That means that you either had, there are two options actually for this, you either had the wrong hypothesis or you executed the experiment wrong. So you never know which one.
Federico Ramallo (25:04) Right.
I see. I see. And how do you validate the execution of your experiments? Through peer reviewers?
Ewelina Kurtys (25:22) yeah, well actually first inside the lab on your own you always have to repeat the experiments. I mean let’s say at least three times it depends on what you do but let’s say at least three times. Also at final spark if we said we’d have done something that means we repeat it several times and we really get reproducible results because especially in biology it’s very easy to
get some nice results once, but then after you don’t get it anymore. So it’s very important reproducibility. It’s actually a huge problem in biomedical research because, you know, a lot of papers are not reproducible in other labs. But I think this can be reason also for that there are many little details, you know, that to do an experiment good, you need to take care of both many, many little details. And I think that makes also reproducibility more difficult because someone has done this.
they were experts in the field and for others might be difficult because they just don’t know how to perfectly execute this experiment. So it’s very difficult. But generally, let’s say reproducibility in biology is a huge challenge. But you always have to have you have done experiments independently on different days, on different samples. And then you should get if you get similar results, you do some statistics and you can usually statistically say if it’s significant or not the difference.
and then you say okay then you can publish that as a positive result and of course in science it’s a huge problem because usually if you have negative results it’s hard to publish but i think there is some rationale for that many people criticize this but i think it’s not so stupid because actually it’s very difficult to prove that something does not exist because it’s easier to prove you know if you have positive results you say okay it exists it’s like that
But if you have negative results, it doesn’t mean it doesn’t exist. Maybe you just done something wrong. So I think it’s really difficult. You know, I am not so much against it because I know many people criticize this that the negative results are not published. But you know, many times they can be just failed experiments. So it’s hard to say. But yes, you usually try to do your best with statistics, with repetition on your own. And then later you submit to the journal.
And yes, and you have what you mentioned is per reviewed. So that means you have to wait for a few weeks or maybe a few months so that other people read your paper and maybe they give comments and they say if it can be published or not. So that’s how it looks usually.
Federico Ramallo (27:47) Right.
Yeah, it’s very difficult to prove a negative, right? And that’s the challenge.
Ewelina Kurtys (27:53) Yes, exactly. And some people criticize
this. But actually, I think many times it can be just failed experiments. So yeah, the fact that you don’t get results doesn’t mean it doesn’t exist. So yeah, very, very hard.
Federico Ramallo (28:06) Right. Right.
And then the other issue is getting into publications on a science newspaper. That’s also challenging, right? Because you have to go through the peer review process and getting all there until it can actually be published, right? Which I can understand the reasoning around that, right?
Ewelina Kurtys (28:23) Yes, it’s always a long process.
Federico Ramallo (28:29) then anybody could publish anything and then the value, the noise starts to increase and then the value of those publications, of those papers kind of diminish, right?
Ewelina Kurtys (28:40) Yes, absolutely.
think peer review is very important because, yes, I think it’s really important. yeah, it takes a long time, but also researchers actually are doing this for free most of the time. So when you are academic, you want to become professor, then usually you are volunteering for reviewing the articles and you practically do this for free. So it’s a lot of work.
So yeah, think that’s maybe the reason why it takes so much time.
Federico Ramallo (29:08) you mentioned a little bit before about using imaging technologies, what drew you to combine it with neuroscience?
Ewelina Kurtys (29:10) Mm.
I think imaging is very powerful for neuroscience. You can measure in vivo what happens, not in the sliced brain, but in the real living brain. So yeah, I think it’s very powerful and it allows you to do research on humans. That’s very important because…
you don’t do only animal research, but then you can translate easily to human because you basically can do the same method. So that’s actually easier because what is the objective of all the medical studies most of the time is to find out how to cure human brain. So that’s always important that things are easy to translate. So yes, and it’s very non-invasive. It’s let’s say,
painless method to study what happens. And it’s used in neuroscience, also in psychology, in any actual research which are connected to the, which are about human brain, it can be used to measure activities. And I think it helps us to understand how brain works.
Now there are even some kind of attempts to kind of read human thinking, but of course it’s attempt only, because you can do classification, only classification. So for example, if you give to the subject like 10 different images or sounds, music, then you can classify which one
the subject is thinking about based on imaging. But only if you know previously which music or which… So it’s not like totally reading your brain. It’s only… It can guess which one you’re thinking about from predefined list. So yeah, that’s such interesting stuff. So yeah, I think it’s very powerful and it’s used more and more and not only in research, also a lot in diagnostic because also it’s used a lot to, for example, detect tumor and stuff like that.
Federico Ramallo (30:53) Right.
It is
Right, right, yeah, I understand that what they’re doing is kind of blindly, compare the images, The scanning images, right? It’s like, I don’t understand what’s going on, but if they look the same, it ⁓ kind of goes that way.
Ewelina Kurtys (31:24) Yes, in many ways it’s like
that. When you do classification, yes, then you can just say that when you think about each music piece, then your brain activity is a bit different and you can compare. So sometimes you know what is going on, sometimes you don’t really know everything.
So yes, sometimes you blindly compare. Sometimes you know exactly what you’re looking for, especially for diagnostic. When you diagnose, then let’s say you should know what you’re looking for. But if for research, sometimes you can, you don’t really know exactly.
Federico Ramallo (31:54) Right, right. How does the neuroscience compare when you work with animals versus humans? Is there a big difference? easy to translate? If something happens in animals, then we can safely assume something similar happens in humans? Does that translation exist or it doesn’t?
Ewelina Kurtys (32:15) I always work with animals, so I think it’s maybe a little bit easier. I think with humans you have much more bureaucracy. is then becomes a lot of paperwork when you work with human studies, not so much experimenting. You collect some little data, but usually it’s lot of stuff, a lot of ethical approvals because the human studies are the hardest. Generally, this is big problem in biomedical research. It’s not easy to translate.
You might have some drugs, stuff working in animals and later in humans they don’t have to work the same way. Because of course animals they have different enzymes, different metabolism, different doses. There are a lot of differences, so yes it’s just a model, but many things don’t work the same way. So it’s always a risk when you test something in the humans that it might not work.
Federico Ramallo (33:03) you
right, right I was reading about full scan full body scan, right and one of the reasons people were suggesting not to do those type of full body scans was there’s always something wrong with your body, right? so that will trigger false diagnostics, right? or would scare people, right? so even though it could be a preventive
mechanism, right, to prevent some early detection of any sickness, but it also opens the door to all these false negative issues. I’m wondering how much of that situation, you know, like every human body is different, how much of that actually would affect your neuroscience research, right?
because at the end of the day need something repeatable and if you have a non-deterministic human body subject, if you like, then how can you make that distinction between is this a variation of the patient or is this a variation of the actual sensor you’re trying to prove
Ewelina Kurtys (34:11) Actually, don’t agree, to be honest, with this full-body scan. I think even if you have false positive results, you can always double-check. And I think it’s better to know than not to know. Especially many cancers, you know, they are… I mean, it’s so much important to have early diagnostic. So I generally think that it’s good to do preventive studies. The only problem is that there are many ways of imaging. So if you do full-body scan, which techniques do you use?
because there so many different techniques and they all show something different. So actually when you do this full body scan, it depends a lot which technique you do and you cannot always see everything. So I think it’s important to use maybe, I’m not sure it’s necessary, definitely not necessary to do PET for diagnostic, just maybe it’s enough to do MRI, but that’s good for brain.
but for the rest of the body there are other techniques which are cheaper, for example ultrasound. So I think it’s important because I know there is this marketing of doing full body MRI maybe, but I think that’s a little bit marketing because at the end there are many techniques and they all show something different.
So if you would like to be really sure, then you would have to use different scans. I think that sometimes it’s marketing. But generally it’s good to do diagnostic because the earlier you know, the earlier you know anything is wrong. Even if it’s, I think it’s better to get false positive than not to know when you are sick and then, you know, get terminal. So that’s my opinion at least. And when you do the studies,
Federico Ramallo (35:24) Right.
Ewelina Kurtys (35:42) you always have to have different subjects. And that’s kind of like maybe kind of helping you to solve this problem. Because, okay, you can have false positive maybe with one subject, but if you repeat a lot, that’s why repetition is so important in bio, any kind of bio studies that you have to always repeat. And if you get results consistently, then you can say you have something. Otherwise,
It can be just a coincidence, indeed.
Federico Ramallo (36:10) Right, right, yeah. I still remember in my statistics class where they would talk about a sample that is statistically representative, right? Which is basically with volume, then you can get all the specific samples, deviances, out of the actual metric, right? Yeah.
Ewelina Kurtys (36:22) Yes, always,
Yes, absolutely.
Federico Ramallo (36:33) and yeah, I agree with you on the full board scan I’ve seen the same behavior with the Apple watch and these health trackers where people get obsessed about the numbers and the metrics and the Apple watch filling the rings and I think it’s important to know to have that information and know rather than not knowing
Ewelina Kurtys (36:47) Mm-hmm.
Federico Ramallo (36:56) But then what do do with that information? Do you take it as an absolute? Do you overreact for any… I haven’t filled in the rings for the last two days or whatever I don’t think we should react to those metrics like that because that’s not healthy But being able to know and being able to be consistent on improving
the numbers towards the goal we want to do, I think that’s much better than just obsessing with the metrics,
Ewelina Kurtys (37:26) Absolutely.
Federico Ramallo (37:28) yeah, yeah, and I agree with you also on the you know, it’s better to have a false positive and work it work it out than than not right Yeah, yeah I also seen that when you’re doing imaging it also can define the level of detail that you you’re getting so if you’re doing a full body scan you could get a low resolution and then focus on the parts that you you want to focus the most right which you know
Ewelina Kurtys (37:52) Yes.
Federico Ramallo (37:54) that could also affect the results.
Ewelina Kurtys (37:57) Yes, absolutely.
Federico Ramallo (37:58) So you co-author on gene imaging and CAR T cell therapy, right? What was the biggest insights for those studies?
Ewelina Kurtys (38:08) T cells. Okay, this is on cancer imaging. Actually, I continued in London as a postdoc in cancer imaging. And CAR T cells is very cool new technology on which we were working on preclinical studies a few years ago, but now I see a lot of studies in humans also and they are effective. So this is when you use the genes from your own cancer, kind of, from your own body. You use the genes to modify T cells.
and then these modified T cells you inject to the human body, to the patient, and they should kill specifically the tumor in this patient. So this is very specific. This is example of personalized medicine.
where you had a treatment exactly for one patient. So this is most personalized possible that it gets treatment specifically for this one. And I think in cancer is often necessary because every cancer is different. This is actually terrible. And the variety because it’s genetic mutation. So mutation is random. So that’s why what comes out is…
always a little bit different. So that’s why it’s so much important to have personalized medicine. And also it’s important for any type of metastasis because the scar T cells, they of course, they are flowing in the body so they can go anywhere. So they can also help with the tumors which are spread over the body, which is huge problem. And actually in many cases they are terminal.
Federico Ramallo (39:19) you
Ewelina Kurtys (39:35) because if you have metastasis, they can be everywhere, you cannot eliminate the cancer from the body. But with CAR T cells, there is a hope for doing even this. So yeah, that was interesting. I was doing a preclinical on mice just to prove that CAR T cells can arrive to the tumor. So we were using imaging to show it.
Federico Ramallo (39:53) wow, very interesting.
Ewelina Kurtys (39:55) Yes, just to confirm
that they can really arrive where they should.
Federico Ramallo (40:00) Right, so basically you can attack on a cell level.
Yeah, because that’s always the issue with with you know, we sell tissue right with a cancerous tissue, right? If you if you cut too little then you’re leaving, Unhealthy tissue, but if you cut too much, you’re cutting healthy tissue, right?
Ewelina Kurtys (40:19) Yes, absolutely. But with CAR
T-cells, it’s kind of like super precision because this is biology. So your own T-cells are fighting. So this is much more precise. And also it can go everywhere in the body. So even if the tumor serves any different part of the body, it can still arrive there.
Federico Ramallo (40:35) Yeah and also it’s less invasive, more precise,
You wrote a paper about the open and remotely accessible neural platform, right?
Ewelina Kurtys (40:44) Yes, this is very
relevant paper for our discussion today. This is actually about Final Spark. This is the latest publication and it’s actually very successful because we published this in Frontiers, in peer-reviewed journal. It was accepted after the first, you know, sometimes you have to try many different journals, but here we didn’t have to. We submitted to Frontiers, to the special edition and we were accepted.
Federico Ramallo (40:48) Hehehe.
Ewelina Kurtys (41:09) And what is very interesting about this paper that it’s in top 1 % of all the papers read in this journal. So I’m sure it’s in big part due to our efforts because we try to promote this, but we are very happy for this. And this paper is not really experiment, it’s more description of what we have in the lab because our lab is available remotely for researchers all over the world. So basically you can access through internet browser.
and you can do experiments, collect data, analyze the data, do everything, do full experiment in our lab without ever entering in the lab.
So you can do this from any place in the world. And we try to describe as much in details what we exactly have because people often come to us, of course, with expectations which are not possible to fulfill. So we want that all the engineers or scientists who would like to collaborate with us, that they know exactly what we have in the lab. What can we do? What is possible today? So that we have everything automated 24 hours per day, seven days per week. You can run experiments.
sending electrical signals to neurons, you can receive the response. And yes, it’s a very robust tool for biocomputing, for prototyping, for learning how we can program neurons.
Federico Ramallo (42:24) Interesting, I would love to learn more about that. I find it fascinating. mean, I’m a software engineer, you know, being able to automate, you know, living tissue, living neurons research, I think that’s fascinating.
Ewelina Kurtys (42:39) Thank you so much. are very happy to hear that we get messages from people from everywhere in the world. There is a lot of excitement about this project.
Federico Ramallo (42:50) Yeah, and I’m just thinking about the possibilities, because now you can access, you can have a globally distributed team of scientists working on the research, and as you said, 24-7, so that’s another huge benefit.
Ewelina Kurtys (43:02) Mm-hmm.
Yes, and it also shows that our research is scalable because you can easily incorporate new groups in the lab without increasing the floor space so much.
Federico Ramallo (43:21) Right, and repeatable, because now you can repeat it as many times That’s amazing
Ewelina Kurtys (43:23) Yes.
Yes.
Federico Ramallo (43:27) So how do you envision researchers around the world are going to use this remote biocomputing resource?
Ewelina Kurtys (43:35) So we have already 10 universities which are using our lab for free and they have some ideas, for example about how to research connectivity inside the neurospheres because it’s 3D structures and we have signals only from the surface, but there are some mathematical methods to figure out what is inside.
And what is also surprising for us, nice surprise and side effect, which we didn’t expect, is that we have first commercial clients, because there are people who reach out to us, companies, individuals, startups, who are working with us, and they pay us to get subscription, to get access to the lab and try to do some R &D on our neurons.
Federico Ramallo (44:17) Right, amazing.
Ewelina Kurtys (44:21) So sometimes,
know, sometimes there are startups who have some novel ideas related to biocomputing. And sometimes there are big corporations who have R &D teams and these R &D teams want to know what is on the edge of science. And they like to try out the new technologies, even if they don’t work yet, but they want to know what will be possible in a few years and get familiar with it.
Federico Ramallo (44:48) Right.
What were the technical or ethical considerations you had to address for publishing on living neuron-based computing?
Ewelina Kurtys (44:54) So, well, technical and ethical, are two different things. So technical a lot, there is huge amount of stuff you have to do to make things possible and we are still working on improving our lab. And for ethical, we get sometimes questions from the public because project become popular. So people are starting to ask questions and this is why we reach out to philosophers who are working on technologies.
so that they help us to answer some questions related to biocomputing because of course we want that our work is accepted by society. So if there is anything we have to do, we will be happy to do this to make sure that it’s ethical what we do.
Federico Ramallo (45:32) Right, right. So we’re running out of time, so I’ll ask one last question before we wrap it up. how do you imagine the future of science in the next 10, 20 years? And what role do you see living systems as biocomputing, you think, are going to play?
Ewelina Kurtys (45:51) So I think a lot will change in the near future because of generative AI, because of the LLMs, because they give new very potent possibilities for automation and even academic research, experimental research engineering can be at least partly automated. So I think it’s a huge change and I’m curious how it will work out.
and how the future will look like. So in around 10 years, we want to build bio-servers. So this will be centrally available, bio-computers, which you will be able to connect to, like to the cloud computing today. And this will be cheap computational power. So this is what we aim for.
Federico Ramallo (46:29) Amazing, amazing. So before we wrap it up, Ewelina, any last remark you want to share with our audience?
Ewelina Kurtys (46:37) So I encourage anyone to reach out to us if they have any question and to check our website finalspark.com. We also have Discord community so you are very welcome everyone to join. There are a lot of discussions there also technical. So I’m sure it’s a lot of fun.
Federico Ramallo (46:53) Amazing, amazing. I’ll check it out. I think what you’re doing is amazing. And I’m looking forward to hear more news, ⁓ great news about what you’re building. Thank you, Ewelina for being here today.
Ewelina Kurtys (46:58) Thank you.
Thank you so much.
Thank you.