Balint Varkuti on Turning Neuromodulation Technologies Into Brain-Computer-Interfaces Using Software by CereGate - Neural Implant podcast - the people behind Brain-Machine Interface revolutions  ## Metadata - Author: **Neural Implant podcast - the people behind Brain-Machine Interface revolutions** - Full Title: Balint Varkuti on Turning Neuromodulation Technologies Into Brain-Computer-Interfaces Using Software by CereGate - Category: #podcasts - URL: https://share.snipd.com/episode/cce922e6-54a2-44c5-9e53-98afb3517e94 ## Highlights - The Secret Sauce of Brain Stimulation Key takeaways: - Other companies are already making electrodes, so this company focuses on the specialty of software - The software has the ability to shape the field of stimulation and structure the signal for conscious perception - The goal is to create comfortable and stable perceptions that are easy to understand - The company uses machine learning and predictive models to personalize the interface for each individual patient - The approach is to partner with the brain by sending signals that it can quickly pick up and utilize for beneficial behavior - The company aims to improve movement and treat different symptoms by leveraging the brain's pattern perception and learning abilities Transcript: Speaker 2 And I think it's very, I don't know, informed or very wise to be like, okay, other people are already making the electrodes and we'll let them do that and we'll focus on the specialty of the software. So do you want to talk about this a little bit on obviously software is very difficult to get across in audio only format. But I guess what's the secret sauce? How is it possible or what's special about it that hasn't already been done? Speaker 1 Yeah, in a certain sense, we have looked into this domain and so I've seen that certain sensory phenomena have been around in brain stimulation for a very long time. So parastigia, for example, false positive sensations where you have an unwanted feeling of a somatosensory nature or phosphines in the visual domain or auditory hallucinations in the hearing domain. So we knew that through brain stimulation, we sometimes are caught in this wire, so to speak and trigger something that the patient can perceive, but what in the clinical domain is unwanted. So for the past 30 years, we've been doing everything in programming to avoid these unwanted side effects. But what we discovered is that there's a whole universe of stability and reliability and reproducibility in these domains. So from the locations where the implant, which is off the shelf, is implanted in a classical surgical trajectory. So you don't have to change anything about the safety profile and the experience for the patient as far as the surgery is concerned. From there, we can actually tap into the rich fiber systems that are in the neighborhood of classical clinical targets. So if we think of the thalamus, then obviously there is a bunch of axonal fiber pathways passing there. The subthalamic nucleus is also in, you could almost call it a wire box of different types of afferent and afferent connections. So in other words, if you know how to shape the field of stimulation, if you know how to structure the signal, then even besides classical therapeutic tonic brain stimulation, you can tap into these systems and create something that the patients can consciously perceive. And the really tricky thing is you don't want it to be uncomfortable. So imagine if we were encoding letters and A would be a bright flash in your eye and B would be allowed to be being torn that would probably work on paper, but it would be a really horrible interface, you would be sitting in your car and driving it against the wall if you get these types of stimulation. So the art is staying really with perceptions that are not uncomfortable, that are easy to perceive, that are easy to understand, and that are stable at the same time. And so we have found out through doing these types of experiments in the past four years with more than 50 patients up to now how basically to do this, we put a lot of this data into a database, we utilize machine learning and other types of approaches in order to have predictive models that can actually tell us well, what type of interface shape, what type of personalized interface is best suited for the individual patient that we see in front of us. And here we really, I want to say partner with the brain. So other approaches where you almost directly try to emulate writing natural neural code, which I think no company on this planet does today, we basically take a different approach. We partner with the brain in the sense that the brain is a fantastic signal, the composer, that's what it does in a newborn baby, the brain also has pre wiring, but still it has to process the stream of sensory information for the first time, get good at it to tell apart shapes, etc. So the brain is really naturally wired for pattern perception for learning. And that's really what we do. So we send signals that the brain very quickly can pick up. It picks them up as behaviorally beneficial. So there is information in them that it can use and need in order, for example, to improve movement or to improve other types of symptoms that we might be desiring to treat. And that is really the approach. So on the one hand, intelligent shaping of the waveforms and stimulation patterns that work from these implants, the other hand, a clinical paradigm that really works with the patient and where the patient's very quickly, literally in a matter of minutes, can utilize these inputs and gain skills or gain abilities that they probably have been missing, such as an artificial sense of balance to avoid falls or even information that is akin to hearing for patients who probably are compromised with respect to the cochlea or cranial nerves and are probably not candidates for classical cochlea implant surgery. Interesting. Speaker 2 Yeah. ([Time 0:02:41](https://share.snipd.com/snip/29a8b848-3c13-4298-957d-d5081c3b9535)) - Exploring the B2B Business Model of a Deep Tech Company Key takeaways: • The speaker mentions finding an intuitive method that is like plugging into a USB port for quick reception. • The company follows a business-to-business (B2B) model and sells services or software to other electrode or implantation companies. • The justification behind the B2B model is not fully explained, but it involves licensing fees or some percentage-based arrangement. • The company positions itself as a deep tech company with the goal of having their brand, Surrogate, integrated into products similar to Intel inside. Transcript: Speaker 2 Yeah, that's really exciting. And yeah, finding this intuitive method, almost plugging in very quickly on almost like a port. I'm almost thinking like a USB port that he's ready to receive something and it was just sitting there the whole time. Yeah, so I want to talk about your business model. You're a business to business company. So you're actually selling your services or your software to other electrode companies or the actual implantation companies. I guess what's the justification behind this? You talked about it a little bit, but how does it work? What do you ask of that company that you're selling to? Is it like a licensing fee or is it a percentage thing or you don't have to tell anything that you don't want to any company secrets? But how does it work? How can a company work with you? Speaker 1 Yes, the business model is B2B. We are essentially a deep tech company. And what we're going for is similarly, like you had in some instances you had Intel inside or some other brand inside. Similarly, we have surrogate inside and some products. ([Time 0:18:18](https://share.snipd.com/snip/bde40daf-0ef4-4898-831d-52726bdb6c57))