Advisor Spotlight: Faccio

What a physicist’s drive brings to the challenges of brain imaging

At e184, we see great potential in the fundamentals. While others build brain-computer interfaces based on existing technologies, we are working on new sensors with game-changing implications. By leveraging cutting-edge physics, we will be able to augment human capabilities without the need for implants.

Among our advisors, no-one exemplifies this aspect of our work like Daniele Faccio. A Professor of Quantum Technologies at the University of Glasgow, Faccio started his career working on optical telecom devices before doing a PhD in high-intensity laser physics at the University of Nice Sophia Antipolis in France. Before starting the Extreme Light Group in Glasgow, he held positions at the University of Insubria in Italy and Heriot-Watt University, Edinburgh. In that time, he expanded his work from laser and optical physics to modeling black holes and testing fundamental aspects of quantum mechanics. Now, a major focus of his lab is on applying the technologies developed there to advance brain imaging, using functional near-infrared spectroscopy, or fNIRS. We asked him about the lines tying this work together, and the developments he sees most encouraging for the future.

What follows is an edited and curated version of our discussion.

Your work covers a very wide range of topics these days, from fundamentals of quantum mechanics to new imaging techniques for the heart and brain. Thinking back to when you started your PhD in 2004, how much of that do you think you could have imagined then?

My training was as a laser physicist doing strong-field physics, working with very big lasers and looking at light-matter interaction at very extreme intensities. As a physicist it’s natural to gravitate toward big questions, and in 2008 a Science paper by Ulf Leonhardt came out describing how one could use the exact same kind of physics I was interested in with these big lasers to simulate black hole physics in the lab. And for me that was a wakeup call. I was super excited about this, and that led to a total pivot in the direction of my research.

When did you start working on brain imaging, and what motivated you there?

I had moved to Edinburgh and started collaborating with Ulf Leonhardt, who was at St. Andrews University, just an hour away. I got some successful grants, some good results, fairly interesting physics, but started at some point to want to move on. This was all about simulating black hole physics. You’re not creating a real black hole in the lab. And so I had the desire to do something a bit more grounded in the real thing.

We’d started to play around with what was back then a new generation of cameras that could detect single photons. Single-photon sensors as in a single pixel had been around for a long time, but the ability to have large arrays of these cameras that could actually take images was relatively new. And we started using those to look at the analogue black holes we were generating. Then, coming back to the point of view “What can we do that’s real?”, we started to think about what else can we do with these cameras.

And they have three features.

One is, they are sensitive to single photons, of course.

The second is, because they are operating what is called Geiger mode, so they’re detecting light in its particle form, it gives a click every time you detect a photon. That can be used to give you very very high timing-precision in terms of when the photon actually hits the sensor. You can start a clock and then you can see when the photons are hitting the detector, and that allows you to then build these movies at a trillion frames per second.

That’s one aspect, and the other aspect is that not only can you have this trillion-frame-per-second resolution, but you can also keep on reacquiring these movies at very high sampling rates. I can take a thousand frames a second, and each of those frames will have encoded a trillion frames per second.

So you have got these three aspects: very high frame-rates, single photon sensitivity, and this trillion frames per second capability. So what else can we do with that?

Then I picked up on some work that Ramesh Raskar was doing at the MIT Media Lab. He was trying to look around corners. A lot of people are working on this now, but back then it was quite a revolutionary concept and they were using a different kind of technology. It was very clunky and they had to take data for a full day to be able to image something behind a corner. I think, with these cameras, we can do this in a second! We can do it in real time. And so that’s what we set out to do.

And I think we gave the first demonstration that non-line-of-sight imaging wasn’t just a curiosity. It is something that could have real world impact. We showed that we could track an object moving behind a wall in real-time. And then we started looking a bit more carefully, and you realized that the same maths that you need, the computational imaging techniques involved, the same computation needed to retrieve the image of something behind a corner or around a wall, is the same that you would need if you wanted to just look directly through a wall.

And specifically, when I say through a wall, what I mean is that many objects that look opaque, including the human head, clouds, the snow, sugar, salt, essentially, anything that’s white, is opaque not because it’s absorbing, because otherwise it would be black. It’s opaque because it’s scattering. And so if you look at a cloud, it’s white at the top, and then it does become black at the bottom. But that’s because all the light in this diffusion process and the scattering has been back-reflected out to space, and so there’s no light at the bottom. But essentially, to the first approximation it is transparent, just very highly diffusive. It looks opaque to us, but using these time-resolved techniques we showed that it is possible.

If there is one common thread in my research all the way from 2004 to today, it is time-domain. Even back then when I was looking at strong field physics, you have these ultrashort 30-second laser pulses, you need very evolved time-domain techniques to be able to capture what is going on. Same with the black holes and optical fibers, and now the same with the brain imaging. It’s that common thread. I know how to do time-domain imaging.

Then the real trigger I think for me was one of those family stories you often hear from neuroscientists. My mother had a stroke, and of course it’s a tragic moment, but as a scientist what really struck me was observing the obvious fact that the brain was intact and functional, but my mother couldn’t articulate words. And that got me hooked. What is going on? How is that even possible? I had as a non-expert in brain lesions always thought that if you had a brain lesion, some kind of malfunction, that would impact all of you and wouldn’t have this sort of separated effect, right? But everything was intact, she just couldn’t speak. Very weird. So I gradually got really drawn into the problem.

So then you look around, what have people been doing and what do they know? And we know a lot. Thanks to functional MRI, neuroscientists now know what ADHD or depression looks like in the brain. Adrian Owen has been able to show that people who have locked-in syndrome, you can ask them to reply yes or no by asking them to imagine that they’re playing tennis, and that’s activating their motor cortex. You can see that. And these are remarkable things in terms of the impact that they can have on people.

But at the same time, the richer countries have the most access to these MRI machines. The US has about 40 MRI machines per million people. The UK, we’re doing less well, we’re doing quite badly, less than 10 per million people. And then the question is, what if I could have a wearable MRI maybe at one hundredth, or why not one thousandth the cost?

In 2024 you had this paper with an amazing title, “Photon transport through the entire adult human head”. What gave you the idea to do to try to do something like that?

I guess it’s the physicist’s sort of taking up the challenge of “You can’t do that”.

There’s a famous handbook in TCSPC, time-correlated single-photon counting, and at some point they say, “Can we use these single-photon sensors and TCSPC for photons that have gone all the way through the head?” And they use the words that you should never use: they said “This is impossible”. I said, “Okay, is it now?”

What we’d been trying to do is, we were taking slabs of polystyrene, this yellow insulating foam that builders use. Just by magic, it turns out that it’s got very similar scattering absorption coefficients to brain tissue. We’re just trying to say how thick can we make this foam and still be able to see inside it. So it’s two and a half centimeters, and then it was working really well. They say, “We should stop here.” I said, “Why? Keep going.” Now five centimeters. It’s still working. I said, “Is it now? So, make it go thicker, 10 cm.” So, how thick can we make this stuff and keep going? And then you see this claim that through the whole head is impossible. So, okay, that may not be true. Let’s give it a shot.

They weren’t wrong. It’s very hard. It was many years, a lot of pain, people getting very frustrated and saying after two years I don’t want to deal with this anymore, I’ve got to work on something else, but we kept on insisting and didn’t give up.

fNIRS today is currently limited to probing just the outer surface of the brain. At the moment, this is the one big advantage MRI over fNIRS: MRI will give you a 3D volume of the full brain, and fNIRS can’t do that yet. So that is one of the reasons for trying to do that: can we see deeper?

Where do you place yourself in the world of quantum technologies?

Anybody really working in this area, you sort of sit in this weird zone where you’re overlapping a little bit with many areas. You overlap a bit with neuroscience, you overlap a bit with engineering technology. You overlap a little bit with the quantum technology, and I was playing around with fundamentals.

This is why in the lab we still have people working on very fundamental questions in quantum mechanics. Can rotation or gravity induce entanglement between photons that otherwise start off their life from completely independent places? Do they become entangled thanks to the action of rotation, or gravity? And one reason we’re still pushing that forward is, the technology that you learn how to build and develop there is actually of interest for what we’re doing.

So I’d say strictly, trying to develop a wearable MRI, this healthcare technologist-neuroscience space is probably not quantum technology, but you need the backdrop of the quantum technology and that drive to understand fundamental questions to feed into what you’re doing with the BCI work. I think what I’m trying to hold on to is I don’t want to forget that the fact that the reason I’m here today is because of that fundamental work that we were doing, and therefore I think it still has a place.

Let’s talk more generally about brain imaging. What’s exciting right now?

For me it’s definitely the time-domain aspect. There’s only a couple of companies worldwide commercializing that technology. You’ve got PIONIRS in Italy. It’s a small company building a fairly small device, but with a fairly high-quality signal-to-noise ratio. Then you’ve got Kernel in the US, that instead has gone down the full-head wearable direction. We have both devices in the lab, and in a sense I think the fact that this technology has got to the point that it’s got sufficient interest that people are willing to fund that, that it can become commercial, that I find exciting because now you got this loop where the fundamental research has led to commercial activity, and thanks to that commercial activity we can now go back and do more fundamental research. So that commercial push, that then loops back into the research, and that at the moment is the thing that’s exciting.

As an advisor for e184, you bring your expertise to advance our research. What attracted you to work with e184?

It was the visionary approach, and the openness. It was clear what they wanted, and they were open to discussing. Money wasn’t really the objection, but the fact that maybe a technology doesn’t exist yet also wasn’t an objection. There was just that clear vision of, sure, the technology doesn’t exist, and we are here to develop it. Because we believe that wearable neural interfaces are the future.