EMMANUEL GIANNELIS: Good afternoon. Welcome to another lecture in the Charter Day series of lectures and events at Cornell University. My name is Emmanuel Giannelis. I'm a professor of material science of engineering and associate dean for research and graduate studies for the college of engineering.
My role this afternoon is to introduce two of our star faculty in the college. And so, on behalf of our dean, Lance Collins, and the whole college of engineering, I would really like to introduce these two faculty members who will be doing the presentation.
The first is Chris Xu and Chris Schaffer, or as amicably called by their lab team members as Chris squared. Chris came from China. He was born in Shanghai, raised in Shanghai, studied after that at University-- physics-- got his undergraduate degree there. And then came to Cornell to study with a distinguished faculty member, here in the department of applied engineering and physics, Professor Bud Webb, who is, I would say, a giant at the university, with the kinds of things that he's done over the years.
After his PhD degree, he went to [INAUDIBLE] to finish his post doc. And then as a member of the staff and eventually came back to Cornell and joined the faculty here in 2002. And over the years, he's been involved with developing optical techniques for specialized imaging, and more recently for imaging the brain.
Chris Schaffer was born in Florida, Jacksonville, Florida, got his undergraduate degree also in physics from the university of Florida, went into Harvard to get a PhD. And then to post doctoral at the University of California, San Diego. And then he joined the Cornell faculty in 2006.
And his group is also involved in developing imaging techniques to follow different processes in the brain. And I think, these too, basically have a common interest, and as it is the culture here at Cornell, is not very difficult to actually find people who have similar ideas and similar interests and get going on a particular project.
But I like to think that this was only part of what got them together. I think the other important thing is their love for water. Chris Schaffer is a surfer. And he basically tells me that he has surfed on some pretty places, including some places down in New Mexico.
Chris Xu is a swimmer. And he also tells me that he has tried to swim more than the distance between US and China within his lifetime. So we're really talking about people who are not only great scientists but also over achievers. And I think when they start telling you about their work that they are doing-- this cutting edge work they are doing to basically image the brain, I think you'll understand why. With that, I would like to have them do the talking, please?
CHRIS XU: Thank you very much, Emmanuel. This is really a collaborative of work between Schaffer's group and I. And in fact, Chris Schaffer really introduced me to the brain imaging field.
So again, it's really a childhood dream to pursue this. Back in high school, in 1985, I thought there's only two things that's really mysterious-- that is the human brain and the universe. And you can see, I was very much on track when I studied the physics and come to Cornell. You can see, that's me, many years ago, a lot younger. And I was doing microscopy for Biological Imaging-- so very much on track to study the brain.
But then, of course, a detour through telecom field, because there was a lot of money to be made in the early 2000's in the telecom field. So that's a giant detour. But now back, almost 30 years later, back into the brain imaging game. So it's really a privilege and it's very exciting to reengage your childhood dream.
So a little bit about the brain. This is the baseline for the human brain is about one kilogram, and the size of about 1,000 cubic centimeter. It only consumes about 15 watts. If you think about a human brain as a light bulb, it's not going to be very bright. It's only 15 watts.
But on the other hand, it has 100 billion neurons. Each is capable of firing at a 1 kilohertz, at 1,000 times per second. And in addition to that, the brain has one trillion cells and 100 trillion connections. So in that sense, the brain is really like a universe. The numbers are staggering.
Of course, if I want to study brain, I want to study a human. That's very difficult. For example, how many people will be volunteering for me to let me poke on your brain? Not that many, right? So we do awake and behaving animal imaging-- try to use animal models to study how the brain functions.
Now, what do we know? On the one hand, we know really well, for maybe up to 100 neurons how they work. On the other side, we also know, actually, approximately, how animal human behave. For example, if I punch you on the face, you'd be very angry. That's the kind of knowledge we have. It's behavior knowledge.
But in between, you can see a hundred neurons we know pretty well. Animal and humans-- 100 billion neurons. So we have about a billion time the difference. It's a blind gap about 10 to the ninth. That's really the difficulty of how to understand how the brain works.
It's really good to have the biggest boss behind your back. So this is Obama, a few years ago. Two years ago, he announced in his state of union address, our scientists are engaged in mapping the human brain. And of course, a few months later, he announced the brain initiative. So it's actually pretty interesting to listen to him.
- To have the honor of welcoming you here to the East Room of the White House, for a very special scientific announcement by the president. So without further ado--
CHRIS XU: He will comment on his physics grade. It's very interesting.
- It is a great personal privilege and a high honor to introduce our scientist in chief, the president of the United States, Barack Obama.
- And I'm glad I've been promoted to scientist in chief. Given my grades in physics, I'm not sure it's deserving. But I hold science in proper esteem. We thought it was appropriate to have him here to announce the next great American project, and that's what we're calling The Brain Initiative.
CHRIS XU: So that's it. So the Brain Initiative. So what's the difficulty of understand the brain? So if you think about engineering, we are in an engineering school-- you think about a computer. On the one side, we have one transistor with 100 other transistors will make a circuit, for example, an operational amplifier.
And 100 of those circuits is going to come together to make a chip. And hundreds of those chips we'll make into a computer. It's a complicated machine. There's no mystery. We understand exactly how to go from one transistor to a computer. Now we can break it down from the computer to a single transistor. No mystery. Everything is solved by physics.
Now, on the other hand, once again, microscopically, at a few neuronal level, we know things extremely well. On the other side, we can think about complicated human and animal behavior. Now in between, what you really need for this brain initiative is to build the tools to bridge the microscopic observation and macroscopic behavior.
I'm not sure you recognize this bridge. This is Thurston Avenue Bridge. At Cornell, we're very good at building the bridge, so that's what we've been doing. So why optical imaging? That's a tool we picked to use. The most important thing is that photons, optics are, in general, not invasive on animals and humans.
For example, you take a walk, it's pretty nice outside, you come back alive. That means light is not invasive. That's what we need. You want to look at the brain function without killing it, because a dead brain is not very interesting. And optics provides the speed and spatial resolution necessary to look at dynamic functions of the brain and study the brain at the single neuron and neuronal processes level.
So light-- with light the size of 1% of the diameter of your hair-- that's just enough to resolve a single neuron and neuronal processes. So we have both a spatial and temporal resolution to study how the brain works.
But the biggest problem for optics is that optics does not penetrate. If you look at me, I look at you, you look at your neighbor, we see each other really well. But we only see each other at the skin level. We don't see anything beyond that. But that could be a very good thing, right? Because behind the skin that could be pretty ugly.
But the other hand, this is a huge problem. A mouse brain is a one cubic centimeter. A human brain a 1,000 cubic centimeter. If you can only look at the superficial layers, you don't understand very much about how brain works. So for the last eight years or so, our group has tried to take light and make it go deeper into the tissue, going deeper into the brain.
Now what kind of technology do we use? It's an interesting technology with a long history, and being invented at Cornell University. So it's really-- I always find it's very difficult to outsmart the great ones, like Albert Einstein. In his very first paper in, 1905-- that's what-- 110 years ago-- he already hypothesized that it's possible to nonlinear excitation.
Basically, as soon as he realized, light is a quantum in nature, he already speculated that it's possible through non-linear optics, which really underpins what we do here. Of course, it takes a long time. You first need a real theory called quantum mechanics. 30 years or so-- Maria Goeppert-Mayer-- both of them won Nobel Prize, by the way. And you see, no experiment. To do this non-linear optics experiment, it takes another 30 years for the invention of the laser.
In fact, the invention of the laser is also Nobel Prize worthy-- is Nobel Prize winner for Charles Townes. Now, to make this technology practical, it takes the three people here at Cornell, about 25 years ago, to using femtosecond laser and the computer technology to make this non-linear contrast a practical thing for biological imaging.
And a lot of people, including myself, at this point, is also working really hard to make this technology into the clinical. This could be something, again, in the next 10 years or so. So how does this technology work? A very brief introduction. This is the fluorescence molecule. It's a very ubiquitous molecule. And it's also, by the way, FDA approved. You can drink this. You will become slightly green, but no damage-- absolutely.
Now typically, when photon excitation goes like this-- you shine blue photon, so shorter wavelength photon, excite this molecule. So photon plus a molecule become excited state of molecule and a green fluorescence emission. This is a very simple thing everybody knows.
In this case, the fluorescence intensity goes as incident light intensity, linearly proportional. For two photons, it's a very simple idea. So the only difference is that two red photons will combine their energy shining onto this molecule and is getting fluorescence excitation this way. So fluorescence is here. The excitation photon now is longer wavelength, because of lower photon energy, and because two photon is needed to excite this molecule.
In this case, the fluorescence goes as the intensity of the excitation squared. So it's non-linear. So 1 plus 1, in this case greater than 2, in some sense. And it turns out, this longer wave length, redder and redder photons, and not only excitation, in this case two photon excitation, is really essential if you are to imaging deep into tissue.
So this is what it looks like. Again, this is a fluorescein solution. Again, you can drink if you want. Single photon excitation-- blue light in. You can see here's the focal plane. In a single photon case, you can see excitation both above and below the focal plane. In a two photon case, only the focal point is bright.
This localization, I can't explain in too much detail, but this localization-- this localized, three-dimensional excitation, is really essential if you want to see beyond the skin level.
So what is good for this multiphoton microscope, as we just talk about. Now, one is exactly this-- higher spatial resolution, deep into the tissue. You can shoot noninvasive imaging at the spatial resolution less than 1 micron. Again, this is less than 1% of the diameter of your hair. And deep inside scattering tissue, we can do one to two millimeter, at this point.
Now another very good thing we don't have time to talk about, is to move this into the clinical. We can now actually image, for example, human patients with no exogenous stains and potential for medical diagnostics, such as early cancer detection, for example. So we don't have-- we will skip on that.
And this-- really the good thing for this technology, is that it provides functional information of biological tissue in the native environment. Again, if you are to look at the function of a neuron, you want to look at a neuron within the brain, if you take the neuron outside the brain, you can study a lot of details. But again, it's not very interesting. It's just a neuron in a Petri dish.
This is how the microscope works. A laser-- some laser scanning. So we move the beam just like this. You focus into the sample, like a simple dot, like this over here. And see right in that corner-- engineering 25 years ago. And achieved about 20 million or so revenue for Cornell. And the market size for this, we roughly estimate about $2 billion. It's not a small invention right here.
And again, just this year, for example, the pioneers in this field for four people. Three of them, actually, my former crewmates and former Bell Lab colleagues, won the big one million euro Brain Prize, just this year. So it's really the best tool, if you want to image the brain. So this is what we can do.
So showing here is a deep image of the brain vasculature. In this case, the mouse-- we open cranial window. We stain the blood plasma with a red fluorescent tracer. In this case, the plasma is stained, but the blood cell are not stained. So on the left side, is a movie showing the fly through from the surface of the brain into the deep down.
So this is the depth from the surface of the brain. So you can see the blood vessels moving. So each horizontal line is a horizontal capillary. The penetrating vessels are like a dot, like this. And sometimes, you can see the red blood cells. The cells are excluded. The cells are not stained. So you see those features sometimes. You can go quite deep with this technology.
And of course, after you take those stacks, you can do a three dimensional reconstruction. So this is the mouse brain vasculature in a living mouse. So again, this is about one millimeter deep, about 500 microns across. So you can do a three-dimensional resolution of the brain vasculature. So now I'm going to turn the mic to Chris Schaffer who will give you one example of what we can do with this technology.
CHRIS SCHAFFER: Great, thank you. So if we can see down into the brain with cell-level resolution, inside a live animal, what can we do with it? And today, I'm going to talk to you a little bit about some work that mine and Nozomi Nishimura's lab have done over the last few years looking at Alzheimer's disease.
How many people here-- I just want to do a quick poll. How many people here-- raise your hands-- if you have somebody in your family with Alzheimer's disease? Raise your family-- hands, and hold it up. Now I just want everyone to kind of look around and get a sense of how many hands are up. And those of you who have your hands down, they will be up one day. So this is a disease that is really going to affect everybody.
So this disease has been discussed or been-- the first paper on this disease was published by Dr. Alzheimer, back in 1907. So the disease has been with us for more than 100 years. But it's truly the pandemic of the 21st century. And some of the statistics about this disease are really scary. So if you're not already scared of Alzheimer's disease, let me help you out.
So for people over 65, more than 1 in 10 have Alzheimer's disease. By 85 that's half of people have Alzheimer's disease. In 2012, this resulted in $200 billion in direct medical expenses, which given the aging demographic-- this is in the US. With the aging demographics of the US, that will be $1.1 trillion dollars in 2012 dollars, by 2050.
This disease affects not just the patients, but also the caregivers. In 2012, there were 15 million Americans that provided 17 billion hours of unpaid care to their family members. And these caregivers suffer problems, as well, health problems and depression as a result. So this is a very serious disease, both for the people who have it, their loved ones, and for our society.
Currently, there are no drugs that modify the progression of Alzheimer's disease. So there's nothing that can be done right now. Now, Alzheimer's disease is a progressive disease that takes a very long time. There's a pre-clinical stage where there's minimal cognitive effects. But this can go on for decades. And during that time, during those 10, 20 years there's already pathological changes occurring in the brain that are predisposing or leading to the memory problems associated with Alzheimer's disease.
This typically progresses to what is called mild cognitive impairment. In this case, there's notable cognitive dysfunction to both the patient and to their family. This doesn't mean that you just can't remember where you put your keys because you were busy. That's not what we're talking about here. What we're talking about is, you ask a question, you get an answer, and you can't remember that five minutes later, and you ask the same question again. We're talking about those kinds of memory impairments.
About half of the people with Alzheimer's-- with mild cognitive impairment progress to Alzheimer's disease. The other half die of something else first. Full blown Alzheimer's disease, you basically have an impaired ability to function in daily life and need help with just getting through the day. To give you a sense of how bad it can be, this is a Auguste Dieter, the first patient who was diagnosed with Alzheimer's disease by Dr. Alzheimer himself.
In addition to looking dejected right here, if you look in this kind of upper right, this is a piece of paper where she was asked to write her name, and it took her three tries to get her first name, Auguste, but she couldn't remember her last name. She couldn't do it. So that's sort of what this disease is.
Now, to tell you a little bit about how it works, if you look at the brain of an Alzheimer's patient, there is a massive atrophy. The brain is significantly smaller in a patient that has Alzheimer's disease compared to a healthy patient. And this smaller brain is due to a whole bunch of neurons in the brain dying. And it looks like the thing that causes the death of those neurons are these aggregates in the brain called amyloid plaques, that you you've probably heard of-- or senile plaques or amyloid plaques or amyloid beta plaques. And these accumulate throughout the brain in this disease. And they're neurotoxic and kill neurons.
Those plaques are made of a small peptide called amyloid beta, which is basically, it's sort of a waste product of a protein called amyloid precursor protein, which is important in maintaining synaptic connections. This amyloid beta is a piece that gets cleaved off and secreted into the extracellular space. Amyloid beta-- it's produced primarily by neurons, excreted into the extracellular space, and it's dominantly cleared out of the brain through the vasculature.
So all of us, all the time, are constantly making this amyloid beta. It's spewing out into the brain, and it's being cleared out of your brain through the vasculature. And as long as that clearance keeps up with the production, you're OK. But if things change in your life or production goes up or clearance goes down and you produce more of this peptide, then it turns out the more of this peptide that's around, it has a propensity to aggregate.
So if two of these monomers come together, they'll stick. The higher the concentration, the more likely to come together. If they come together and stick, it produces these aggregates, which one, can't be cleared from the brain, and two, which are very neurotoxic, so they kill neurons, and they eventually aggregate into those amyloid plaques that I showed you earlier. So this is sort of the broad understanding of the development of Alzheimer's disease.
But one symptom that has been known about for 30 or 40 years in Alzheimer's disease, but it's remained unexplained, is the fact that patients with Alzheimer's disease have a dramatically decreased brain blood flow. And it's a lot. So patients with Alzheimer's disease, have about a 30% reduction in brain blood flow. That's sort of like-- you know when you're laying down, and you stand up too quickly and you feel dizzy for a second, that's about a 30% reduction in brain blood flow. In that case, it only lasts for a couple of seconds. In Alzheimer patients, their brain blood flow is reduced like that all the time.
So this has been shown in humans. It's also been shown in mouse models of Alzheimer's disease. So these are mice that have been genetically engineered. They carry human proteins that predispose them to develop Alzheimer's disease. And they develop the disease. They get amyloid plaques. They get brain atrophy. They can't remember their name. So these blood flow changes are also an early feature of the disease.
If this was something that was happening at the very end of the disease, maybe it'd be interesting to understand and treat it, but it's a very early feature of the disease, often before there's cognitive impairment even, so that suggests that maybe it plays in exacerbating or initiating role in the disease. But although this blood flow decrease is likely cognitively significant, and later I'm going to tell you it's something that could drive the progression of the disease, it's remained unexplained.
So this is where this really powerful brain imaging can help. So even in human studies, people know that it's not a systemic problem. It's not your heart. And at least until very late in the disease, it's not big blood vessels in the brain. So that leads the question well, maybe it's the smallest blood vessels? But how are we going to be able to see this?
This is a cross section of the brain of a rat where you can see all of these small blood vessels filled with ink. We basically have to be able to map the function of all of these blood vessels in a live animal in order to be able to understand what causes this brain blood flow decrease. But with this non-linear microscopy that Chris talked about, it turns out that it's actually not that hard.
This is an image of the vasculature in the brain of a mouse that has Alzheimer's disease, and the little green dots show a fluorescent dye that binds to these amyloid plaques. So this mouse has-- you can see the blood vessels and we can see those amyloid plaques. And when we looked in this mouse, very similar to humans, we didn't see any problems in the large blood vessels in the brain-- no signs of clotting or constriction or anything like that.
But when we looked at the very tiniest blood vessels in the brain we noticed something. So earlier, Chris told you that we can inject a dye that labels the bloods-- that labels the liquid part of the blood, but not the red blood cells. So when we do that, if you look at this little capillary right here-- let see if we can go back-- so if you look at this capillary right here, you can see these little black dots that are kind of moving frame to frame, and those are red blood cells that are flowing through that blood vessel.
But if you look at this one right here, you'll notice that, that cell is stuck, so that capillary segment has essentially no blood flow. There's something in there that's blocking it and preventing blood flow. So when we compared Alzheimer mice, so here on the left, Alzheimer mice to non Alzheimer mouse, we find in the non Alzheimer and the normal mice, that are only about a quarter of a percent of capillaries are not flowing, at any given time.
So this is different than other organs in your body. So in your brain, basically all your blood vessels are flowing all the time. In your muscles, that's not true. At rest, only about a third to a half of the capillaries have blood flow. And when you exercise, and contract the muscles, some more capillaries open up and you have more flow. But in your brain, they're basically all flowing.
In the Alzheimer disease mice, though, there's an elevation to almost 2% of the capillaries not flowing. Now I know what you're thinking, 2%, oh come on, I still got 98% of the blood vessels. I'll be fine. I'm going to try to show you in a few minutes about how that 2% could actually be quite important because of the structure of the vascular network in the brain.
So about 2% of the capillaries are stalled. One question is what causes these stalls. Why isn't blood flow flowing? So we developed imaging strategies, again using nonlinear microscopy, that allows us to distinguish between leukocytes, red blood cells, clots, different kinds of things that could cause these capillary stalls. And it turns out that about 2/3 of the stalls are caused by a white blood cell that's stuck in that capillary segment, probably due to the inflammatory effects of amyloid beta.
Now let's go back to this question. How much does 2% of my capillaries in my brain not flowing really matter? So it turns out, that Nozomi Nishimura and I, back when we were at UCSD, we did an experiment where we explicitly measured how much blood flow changes if you block a single capillary. So we had a fancy laser based method to occlude an individual capillary segment.
And by tracing red blood cells, we could map the blood flow speeds in a network of capillaries before and after we blocked blood flow in one capillary. And when we did this, we get a relatively intuitive result. So if this is a vascular network, and we block this vessel here-- imagine we look at vessels that are one branch, two branch, or three or four branches downstream from the one we blocked. So this is the blood flow speed after the occlusion is a fraction of the baseline value. If it completely stopped, it would be zero. If it was the same before and after, it would be one.
And one branch downstream from an occluded capillary, the blood flow speed is only about 10% of the baseline value. Two branches, it's a quarter. Three and four branches it's a half. So now let's look with these kinds of blood flow changes, at how much of an impact 2% of the capillaries stalled would have.
So this is the vasculature in a reasonably sized chunk of cortex in a mouse. We used up one under-graduate in manually tracing this. And the blue indicates the location of amyloid plaques. And then, in this mouse, these were the capillaries that showed stalled blood flow. So they had these leukocytes that were plugging them.
These are one branch downstream, so they'd be at 10% of the baseline value. Two branches downstream, so it's a quarter. Three and four branches downstream, so it's a half. So you start to get the sense that because of the topological structure of the network, that even a few percent of the capillaries being stalled can lead to significant decreases in blood flow.
And in fact, if you just do a back of envelope calculation, and sum this up, this predicts about a 20% decrease in blood flow overall. And very recently we've done an experiment where we basically killed all the leukocytes in a mouse, and we measured it at baseline. In this case, there were about a little over 1.5% of capillaries stalled, and we measured the blood flow speed in a few vessels. And then we killed all the leukocytes, the number of stalls went down, and this blood flow came up dramatically.
So why are we interested in this. So there's a couple of reasons. One, I told you earlier on, that much of a decrease in brain blood flow could have a cognitive impact by itself. So if you fix this, it could be that patients with Alzheimer's disease could already-- their brain function could be improved.
But maybe more importantly, it might actually be something that changes the progression of the disease. So there's about a half-- I told you that the way you get Alzheimer's disease is you have too much amyloid beta in the brain and that it tends to aggregate. And those aggregates are neurotoxic. And you're going to get those aggregates when either the production of amyloid beta exceeds the clearance or the production stays the same, but the clearance shuts down.
Now, there's a small-- about 0.5% of cases of Alzheimer's disease are associated with mutations that have been identified that dramatically increase the production of amyloid beta. But about 99.5% of cases of Alzheimer's disease are what's called sporadic. There's no known mutations that increase production of amyloid beta, and it looks like as you age, you decrease the clearance of amyloid beta. Everybody does. Some people it decreases too much, and then they get Alzheimer's disease.
Now the other thing I told you, though, was that one of the dominant clearance pathways for amyloid beta out of the brain is through the vasculature. So if I have the blood flow decreasing by 20% because of these plugged capillaries, that's sort of like plugging the drain in the bathtub. So it's flowing out more slowly. If the amyloid beta is flowing out more slowly, it can increase the concentration and drive the disease.
So this is how we think about this now. So decreased clearance of amyloid beta increases the number of these amyloid aggregates, which cause neuronal dysfunction and death. In addition, those amyloid beta aggregates drive vascular inflammation that leads to leukocytes plugging in the capillaries and causes decreased brain blood flow. This decreased brain blood flow could have a direct effect on neural function. But in addition, it will decrease the clearance of amyloid beta producing a positive feedback cascade here.
Now anytime you find a positive feedback cascade in the disease, it's time to break out the really good champagne, because this suggests that if we could block this on one side or the other, we could do more than just improve brain blood flow in Alzheimer patients, maybe we could delay the onset, or slow the progression of Alzheimer's disease enough that you can die of a heart attack, like you're supposed to.
Now, one thing I should tell you here, though, is all of the studies that I've just described were in the cortex of the mouse, the part of the mouse that's associated with motor function and sensation and higher executive functions, things like that. But the part of the brain that Alzheimer's disease affects first is actually a deeper lying structure called the hippocampus, and that's why it affects memory first.
The hippocampus is the part of your brain that facilitates the formation of short term into long term memory. And we can't study it there yet, because it's too deep down in the brain. So next I want to turn back to Chris Xu, who's going to talk about some very new technology, in the last year or two, that will enable even deeper imaging.
CHRIS XU: Well as usual, Chris always give me motivation to keep pushing our technological development, as you can see. So this is where we were at a few years ago. And we can image deep. We still have-- at a time, we're at like world record depths. It's like, oh wow, that's very deep. Of course, Chris says, not enough. We want to see hippocampus.
So you can see this where we can see the brain vasculature. And we can get very nice images all the way down to maybe about a millimeter. And then you can see the contrast-- 800 microns-- very nice image. 900-- maybe. A millimeter or so and beyond-- almost nothing. What is going on?
It turns out, it's a big problem. That is the brain anatomy. This is a mouse brain, a coronal section. The light green area is the cortical layers, Chris just talked about. Underneath of this, is a fatty layer, is myelinated axon-- very dense. So it's very difficult to penetrate this. It's like a thick layer of fat-- very hard to go through. In this here, lies the hippocampus, which is critical for memory formation.
So how can we do that? It's not like we couldn't penetrate why matter, seems like, with the technology we have. And in addition, since we're physicists, we take a calculation-- you can see, we actually reached the fundamental limit of this technology. If you use this technology, that's a way you can go. You can see the cortical layers and that's about it. So how can we do deeper? How can we do this?
In Silicon Valley, they call this ideation process. They have a quick think, think. So this is what we can do with long wave length and two photon, we can actually create a world record penetration depth. Now what's next then? Long wavelength is very good. Two photons are really good. I give you a hint. Use longer wavelength. Very simple, right? But what else can we do?
That's still not enough. Of course-- three photon excitation. So long as good, longer is better. Two is good, three actually is better. And it really works. So let me show you how this works. So first, you have to create an excitation source for this type of new technology-- the three photon microscopy.
And it turns out, my five years in telecom kind of pays off over here. It turns out that telecom technology, in particular this photonic crystal rod, is really perfectly matched for this kind of laser source. That's what we did. With a telecom laser, telecom photonic crystal rod to generate literally a megawatt peak power laser, and take this going through a microscope-- little bit of optical engineering to make sure the optics are not opaque. We put the mouse over here. This is the brain. The mouse sits right here on this stage.
And this mouse, got a brainbow mouse, essentially all the neurons express different colors. So it's like a rainbow color, and they call this a brainbow mouse. And this technology of staining this brain was genetically engineered for us and proteins and actually won Nobel Prize in 2008.
So this is what we can do now. Again, this is a mammal-- six layer of neocortical layers, about a millimeter deep. And this is called external capsule, the white matter-- very difficult to penetrate. And this is the SP layer for hippocampus, which a lot of place cells, like how you remember your orientation, for example, the neurons are. And this again, showing you the fly through, from the top of the surface of the brain, going down through the layers, going through the white matter and reaching all the way to the hippocampus.
So again, this neuron all stained in red is the red fluorescent protein, label neurons. And then the blue here, actually showing the third harmonic generation, these are the blood vessels, the vasculature. And the scale bar, here, is about 50 microns. So the cross over here is about 150 microns also.
So you can see we can go quite deep. And you can see the white matter will come up, and we can actually go through this, and reach in the hippocampus. And you will see the white matter [INAUDIBLE] the axons, are very nicely through these linear striated features like this. These are the white matters. Actually, for the first time, we can actually visualize these white matters in a living mouse.
And you can see the neurons underneath this white matter. So the first time we can actually break through this wall, going from the cortical layers all the way to the subcortical structure. So three dimensional reconstruction, all the neurons you can see-- very nice neuron bodies, neuron processes, the white matter layer is showing the third harmonic generation. It's not only in the optical technique. So you can see the nice myelinated structures.
And these are the neurons in different layers of the mouse brain. And we can see we have very high spatial resolution is again about 1 micron, again about 1% of the diameter of your hair. Even the z resolution about 5 microns. So we could easily section through individual neurons.
Of course, neuroscientists want to know not just where the neurons are, they also want to see if the neuron is active. They want to see neuronal activity, neuronal functions. The technology we use is called calcium indicators. In particular, genetically engineered fluorescent protein based, calcium probes. So in this case, for example, even 15 years ago, people have proved that every time a neuron spikes-- you can see neuronal spike, like this-- every time a neuron spikes, the channels and the membrane is going to open and the calcium outside of cell is going to rush in. It increases the intracellular calcium concentration.
And when that happens, these indicators will become much, much brighter. For example, the GCaMP we used is about 100 times brighter in a high calcium concentration. And this is very robust for imaging the brain activity. Actually, people in [INAUDIBLE], for example, spent $10 million to literally just to purify and improve this particular [? cause ?] indicator.
And this is what we can do now to see neuronal activity. So this again, is the x, y frame of neurons at different depth. In this particular location, the cortical layer is pretty deep, going to about 1.1 millimeter. I show you two movies about neuronal activity in the mouse brain. And this is about 400 microns in. You can see the flashes. Every time a neuron flashes, that means a neuron is having an action potential. It's a calcium indicator over here.
The next one is a little bit deeper-- 770 microns in. You can see the neuron body, neuronal processes. They're flashing like this. And of course, you could quantify this in a time trace. So this is fluorescence as a function of time. So every time these fluorescence fluctuates, that's an indication of the activity of these particular neurons. And this is literally called watching mouse thinking. Cause every time mouse is active, these neurons fires.
Now this is in the cortical layers. We can also look at the brain activity in the cortex-- in the hippocampus. Again the neocortical layers, the white matter in the subcortical structures. Three dimensional reconstruction of the neurons from the layer 6 down to the hippocampus SP layer. And the blue hue indicating the white matter superimposed third harmonic generation, again by the myelinated axons.
And over here, at about 1.1 millimeter down, this is the hippocampus. And you can see the movie, again say-- watch mouse thinking [INAUDIBLE] deeply. Now we can monitor the mouse activity all the way down here. It's a five minute movie sped up by a factor of 10. So you can see the neuron flashes. It's a very densely of neurons in the SP layer. So about 25 neurons in a field of view of about 175 microns. So you can see the flashes. So literally, we're watching a mouse thinking, at this point.
And now we can also imaging through the bone. Now all the images I've shown you so far before is that we have to do a craniotomy, we have to remove a piece of the skull on the top of the head, two images-- image structure and function below. In this case, you can see a technology, we can also image through the bone. So this is the bone layer of 150 to 200 microns thick. Now we can see the blood vessels underneath the bone without removing this piece of bone. A movie fly through showing here. These are actually called osteo sites.
The thing-- the cells actually makes the bone. And now going through the bone, through the blood vessels, underneath the brain structure. So we also have a technology now to be even less invasive to leave the skull intact while looking at the structure and function underneath.
Of course, we want look at the brain activity in awake and behaving animal. One way to do, is I show you several devices that we have developed, mostly for hospital use. Small microscopes to insert into human. But we can also put these on mouse head, so mouse can carry it around-- with a fiber tether, you can watch mouse activity while it's running around.
Another way is we can also put the mouse with [? hat ?] fixed geometry. So we put a giant microscope on the top, let the mouse moving on a big ball, like a treadmill. So by doing so, the mouse can actually navigate. For example, you can let a mouse watch a movie, so it can navigate to different places or getting feed with food and water, while you can watch it's brain activity. So we look at mouse while it's awake and behaving.
So for this of course, we now can image the hippocampus. Of course, Chris has other ideas-- want to look at a spinal cord, which is equally difficult to look at. So I will let Chris talk to you about one application of this. So for the first time, now, we can actually look at the motor neurons in a spinal cord.
CHRIS SCHAFFER: Thank you. So I know the talk said brain, but really what we meant is central nervous system. And your central nervous system includes not just your brain, but also your spinal cord. And in fact, many of your most critical functions are coordinated in the spinal cord. So just like being able to image into the brain gives us the ability to study disease processes and normal function, being able to image into the spinal cord, would do the same.
A couple of years ago there weren't any good surgical methods to get long term optical access to the spinal cord and the mouth. So my group developed a little procedure that allows us to clamp a small plate over a few vertebrae that are essentially fused to each other with metal bars. And this is looking about three months after this thing has been implanted.
The mouse can-- you can still see the spinal cord, the dorsal vein here that drains blood from the spinal cord and the spinal cord tissue. And these mice do pretty well. This guy is trying to escape here. So they can run around like this for months with that window that gives us the ability to look into their spinal cord.
Now, the spinal cord, that white matter that Chris was talking about, instead of it being down deep in the tissue, it's actually right up on the surface. So as your brain goes down the spinal cord, sort of the gray matter, the part that has the neurons that do the firing, folds underneath, and the white matter, which was deep, comes up on the surface.
And so when we do two photon imaging in a mouse with-- using that chamber, we can see, in this case for example, axons here imaged over the course of almost two months, and blood vessels labeled with an intravenous injection of dye. And using this, we can study disease processes. Here, for example, is looking over the course of about a month at axon degeneration after a very mild spinal cord injury.
So, here, there's a bunch of axons in the spinal cord. We cut them. And then we watch how they change over time. You can lots of things-- some axons that die back back quite quickly, like this one with the yellow arrow, some that died back more slowly, like this blue one takes almost a week before it disappears, some axons that start to grow, here, but grow sort of aberrantly and then die back off.
But one thing I'd like you to focus on is this axon with the little red arrow that persists right at the lesion site, over the course of about a month. I call these Superman axons, and it's a very rare finding. It's only a few percent of the axons that are able to persist right at the lesion site. And right now, with standard approaches for studying spinal cord injury, nobody would have been able to catch these. You wouldn't have seen these axons that are really able to persist right at the lesion site, because they're so rare, and you just have a histological approach to find them.
But we thought maybe these axons that persist well at the lesion site, could be ones that maybe have a little bit more regenerative capacity. And in fact, when we did an experiment where we added a simple drug that just promotes broad growth patterns, we find that only the axons that are able to persist at the lesion site, are the ones that sort of sprout robustly. Now this isn't a cure for spinal cord injury. Neurons sprouting like this is probably going to cause more problems than good things.
But nonetheless, it is suggestion that this kind of technology, the ability to explicitly study the heterogeneity of the response, to be able to study it over time, would allow you to see things that are just not visible otherwise. But that-- but with two photon microscopy, we really can't see deeper than just these axons that are on the surface of the spinal cord. We can't see down through that very fatty white matter to the neurons that sit underneath.
And those neurons underneath are very important, in terms of their function. So for example, when you decide to walk, it's not like your brain says left leg, right leg, left leg, or quadricep, hamstring, quad-- you know, it would be so annoying. You'd never get anywhere. So when you walk, instead what happens, is your brain says walk. And then when your brain says walk, there's a pattern of neuron-- there's some neurons in your spinal cord that get activated by that, and they start this sort of rhythmic pattern of activity. Those are called central pattern generators.
And as they fire, they coordinate the motor neurons that contract muscles in a rhythmic way to produce a basic motion like walking. Right now, it's not very clear how these circuits work. And it's even less clear how they fall apart in disease states, say after a spinal cord injury, or an ALS, or other motor neuron disorders. And the reason is that we haven't had tools to be able to study them in a live animal.
These circuits are studied actually by a very well known faculty member in the neurobiology behavior department here called Ron Harris-Warwick. He's one of the people that sort of discovered these and worked out a lot of their patterns of activity, but he's been forced to do that in preparations where he takes the spinal cord out of an animal and puts it in a dish to study it. So it's not connected to an animal any more. It's not doing-- probably most of it is right, but nonetheless it's not in an animal.
So with this three photon imaging, it turns out we're able to image through that white matter up at the brain surface. This is looking at the vasculature down almost halfway through the spinal cord. And when we have that, we can actually image neural activity. So this is--this data is very fresh, about a month old. And this is the first imaging of neural activity in the spinal cord of a mouse.
So these are a few sensory neurons that sat in the spinal cord. And Yu-Ting Cheng is a graduate student in my lab. He basically pulled on the mouse's foot a couple of times. And when he pulled on the foot, that gives a sensation that's recorded by these neurons.
Now similar to Chris, I think it's only interesting to study these circuits in animal-- in cases where they're doing what they're supposed to do. And these circuits are supposed to help the mouse move. So we really need the capability to be able to do this in a live mouse that's awake and functioning.
And just like with the spinal cord prep, it's possible to train mice to be held by this implant on their spine and run on a treadmill while we're imaging them with this two photon microscope-- three photon microscope.
So just to conclude, Cornell's invention of non-linear microscopy has really been something that's completely revolutionized the study of the central nervous system. Just to give you a sense of the impact, there's over 2,000 of these non-linear microscopes installed in research labs around the world now. And about 1,000 papers per year are published that use this technology to explore not just the brain, but many aspects of physiology.
Cornell continues to lead the world in both the development and application of this technology. We showed you a couple of different stories today. One, this capillary plugging as a mechanism for decreased brain blood flow and Alzheimer's disease. And we also saw you-- showed direct imaging of neural activity in the hippocampus and the spinal cord, both achieved actually within the last six months or so, and both first time in the world.
And I think this is just the beginning. I think this continued focus on building tools that allow us to get deeper, to see new anatomy, that allow us to work in animals that are awake and their brain is functioning while we're studying them, and the ability to get more information per image that we take. There's a number of us here at Cornell that are pushing all of these frontiers, and I believe this will be something that continues to help us understand how the brain works, and understand how it fails in disease.
So with that, I think Chris and I would like to both thank you for your attention and we're happy to take any questions.
EMMANUEL GIANNELIS: Yes?
AUDIENCE: I had heard just a little while ago, that the only definitive way of diagnosing Alzheimer's was after death at an autopsy. Has that changed any? Do you have different means by using some of this stuff that can actually tell when somebody has Alzheimer's other than behavioral?
CHRIS SCHAFFER: Yep. Thank you for the question. So the question is sort of how do you diagnose Alzheimer's disease? And it turns out that being diagnosed with Alzheimer's disease right now is based on a clinical exam. So you chat with a physician, they ask you some questions that assess your memory and other functions, and you're declared to have Alzheimer's or not.
Unfortunately, there's a number of different things that can lead to dementia and not all of them are Alzheimer's disease. And it's not always easy to distinguish between them from a clinical diagnostic right now. The only current definitive diagnostic is postmortem.
There are some diagnostic strategies that are in trials right now that are based on PET imaging of something that binds to this amyloid beta, and they show some promise. But right now the definitive diagnostic is postmortem Fortunately, that doesn't really matter, because there's nothing you can do anyway, so you just--
AUDIENCE: So this is a a follow-up-- so you can't tell whether the blood flow is decreased and shows that the amyloid plaque is building up-- [INAUDIBLE]
CHRIS SCHAFFER: Yeah. So the question is whether or not you could measure that blood flow decrease in humans, and you absolutely could. But I don't think that, that blood flow decrease is going to be specific to Alzheimer's disease. I think that there's going to be many other conditions that will lead to a similar phenomenon. And in fact, we've identified some of those already.
AUDIENCE: When you talked about the leukocytes [INAUDIBLE], has anybody looked at the possibility of vasodilators to open up those capillaries and would have that have any kind of an effect if you put early Alzheimer's people on vasodilator?
CHRIS SCHAFFER: Right. So the question is whether you could maybe use a vasodilator or some other thing to block this early plugging of Alzheimer's disease. We're absolutely thinking about that. I don't think a vasodilator is probably the way to go, because they tend to act in the arterials rather than in the capillary, so you really wouldn't free things up.
And it actually looks like those leukocytes are sticking because of a specific molecular interaction. And in collaboration with Constantino Iadecola who's a leading researcher down at Weill Cornell Medical College, we're starting to explore the upstream molecular pathways that are leading from amyloid beta aggregates to leukocytes sticking. And we hope that somewhere in there, we would identify a therapeutic target. But I think we want to-- I think we want to look carefully in that molecular cascade to find a therapeutic target rather than just sort of trying things.
EMMANUEL GIANNELIS: Let's take one more question on the official program, and then I'd invite you to come down and ask them directly to the speakers please.
AUDIENCE: Just sort of from a lay perspective-- two photon, three photon, one echo to six, but-- MRI, X-Ray-- where does this fit into the scheme of other things that we hear about?
CHRIS XU: Yeah. So first of all, compared to what we do with MRI and X-ray is really-- the difference is resolution and speed. So for example, MRI and X-ray, the best resolution you get for the brain-- if you do fMRI, your resolution probably is one cubic millimeter. That contains about 100,000 neurons. That's the finest resolution can get from those technologies.
And we can get a single neuron--a single neuron-- a process kind of a resolution. And in addition then, the speed is very different. For an fMRI, you'd probably wait for seconds, at least, for the hemodynamics to kick in. Well over here, we can do tens of hertz. So the time resolution is also different.
In terms of ho many photons can we go? We already did some study on this. At this point, three would be really good. It's practical to do and also it probably can give you to the three millimeter kind of a depth. Until we reach that point, I'm not sure we need to go anything more. And also going more, your return is actually getting less. It's like hitting diminishing return-- going from one to two-- huge return. Two to three-- still very big return. Three to four-- the return gets smaller and you saturate, essentially. That's what it is.
EMMANUEL GIANNELIS: Very good. Let me suggest that we thank Chris and Chris for this exciting talk at the intersection--
CHRIS XU: Thank you very much.
We've received your request
You will be notified by email when the transcript and captions are available. The process may take up to 5 business days. Please contact firstname.lastname@example.org if you have any questions about this request.
One of the grand challenges of the 21st century is to decipher the neural mechanisms that underlie brain function and consciousness and to elucidate the causes of devastating brain disorders such as Alzheimer's disease.
In collaboration with neuroscientists and neurologists at Cornell University and Weill Cornell Medical College, researchers in the College of Engineering at Cornell are developing advanced optical tools to enable detailed studies of brain function and dysfunction.
Chris Xu, professor of applied engineering and physics, and Chris Schaffer, associate professor of biomedical engineering, describe these powerful new techniques and the neurological findings they have enabled. Part of Cornell's sesquicentennial celebration, April 24-27, 2015.