June 6th 2021 | E055 How Do Brain-Computer Interfaces Work And What Can They Achieve? Career In BCIs As a PhD
Dr. Davide Valeriani is a senior machine learning scientist at Neurable, a Boston-based startup developing EEG-based brain-computer interfaces (BCIs) for enhancing focus and controlling external devices. He obtained his bachelor and master degrees in computer engineering from the University of Parma (Italy), and his PhD in computing and electronic systems from the University of Essex (UK), developing collaborative BCIs for improving group decision-making.
Between 2018 and 2021, he was a postdoc at Mass Eye and Ear / Harvard Medical School, working on deep learning algorithms to diagnose dystonia from MRI data. Dr. Valeriani’s research interests are in the areas of computational neuroscience, machine learning, brain-computer interfaces and decision-making. He is also a passionate mentor, loves cooking and playing with his cat Gin.
Davide’s LinkedIn profile: https://www.linkedin.com/in/davidevaleriani/
The episode was recorded on June 5th, 2021. This material represents the speaker’s personal views and not the views of their current or former employer(s).
Natalia 00:10 Hello, everyone. This is yet another episode of career talks by Welcome Solutions. And in these meetings, we talk with professionals who have fascinating career paths and who are willing to share their life hacks with us. Today, I have the great pleasure to introduce Dr. Davide Valeriani who obtained his Bachelor’s and Master’s degrees in computer engineering from the University University of Parma in Italy, and his PhD in computing and electronic systems from the University of Essex in the UK, developing collaborative BCIs for improving group decision making.
Between 2018 and 2021, he was a postdoc at Mass Eye and Ear Harvard Medical School working on deep learning algorithms to diagnose dystonia from MRI data. Today, he’s a senior machine learning scientist at Neurable, a Boston-based startup developing EEG-based brain-computer interfaces for enhancing focus and controlling external devices. Great to have you, Dr. Davide Valeriani, and thank you so much for joining us and for accepting the invitation. And I have to first say that, I know that you’re in Boston right now.
And] I have to secretly tell you that I am learning how to do the Boston accent but I’m not very good at that yet. It’s my favorite one among all the American accents. I’ve seen Boston Rob on survivor. This’s the accent I want to have. I think I’m halfway but I will only reveal my brand new accent when I make it. Next time, we’ll talk in a Boston accent. But it’s great and it’s my favorite.
Dr. Davide 01:52 Thank you so much for the invitation. I’m glad to be here and to chat a little bit with you. And I’m not an expert on the Boston accent. I still retain my Italian accent here. I can still survive with it.
Natalia 02:09 You can still listen next time. Anyways, great to have you. And please tell us your own story from your own perspective, please.
Dr. Davide 02:20 I grew up with a big interest in computer science and coding. When I was four years old, my dad was bringing home a computer every weekend and I was trying to play with it over the weekend. Back then it was a different type of computing, different advances in computer sciences were before Windows 95. It’s crazy times. That’s how I kind of went on with my career starting from high school and then bachelor’s and master’s degrees in Italy in computer engineering and then wanted to see how I could use my expertise in computer science to advance the knowledge we have in some key areas.
And one of these key areas was neuroscience. I approached the world of neuroscience when the whole brain initiative in the US and similar initiatives in Europe started. And I thought that was a great addition to be able to contribute in that area. I purchased the area of brain-computer interfaces in particular and the idea here is that you want to be able to build new devices that allow you to control a computer or an external device with your brain. That’s what got me fascinated.
I watched this movie in Pacific Rim when there was like people controlling robots with their mind collaboratively. And that’s how I also started my PhD research in collaborative brain-computer interfaces to improve decision making that I did in the UK at the University of Essex. And, after that experience, I moved towards clinical neuroscience and I moved here to Boston back in 2018, where I joined a lab at Harvard Medical School to understand how we can use again computer science to advance the diagnosis of particularly rare disorders like dystonia from structural MRI data.
Brain imaging and deep learning were my expertise there. And then three months ago, I joined a new Neurable. And there the idea was to make sure that I could advance my career into how I can experience BCIs and neurotechnologies from an industry perspective and understand what is missing out there to be able to get a neurotechnology or a BCI in the supermarket and buy it and use it on a daily basis. All these devices, we still very much used in academic labs and research labs.
That’s what I want to contribute. How can I join a company that is aiming right now to make a big BCI that people can buy and can use for specific reasons for getting insights into their brain activity or for controlling external devices with their brain? How can we make that transition?
Natalia 06:08 My first question for you before we get into details related to your work is, what is BCI for people who don’t have experience with BCI, who heard the term but never really worked with it? Could you explain in simple words how it works and why is it important?
Dr. Davide 06:57 BCI stands for a brain-computer interface, sometimes also called a brain-machine interface. And the idea here is a device that allows people to control or communicate or do many different actions with their brain activity. The idea here is a brain-computer interface, meaning a new way to interact with computers as we do with the mouse. With the mouse, what it does is transform our movements into movements on the computer screen. What a BCI does is translate and decode information from your brain activity.
You are thinking about what patterns are there in your brain that are present and transforming or translating those patterns into actions on the computer could be like moving the mouse with your brain. You think about, let’s say, I want to move the mouse to the left, and the mouse moves to the left. That’s a brain-computer interface. It could be the fact that you are working on a daily basis and your brain activity is monitored by the computer and the computer tests.
It seems that you have lost focus on this task, maybe you need to take a break. There are a number of different applications for BCIs. BCIs were born for people with paralysis and people that were locked in that couldn’t have any other way to communicate other than being able to decode information directly from their brain activity. The brain was working properly. But they couldn’t move any muscle, for example, or they couldn’t speak. And so their BCI would enable them to send messages, short messages to the outward there, say like I’m hungry, or yes or no, or short sentences, and those made like a huge impact on people locked in and people with severe paralysis.
That’s when we say I started in the setback in the 70s. Now, we are also working on how we can use BCIs for people that are healthy to improve human performance as additional devices to help them perform better in their daily job or allow them to last longer on tasks that are tiring or get insights from their brain activity or controlling computers with their mind. You can imagine BCI as this novel interaction technology and I’m expecting them to make a huge impact on society almost as much as the computer mouse has done decades ago.
Natalia 10:15 That was also my first experience with BCI. When I was studying Medical Physics over 10 years ago in Warsaw, that was the ambition of everyone in the department who was working with BCI, as their first objective was to enable the design of physically disabled people to watch them interacting with a computer and be more functional in daily life. From what I understood, you’re a new kid on the block. You just start working there three months ago, no more than three months ago. Is this your first experience working in the industry or have you also gone for some internships before?
Dr. Davide 11:02 I did mostly consultancy jobs during my PhD. I was always kind of tried to keep some connections with industry because Computer Science and Engineering is a kind of natural transition and the natural way to do whatever approach, whatever research you do, usually has very tight connections with industry problems. I did some consulting in terms of machine learning when I was back in the UK. And even at Harvard, I was having some connections with some companies, local or national companies to how to use machine learning and neuroscience together.
But it’s my first full-time job in the industry. That’s for sure. And I also feel I’m the new kid there. In the startup world, one of the things I learned is that there is no new kid. You’re expected to work full time on the topic from day one, almost. I feel already fully integrated with the company after just three months.
Natalia 12:24 Can you tell us a bit about the differences that you see? I’m sure that you already spotted some differences in the way BCI is done in academia and beyond academia? I would like to hear a little bit about those differences.
Dr. Davide 12:39 First of all, I do believe and I want to say this message out loud that we need both academia and industry for BCIs to make it to the final supermarket shelf or something. Academia is a great environment for doing cutting-edge research and being able to explore ideas that nobody would maybe fund or nobody would believe in. That’s how most of the breakthroughs came out from academia. And then you always need the second part, which is like an industry, or a federal government component that can fund that idea and to make it to develop it. BCI has been around for almost 50 years right now.
It’s unbelievable because maybe people started thinking that BCIs were born just a couple of years ago, or five years ago when major companies were born. And that’s another key difference most of the work in academia is kept not secret but is kept within academia due to the scientific publications that we do because the average person doesn’t read a scientific paper. And so academics keep talking with other academics about what they do. But this technology doesn’t go out there to the end-user.
And that’s one of the key differences for industry and that’s why we need industry. We need companies able to take ideas from academia and bring them to the market and solve all these challenges that necessarily need to be sold to be able for BCIs in this domain to arrive in the consumer market. And so, one of the differences for example, in BCIs, that I faced from academia to industry is that whenever you want to measure the brain activity, in academia, you say I want to target the motor cortex. What I do is put electrodes in there if I want to use non-invasive devices.
I put an EEG electrode on top of your motor cortex. I record and I run some experiments with many different people. And then I publish a paper. If then I want to run a study in the temporal cortex, I put in an extra there and so on. In industry, you don’t have this much freedom of adapting when you’re designing and developing a wearable device as Neurable does. You don’t have this freedom of saying, like, I want an electric there because, at the end of the day, the consumer out there is going to tell you like, I’m not gonna wear something that has an electrode on top of my forehead.
What you need to do is more engineering work on trying to understand how given all the constraints you have from the technology you are developing, how you can make the maximum of whatever you get out of those sensors. Your sensors are placed in a specific area. Now, you develop algorithms that allow you to get that information from the motor cortex in another way. It’s more like, you know, there are more constraints. But at the same time, you are also in higher reward because, at the end of the day, you will have people paying for buying that product. Whatever you build and develop, you see that you have made a kind of a change for the end-user there.
Natalia 16:37 I couldn’t agree more. That’s also one of the biggest advantages of having a company that sometimes when there’s a Corona crisis, you might have an empty fridge once in a while but once every time you’re paid, you’re paid for creating value straightaway, like no one will pay you a penny if they don’t see value in what you do. This’s very straightforward, the way of creating value and getting paid for it not indirect way, like in academia, that you’re kind of promising something to society, and the society doesn’t have some proxy to give you the money. But this’s like a very vague proxy.
And then it’s all based on a calculated guess that you’re even able to do what you said. But this’s a very vague way of distributing value and producing value and distributing funds. I totally agree with you. It’s harder. It’s more sweat. But it’s also more rewarding. That’s 100% right. Can you tell us a bit about what BCI can and cannot do today? There are lots of promises there. And I mean the hopes are high. But in reality, what is possible with BCI, and what is not visible yet?
Dr. Davide 18:03 That’s a great question. First, I think we should start with how we record the data, how can we detect and extract information from brain activity, and there are two words out there. One is called invasive technology which means that to record activity in the brain, we need to inject and insert specific sensors and electrodes into the brain. This allows you to get more precise measurements from the brain activity but as a drawback, many people are not happy to have brain surgery to have those sensors implanted.
There are a number of research groups and companies out there that are trying to develop BCIs using these invasive technologists but they are currently limited by the fact that only people with some severe disabilities would be happy to have that as a sort of a last resort for them to communicate or to control other devices.
With invasive devices, what we can do is like already control a robotic arm with our brain. We can get information on when a specific area of the brain is active or not. And then use machine learning there to translate that pattern that information into saying this person wants to move an arm or a leg. And so we can control robotic prostheses with that, so we can have people with implanted BCIs that can control exoskeletons or prosthetics out there to move a prosthetic arm with their brain. I mean this’s amazing up to a few years ago, we couldn’t even build this type of technology.
There were a lot of engineering breakthroughs that came out in recent years. But again, the huge problem there is about the acceptability of those technologies for people. It’s difficult to believe for me that even in 10 years, 20 years from now, people would be happy to undertake brain surgery to have these devices implanted. You know, to control a computer mouse with your brain rather than your hand, the risks associated with the surgery are also very high. And there are a number of complexities there and costs.
The other word is what is called the non-invasive BCIs. And that uses other types of technologies for measuring brain activity which as EEG or electroencephalography which is a measurement that you routinely do in hospital settings for checking on particular brain disorders. That’s not invasive in the sense that you just put a sensor on top of your brain in certain areas. And what the sensor does is listen to the activity of millions of neurons in that area that when they are activated at the same time because that area becomes active, they fire so they produce a very small voltage, very small current that these sensors can detect.
These devices are non-invasive, so they are better acceptable for people. But they have the drawback that the signal you get is not as precise and as accurate and as detailed as the one you get from invasive recordings. And so the complexity there moves towards engineering like building algorithms to make sense out of these noisy signals. Having said that, there are these two words. I’m currently working on the second one which is non-invasive BCIs. I believe they would probably deliver something for the end-user much faster than the first word because I think we are much better at solving technical engineering challenges these days than the rates or ethical challenges that the first invasive BCI is posed there.
With non-invasive BCI, what we can do is already detect whether we are focused or not, whether we have lost attention. In certain tasks, we can also perform some control like brain control, using your brain activity to control external devices. This usually requires you to do to issue the command several times. You can imagine saying, I want to move my mouse to the right. I have to think about moving to the right again because the sensors are not that accurate. I want to repeat this several times. But it’s something we can do.
And actually, the company where I’m working right now is building these non-invasive BCIs in the form of headphones. And we currently have a pre-order campaign on Indiegogo for this. These headphones allow you to detect activity from the area around the years and translate that activity into whether you’re focused or not or translate that activity into whether you want to issue specific commands to a computer or not, like saying changing your song on Spotify or something else.
This’s already technically possible. And what we are missing right there is just companies that pick up on this research like neuropathy is doing and bring it to the consumer market. There are things that we can’t do on BCI for example, we can’t read the thoughts of a person. The media say, oh, BCIs can read your mind. No, that’s wrong. We’re not even close there. And it’s probably not what we want to achieve anyway. We’re not able to read the mind of people. What we read is patterns in brain activity that are usually associated with certain mental tasks that you do.
If you are now thinking about I want to eat an apple, I can’t use a BCI to learn that from you. But if you repetitively think I want an apple, I want an apple, and then suddenly you change to, I want to go out for a walk, maybe the BCI can understand that you change your brain activity somehow and you change your mental process, but without the ability to say like, I understand that she wants to go for a walk. There are certain overpromises from the BCI world out there that we are not probably able to achieve anytime soon.
But I think we should start from scratch. Can we have today a device that we put on a pair of headphones and allow us to give insight into our brain activity that we are not otherwise allowed to have with any other device? And the answer is, yes, we can have that device. I mean, other companies out there are working extensively to build these devices for many different use cases. And that’s why I think we are living in an era that would allow us to get a lot of different breakthroughs in the BCIs in the next few years.
Natalia 26:26 With respect to the ethical issues, I can imagine because when I was starting my PhD, I was doing PhD in two different labs. One was the fMRI lab, where I was pretty much writing my whole thesis but the other lab was animal work. And in my first year, I had to go for the animal. It was like a two-week intensive course on ethical issues associated with animal research. I had to go through a public exam and it was hard to even get credentials to work with mice when you have any invasive techniques. And that was all that required a lot of ethical discussions. I can imagine that this is very complicated.
I totally understand why. But actually, I have a question about the resolution. You said with some types of BCIs, today, it’s possible to detect what type of movement you intend to make. But what is the resolution? For instance, can we already predict which finger a person can move on a new arm for instance?
Dr. Davide 27:49 We can go down to the single finger right now, again, with invasive devices is much simpler. With non-invasive ones, going down to the single finger would require you to think about moving that finger repetitively for them for some seconds before that movement is able to be performed by a prosthesis. But it’s doable. And whenever we go back to an arm prosthesis there, we also have to think that we don’t always have only a brain activity that we can use the brain to translate this intention.
But we can also have peripheral nerves that we can tap and we can put sensors on. That would inform whether you want to move your index finger or another finger. Another key area of discussion in BCI is particularly applicable in academia, for example, is that when we talk about BCI, we want to have everything controlled purely by the brain activity, otherwise, it’s not a BCI or is called the hybrid BCI. But for the end-user, again, it doesn’t make a huge difference, whether you move your finger because your brain activity said to move the finger or because your brain activity plus your peripheral muscles or nerves said at the same time I want to move that nerve.
But for the end-users, these are technical details. We don’t want to go into the details there. And this’s another difference I noticed in the transition from academia to the industry. It’s focused on solving a critical problem or solving a particular aspect and delivering a product for the user. The path to get there is not as important as actually delivering something. While in academia, you sometimes are more focused on answering one specific question which is at the very beginning of your research path, and in use sometimes lose details on why is important that I answer only this question and I can’t also deliver something as a product.
Natalia 30:10 I understand. Because in academia, it’s important that what you do is what no one did before. And you have an answer to some questions and it doesn’t really matter. What’s the value on, you know, in the grand scheme of things was the value of that answer for the beneficiaries of the society. I think we were just thinking in different terms. It’s being goal-oriented but differently. If I express myself, well, but okay. It’s just an off-topic question for you right now. Because I’m interested.
Recently, I was following what Neuralink Elon Musk’s company is doing. And recently I heard stories that they had some disagreements between teams because Elon put a team of neuroscientists together with a team of engineers. As we just said, they have different attitudes. Engineers want to build and build and launch products or projects, and scientists want to be correct and methodologically correct and just do everything very slowly but systematically and in detail.
According to the old scientific school, there were some major disagreements and tensions because Elon is also more of an engineer than a scientist. He was keeping the sight of engineers, of course, because he invested like 100 million euros dollars of his own money into the company. I can imagine he wants a product. It makes sense. And I have to say that I always believed in Elon, like in his previous projects, I felt that Tesla, SpaceX, and PayPal were all great projects. I was thinking that probably one of his current projects is also like very rational.
Although I hear mixed opinions about Neuralink. And I have to say that after what you started saying about Dogecoin. Recently, I also started doubting his well-being. It’s funny because I used to go to blockchain conferences before the crisis. And sometimes, you know, guys try to pick you up saying, Hey, I have Bitcoin. They just try to make it and I was always mocking them. I was like, oh, I’m team Dogecoin. I was just making fun. Because like those Dogecoin is a joke so they were like, I have Dogecoin too.
They just didn’t get the joke. I was just kidding. I know it’s a scam. But, now I’m like, oh, like it’s the fourth-biggest coin. What is happening here? Isn’t that I’m out of my mind or Elon is out of his mind. Maybe I have no idea or he’s just mocking everyone. This’s his joke and he’s just amusing himself. I don’t know what to make out of this. But with Neuralink, I heard mixed opinions. I heard critical voices. I’m curious. What is your point of view? How do you see the future of the company? And do you think they took too big a challenge or they’re still able to hit older targets in some way?
Dr. Davide 33:45 There’s a lot of debate on Neuralink because whatever, whenever Elon Musk like tweets or says anything, this gets massively picked up by the media, and media has amplified this. They’d say, products or prototypes that Neuralink has done also because of having a strong or a famous head as a startup. Neuralink helps construct media interest. But having said that, from my perspective, what Neuralink is doing is, so neuro link is a very strong engineering company.
What they are trying to do and show is what neuroscientists and academia have shown in the past few decades. It’s possible from an engineering perspective. There was like a few weeks ago, this popular video of the monkey being able to control a video game with the brain and so on. And people are Wow, this is possible. And he’s, I mean, one of the main debates out there was that this has been shown in academia decades ago. But what people maybe have missed and the media didn’t highlight much is that Neuralink, provided input on the table, is that they showed that this was possible from an engineering perspective.
You can build a device you don’t have. You don’t show an effect in a lab in a research lab but we build a device that’s capable of detecting brain activity in real-time sending it wirelessly to a computer and making sure that the computer reacts accordingly. And they showed it in a monkey, obviously, like, when you transition from a monkey to a human, it’s gonna take decades probably, or years or much more than Elon Musk thing is gonna take. I think it’s a great achievement for a local startup.
And obviously, Neuralink can use all this money that they have for doing these types of experiments that are maybe still not seen as interesting from an end-user perspective. You know, the end-user says, Okay, this’s a cool video, but what can I do? I mean, can I buy this technology now? No. Can I implant something in my brain to do the same as the market? No. All these things are all questions that are not answered yet. The other problem, I think, for Neuralink and how neural link is accepted or not by the public, is that it’s an engineering company.
And I think they didn’t develop too much of the neuroscience perspective out there. I remember, a couple of years ago, when they start advertising their positions in the company, they were saying, like, we look for BCI, scientists, engineers, and we don’t need any neuroscience expertise. We are more interested in the engineering one. And this’s something that people have to understand that neuroscience background is still critical when you want to make neurotechnologies that are acceptable and working out there. I think they are doing a great job from an engineering perspective. They’re probably overpromising a little bit about the rigorous work that they do, or the rigor of the work they do.
They don’t like to have the acceptance by the neuroscience community there. They don’t want that. I think it’s still important. I think you still want the academia to check your work as an industry partner. And make sure it’s rigorous. And what we do is we have a number of different scientific advisors externally from major universities that check whatever we do from a neuroscience perspective. And most of us are PhD in neuroscience or BCIs in addition to our background in engineering. I think it’s critical.
I’m also following what Elon Musk does as much as you do. It’s fine, the number of companies. But I’m a bit skeptical about the fact that if you succeed in one company, that means automatically you’re gonna succeed in another company. And also this idea that Elon Musk is an expert across different fields could be like a great expert in solar panels or electric cars. But from there moving to the financial market and moving to brain-computer interfaces, that’s a huge jump.
Natalia 39:07 I think what happens to Neuralink is maybe too much reliance on data science but the point is that to be a good data scientist, you have to combine the ability to program and to understand the understanding of algorithms with some background knowledge. That’s the difference between a data scientist and a machine learning engineer. In today’s job market, only machine learning engineers are valuable and data scientists are not because writing a script that is just taking the data and molding them with the popular algorithms can be automatized.
You don’t need to have a person sitting there and cashing the salary for it. This’s not a value anymore. What is the value now is to know which algorithm works for which data and how the data was collected and what the data represents and have some background knowledge. If you don’t have that, if you don’t understand the brain, indeed, in this case, and its physiology and neurophysiology, then you don’t understand your data. And if you don’t understand your data, then you could be the best programmer.
Dr. Davide 40:24 I think also for machine learning engineers, it’s critical to understand the data. To know, what you are doing, what is the input of your system, and unfortunately, many machine learning or data scientists professionals out there still believe that. You are tasked to build a black box where you don’t have to know anything about how the data comes in or comes out. Your task is just to train these usually complex, deep learning algorithms. And so then you can go and say, like, I’m in machine learning, whatever.
The problem there’s that this again, goes towards potential advice for people who are listening here that are coming from neuroscience or engineering perspective, like the ability to combine two different expertise, two different fields into one person, let’s say neuroscience and computer science is what is going to be critical in the future and moving forward. The ability of not having as a company to hire a computer scientist to be a machine learning algorithms, but to hire a computer scientist with knowledge of neuroscience, if you want to be near technology, or with knowledge in biology, and if you want to build next-generation algorithms for single-cell analysis or whatever.
Having insight and knowledge about what your data look like is going to be critical. And that’s, I think, what is also missing from some other startups that are companies in BCIs, that people think that, you know, because we have so much EEG data or neural data, and brain data out there, we can just train a usually complex, deep learning algorithm and deep learning is going to solve all our challenges without us knowing how. But that’s what I challenge. That’s, I think, why we still need neuroscientists out there that will help us explain why our system work. There’s this concept of interpretability of machine learning which is also applicable in BCIs.
Natalia 42:44 Okay, great. And I would like to now come back for a while to your career since it’s a career talk and not BCI talk. My question will be, do you have any long-term plans continuing the same area, but beyond academia? But do you have any vision for what you want to do in the next 5, or 10 years?
Dr. Davide 43:11 That’s a never-ending tough question these days. One thing I learned is, for example, that a career in the industry moves much faster than it does in academia. And while in academia, if you want to do a career in academia, you expect to do several years of grad school postdoc, and then maybe you start getting faculty positions and so on. In industry, everything changes so quickly, especially in areas like neurotechnologies, where we don’t have many startups out there or companies working on that.
And we also don’t have many professionals out there that have those combined skills I was talking about. Five to ten years for me is like a long plan. Ideally, I would like at some point to move back to Europe and continue my career there. I want to maybe transition to a more senior role in companies as a major role player in building senior technologists. I like building BCIs in your technologies. I’d love to stay in this area.
But at the same time, I understand that if you have like a combined skill of let’s say computer science and neuroscience, you can always move your career to one of the two skills whenever you want. If you’re tired of building BCI, you can still go and get a job in pure machine learning or get a job in neuroscience or in this type of area with minor work. That’s why I think it’s still critical to get these combined skills. But ideally, I would like to keep working on this and see multiple products, possibly coming out to the market and contributing to making BCIs really available to everyone.
Five or ten years is also longer. I don’t know, where BCIs will be in 10 years from now. If we follow Elon Musk for 10 years, we’ll be probably all on Mars finding water there. But for me, maybe in 10 years, we will have maybe 2, 3, and 4 different neuro technologies available out there with different use cases. And again, I think that field of BCI will be active for at least the next 20, or 30 years. There’s room for me to keep working on my career in that domain, I think.
Natalia 46:09 I understand this’s a very hard question for everyone. Today’s job market is no longer stable. And it won’t be. I don’t think it will ever come back to the state it was maybe a decade ago and we are not getting back.
Dr. Davide 46:26 That’s another key difference there, that in the US if you work, especially if you’re working for a startup, the job market has never been stable. It’s never stable. And that’s both a good and a bad thing coming from the European perspective. It makes me difficult to understand. I don’t have a secure job with tenure with like an indefinite contract.
But at the same time, this helps you like saying, if tomorrow there is a company coming out that looks for a senior position, and you want to apply for that, you can make the transition much quicker. In a couple of weeks, you change companies and this thing happened all the time in startups. It also had the employee, always challenging himself or herself there to move forward and to make the next step and to grow up and go to the next step of his career much faster.
Natalia 47:27 Or this also I am planning not to have employees.
Dr. Davide 47:33 Because employees are always unpredictable, right?
Natalia 47:35 I’m just I’m working with subcontractors. And that works much better for me. And I think, I will work with subcontractors and pay them also royalties from projects. And it works better because it also binds them to the project but not in a way that you’re being an employee to someone does. It’s project-based and it works differently. I feel different because there is no boss and employee relationship. All are equal.
That’s I know how employees think about their employers today. And I’m like, I also work with people who are looking for jobs. I am thinking, you know, do I want to be an employer? Maybe one day I am, but I’m consciously turning the company in a way that I will work with a larger group of people but not on terms of employee-employer.
Dr. Davide 48:43 It takes a lot of time and effort to have employees. From an employee perspective, one of the things, for example, that the reason why I chose Neurable Is that they have a strong team and a very happy environment. We work almost like a family. When you transition from academia to industry, you’re scared of being eaten by these huge corporations or something. And definitely jumping from academia to a startup is usually an easier step. But some startups also are like very intense workday schedules and so on.
It’s kind of a red flag out there. And Neurable is investing a lot in making sure that every employee is happy and is having all the resources he needs and that takes time. Going back to your point, as an employer takes time and takes our sources to ensure that all your employees are happily satisfied.
Natalia 49:50 Lastly, I have to like questions that are the usual suspects. First of all, if you could look back to your career so far, is there anything that you would do differently if you had the chance?
Dr. Davide 50:08 I am actually pretty satisfied. I’m not sure I would probably do maybe some more neuroscience training in my bachelor’s and master’s. Back then I was mostly focused on robotics. I learned all of these modeling, algorithms, and technical engineering details that when you move to BCIs, maybe are not as important as when I started my postdoc I had to open go into books and study brain anatomy because I never learned brain anatomy. I had a very strong engineering background in bachelor’s, master’s, and PhD.
If going back, I would maybe choose a master’s or a PhD more in bioengineering at least, to get some ideas and a glimpse into neuroscience as well. But at the end of the day, I think it’s still up to you. You don’t really need a degree to learn these new things, especially these days. You can go back to online courses or other materials online. You can learn by yourself. That’s how I did it. And I think it paid off pretty well. And one thing I would also love to do would be to travel even more.
I moved around quite a lot. But still, I think it would be good too. You know, for example, I did both my bachelor’s and master’s in Italy. Maybe going back, I would probably do the master’s somewhere else to be able to explore academia in another country, as well. Even though I love my master’s in Italy. I really had a lot of fun. And I learned a lot there. I’m totally happy with it.
Natalia 52:11 Okay, and is there any advice you could give to PhDs who are now thinking of the future?
Dr. Davide 52:19 I mentor a number of students on this that want to get into neuroscience and BCIs these days. And, my two or three pieces of advice would be like, first is travel, learn new things and get to know new cultures, and new people because that eventually also helps you work in an international team or help you transition from one area to another one much easier. Visit places, travel abroad and meet new cultures. That’s one key one. The second one is to learn multiple skills. Going back to my point before, I think, in the future, when we need more computational people so having some computer science background is critical.
It’s going to be critical, even if you want to work on pure neuroscience, or impure by origin. In any scientific area, we’re gonna depend on computers more and more. There’s nothing we can do about that. Having computer skills, but also combining computer skills with some other skills, allows you to become a very specialized individual that is easier to be sold from company to company. And the third aspect is to have fun and enjoy what you do. Don’t forget that. At the end of the day, you spend more than eight hours per day, usually working on your staff.
From PhD onwards, even if you’re not sitting in front of your computer, maybe are thinking about your work anytime. And you need to enjoy doing what you do. You don’t want to wake up the moment you wake up in the morning and say, I have to go to work. I’m so tired. I don’t want to go to work. That’s a red flag. And that was a red flag in my career as well. At some point, I was at that stage. And I said, Okay, maybe now I should stop. And I should think maybe I should change something here because that’s not right.
That makes your day more tiring. And it’s a loop. And then you get into bad health as well. You want to enjoy what you do on a daily basis. You want to be excited about what you do. And you want to have fun with what you do. If you don’t, at some point, just take a step back and start thinking am I making the right step into this company. It is a postdoc for this PhD program, or should I move away?
Natalia 55:10 I couldn’t agree more. I think wishful thinking was also my big mistake in my 20s. I always felt that this’s also what they teach us in academia. You suffer because this’s necessary.
Dr. Davide 55:23 That’s totally wrong.
Natalia 55:27 No, I know. But I needed some time to discover that. It’s great. Thank you so much, Davide, for your insights. Thank you for joining us today and for a wonderful conversation on BCIs.
Dr. Davide 55:44 Thank you. It was a great conversation. I really love to be here.
Natalia 55:47 And every one who would like to ask the questions, please find him on LinkedIn. His LinkedIn profile will be linked below. And of course, if you’re interested in this type of content, please subscribe to the channel. See you next time. Thank you.