Aurelia recently appeared on a podcast speaking about Nectome! Full audio is available here; transcript follows.
Nezir Alic
Welcome to season 2 of New Horizons. The first episode today is on cryonics, and the guest is Aurelia Song. She’s the founder of Nectome, which has recently started offering whole-body preservation to terminally ill patients. Aurelia has spent the last decade working on what might seem like a very daunting problem: how to preserve a human brain well enough that, according to our best current neuroscience, everything that makes you, you is still in there. She developed a technique called aldehyde-stabilized cryopreservation, which won both the small and large mammal prizes from the Brain Preservation Foundation — prizes that had gone unclaimed for years, and that the traditional cryonics industry had failed to win. Before that, she studied at MIT.
We’re going to talk about the science of preservation, how it works, why Aurelia believes that memories are preserved, how revival might happen, how this practice might expand and become more integrated into society, and much more. This was recorded on April 20th, a bit of a delay in releasing it. Please welcome Aurelia Song.
Nezir Alic
So there are two different schools of thought, or two approaches to cryonics, right? Can you explain them, and why you’re more drawn to or confident in one than the other?
Aurelia Song
So, I’d say there’s actually a couple of different schools of thought, more than just two. The original cryonics was started in the 1950s, and it came out of this idea of cryobiology. They had very recently just shown the successful revival of sperm — they work after you freeze them at very cold temperatures — and everyone was really excited about that. Can you imagine: it’s the 1950s, and nothing had been invented that could actually preserve life at all. And then suddenly you have this full end-to-end demo where you can preserve sperm, they can recover, they function fine, and it kind of just comes out of nowhere, right? So there was a lot of excitement at the time that maybe just around the corner you’re going to have humans in suspended animation. You’re going to have the ability to preserve organs, whole people — you had all these sci-fi movies that talked about suspended animation. And I think that was very plausible at the time. If you can preserve life, how hard could it be to preserve more cells? You got it working for one; maybe that’s the hardest part.
So the early cryonics was all about: let’s cool people down with whatever the best technology we have now is, and if that’s kind of similar to what we were already doing that’s successfully preserving sperm, maybe someday people figure out a way to reverse this process. And that was kind of the idea — that it’s better than the alternative. And it was in the context of a huge amount of excitement about cryobiology.
So, it turned out that it’s really hard to preserve whole organs and preserve whole people, such that there was a cryonics winter, or a cryobiology winter even more broadly, and this ended up being years and years where we’ve learned how to preserve more sperm and more types of things, like embryos — there’s people walking around today that were frozen as embryos and recovered. But we’ve really been having a lot of trouble with organs, for various reasons.
My point of view is coming from neuroscience and information theory. So I say, we want to preserve people — and what does it mean to preserve a person? What it means to preserve a person is: if you have different people and you preserve them, they become different artifacts. That is the information-theoretic criteria for preservation in a nutshell. And I’m much less concerned about biological viability, as measured by assays you can do today, and much more concerned with preservation of information. And traditional cryonics starts with saying, let’s make sure the cells are still alive in the sense that they can function metabolically, and just get better and better at that until it’s fully reversible — and in the meantime, use whatever the best thing we currently have. So that’s one kind of approach to thinking about these differences.
Nezir Alic
So if you’re focusing more on the information approach as opposed to the biological one, doesn’t that make revival more difficult?
Aurelia Song
So, taking a step back — what is it that we do at Nectome, right? We are preserving people that are terminally ill. This is as part of the end of their life, it’s whole body, and it’s preserving them with the goal of eventual revival. And so this is something that is only for terminally ill people. It’s something that we’ve studied for 10 years, and we’ve learned that in order to do it properly — and properly here is defined as: can you preserve all the synapses? Can you preserve all the neurons? — in order to do that, it has to be done very quickly postmortem, and we found that you have about 12 minutes to do this successfully. And we figured out protocols that can be done successfully. And so this year we’re starting to actually make this available, where people who are terminally ill can come and they can be preserved.
Now, the question here is thinking about: what does revival look like? What does a scenario like this look like over time? And from my point of view as an archivist, I want to preserve people well enough that our current neuroscience says they’re still there, in information-theoretic terms.
Nezir Alic
I think a big part of your case rests on the idea that memories are stored in physical brain structure, as opposed to, like, electrical activity. How confident are we that we understand where memories are stored, and that this process that you do at Nectome would preserve that?
Aurelia Song
Yeah, so the question is: we’re really good at preserving structure. How do we actually preserve people? We start with chemical fixation, and this cross-links proteins into a solid gel within seconds. You start using fixatives — in about 60 seconds or so, proteins are being cross-linked and all of this structure is retained. And the question is, okay, we can preserve structure really well. We can preserve structure so well that it looks like a neuroscience textbook. But is that good enough? Or is there maybe some type of dynamic activity that we’re not preserving that is a major issue?
And I think a priori, it could go either way, right? If you don’t know anything more about how memory works, I think it’s very reasonable to start by saying, well, we’re very dynamic creatures. All of our neurons are always polarized, they’re always firing. There’s all this activity that’s happening. And if you have a process that stops all of that activity, maybe it’s sort of like the RAM on a computer, where you turn it off and it loses all of that information — and yeah, the physical hardware is there, but the actual pattern of information that you cared about is gone.
And I would say, for this, we actually understand really well that that’s not the case, because we can shut that activity down artificially and people survive this. So this is one of my favorite surgical techniques — it’s actually my most favorite surgical technique in the world. It’s called deep hypothermic circulatory arrest. There were these stories for a long time of people falling in frozen lakes, and they would drown in these frozen lakes, and they’d sit there for an hour or two, and finally people would recover them. And in certain conditions, they would make a full recovery — even though when they pull them out of the lake, they’re not breathing, there’s no heartbeat, they really look like they’re dead. But they’re cold. This is the origin of the phrase “you’re not dead till you’re warm and dead.”
So some doctors in the 1950s who were heart surgeons had this really frustrating problem where they knew how to fix certain issues — like aortic arch repair — the main vessel that comes right off your heart — with fairly simple surgeries. But they couldn’t actually do it fast enough to save the lives of their patients, and they couldn’t really do it while the heart was beating. And so it must have been very frustrating to know you could save someone’s life, and it just be a matter of like 10 or 20 minutes — where if you could just have their heart stop and do the surgery in about 10 to 20 minutes, you could fix it. But if you actually did that, then they wouldn’t survive the process.
So they looked to these stories of these people drowning in frozen lakes. They said, well, let’s do that intentionally, but in a really controlled manner, and see if we can buy the time that we need to do the life-saving surgery, and then get the person back. Does that make sense? You cool someone down, like they drowned in a frozen lake, do the heart surgery really fast, warm them back up — like we’d recovered people from the frozen lakes — and maybe that would work. And it turns out that just works.
So, you can cool people down to fairly low temperatures — 12 degrees Celsius, 16 degrees Celsius. Regular body temperature is 37 degrees Celsius, for reference. And after you cool them down to around 18 or so degrees Celsius, their brain activity stops. If you have an EEG that you use to measure brain activity, it’ll become what they call electrocerebrally silent — no activity that’s detectable by that mechanism. And once you have electrocerebral silence, that’s actually the goal. That’s the criteria where they know, okay, the brain’s not going to be consuming very much energy in this state. So now the heart can stop, and you can do your surgery in about 20, 30 minutes, and then you can fix everything back up together, warm them up carefully, and they’ll survive.
So, it is reasonable to imagine that we could have been like RAM in a computer — where if you cool someone down, yeah, you might be able to bring them back, but all of the attractor states are lost and everything that makes them who they are is gone, and there kind of wouldn’t be anybody there. If that was the case, I would expect people going through deep hypothermic circulatory arrest to essentially have their memories erased. And we find exactly the opposite: that you apparently can shut this stuff down, and then it finds itself again once it warms back up.
This is very convenient for preservation, because it leads to the pretty well-founded conclusion that whatever you need to preserve about a person, it is static. It’s a structural type of thing. It’s got to survive DHCA. It’s got to survive deep anesthesia. It’s got to survive ischemia. It’s not just DHCA that shuts down activity — there’s a lot of other things that do it. If you look at electroconvulsive therapy, for example, people survive that with very little cognitive effects — sometimes some beneficial cognitive effects. And this is not the thing I would expect if memory was dynamic. It’s the thing I would expect if memory was structural. That doesn’t necessarily mean we’re preserving enough structure — for that you need more neuroscience evidence — but it at least makes the entire thing plausible.
If DHCA didn’t exist — or in fact, if DHCA had been tried and we found that it erased people’s memories — then I wouldn’t be doing chemical preservation, because obviously that’s not going to work. I would be doing something that could try to preserve dynamics. It’d be a lot more complicated. We probably wouldn’t have it for 50 more years.
Nezir Alic
Wait, you mentioned — I thought you were fairly confident that enough structure was preserved, because I’ve seen microscope photos, and I thought you concluded that there was good reason to believe that enough structure is preserved for memory. Or is that not certain?
Aurelia Song
I think that we can be positively certain that we can preserve enough structure for memory. So the first part of the argument is, do we need dynamics or not? And the answer is, no, we don’t need dynamics. So we need to preserve some type of structure. And the question is, well, what type of structure do we need to preserve? And for that, you need to look at two things. You’ve got to look at chemistry, and you’ve got to look at neuroscience.
And the overall framework here, again, is that a preservation technique works if it keeps separate things separate. So you have two inputs — maybe they’re very close to each other. You could imagine a person and an exact copy of that person, except they have one single long-term memory different. And then you consider preserving both these people. If you can maintain whatever the physical difference is that creates that difference in memory between those two people, through your preservation technique, then that’s an adequate preservation technique.
So again, a priori, for any particular preservation technique — does it work or not? This is something you need actual knowledge about, for chemistry and for neuroscience. And so the relevant thing for how good are you at preserving structure is fixation chemistry, and the relevant thing for, okay, what is the kind of structure that you need to preserve, is all of the neuroscience that we’ve done over many decades.
The argument we’re trying to put together here is, can we preserve a person, right, in a way where they are reasonably well preserved and we could expect to recreate them or revive them in the future? And first is, what even is the framework to decide whether a preservation technique works? And I would say the correct framework for you to use for that is information theory. It is, do you keep different inputs different? And the test that I would put forward is this kind of minimal memory difference test. If you imagine myself and an exact copy of me, with one long-term memory that’s different, and you preserve both of them — have you kept whatever physical difference there is between those two people that makes them have different behaviors? Have you maintained that through preservation?
So, for example, I know your name is Nezir, okay? And if you ask me tomorrow, who was that person you spoke with? I would say, well, his name was Nezir, and he wore a blue shirt. And you could have worn a red shirt, and you could have introduced yourself as David. And so then that version of me would have said, well, it was David in a red shirt. And even if you performed deep hypothermic circulatory arrest tomorrow and then woke me up from it afterwards, I would still remember who I talked to the day before.
So there’s some physical difference between those two alternative versions — let’s say, Aurelia-red and Aurelia-blue. And we have this preservation technique — is the preservation technique adequate to preserve that physical difference? Keep the things that are different, different. So you may first think, well, do the dynamics matter? We know they don’t, because of DHCA. So then the next question is, all right, how good are you at preserving the physical structure? And then, how subtle is the difference in the physical structure that you need to preserve? And if it’s the case that you preserve at a better resolution than the brain itself is using to encode information, then your preservation technique is very likely to work. And that is what I believe, based on all of the evidence I’ve seen.
And there’s kind of two angles we’ve got to approach it through, and eventually they meet. One is, how well are we preserving structure? And two is, how subtle of a change in structure does the brain need to use to encode information? And they’re both interesting. I’m wondering which one you want to start with first.
Nezir Alic
Start with whichever one you want.
Aurelia Song
Okay, so on the one hand, let’s talk about how good is the preservation technique that we have now, when it’s properly applied, right? So, things that we know about it include that it preserves practically every protein in the entire body. So you’re made of cells, and the cells are in turn made of proteins. And proteins are the kind of business end of the cells. They’re the molecular machinery that actually does most of the actions of life. And they are very densely packed inside cells.
When I started this, I kind of imagined that the insides of cells were mostly watery, and, like, there were a few proteins floating around doing their thing. And the reality is, proteins are packed so tightly into this aqueous environment that it’s kind of a miracle we aren’t solid blocks of polymer to begin with. If it was any other type of polymer, it probably would be solid. And it’s sort of already on the edge of being solid. Think, like when you warm up an egg — it becomes solid, but you didn’t change any of the molecules hardly at all. You just kind of jostled them around, but just that is enough to solidify it.
So when you add a chemical like glutaraldehyde, which grabs all the amines of the proteins and binds them together, it hardly takes anything at all to make all of this stuff solid. And then, once so solidified, all these biomolecules are stuck. So the DNA is stuck, the lipids are stuck, the proteins are bound together and stuck. So pretty much everything that the cell was made of before it was preserved is still there afterwards.
And how do we know that it’s still there? There’s a couple of different papers that talk about this that are really fascinating, but I would point to the entire field of immunohistochemistry as probably the strongest evidence, because neuroscientists, for many years, start with fixation as the first step. And then they use antibodies to label some particular protein that they’re looking for. And pretty much every time we compare — here’s what a protein that’s tagged with some type of fluorescent dye looks like when it’s alive, and here’s what that protein looks like after fixation, when we use antibodies to look for it — we find that the signal’s essentially the same. And when it’s not the same, it’s often because the fixation itself disrupted the antibody, but you can recover this by using different antibodies, or through partial digestion or other various techniques.
When they measure the dry content — like, just total amino acid content of cells after they are fixed — they find that that’s the same as if they quickly freeze cells. So you might wonder, hey, does a lot of stuff get solubilized and removed? And the answer to that, we definitively know, is no — fixation basically captures it all. So at a molecular level, fixation mostly captures everything that you care about.
And then at a level above that, a kind of microscopic level — fixation does very minimal disruption to the structure of the brain, and of cells in the body in general. And we know this because we have electron microscopes. So you can preserve a brain, and you can look at the individual cells that are preserved, and you can see, ah, here’s, like, synapses, here’s cell bodies, here they are connected in a way that makes sense.
There’s some very beautiful work that’s correlative light and electron microscopy — super-resolution light microscopy — where they can see individual synapses with two-photon microscopy. And there’s one particular paper that you might want to link to later, where they looked at one particular neuron in the brain of a mouse, and they mapped out the synaptic connections that it had along a certain length of its axon. And this is while it was alive. Then they fixed that brain, and they very carefully found that exact neuron that they had imaged during life, and they did an electron microscopy panel. A real tour de force of microscopy. You can imagine it might be quite difficult to find it again after you’ve viewed it and fixed it and then processed it. And this looks exactly like you would think it does. The one that’s alive and the one that was fixed — that structure looks the same. The one that’s fixed just has higher resolution, because they were able to use electron microscopy instead of the light microscopy.
So at a molecular level, you’ve got all this stuff. At a structural level, we’re also able to show that the structure is still there. And that’s what fixation does, right? It takes something that’s a living system, and it basically glues the proteins together. And then that resulting form is very durable, and it’s still kind of everything there the way that it was.
I sometimes analogize it to, like, taking a watch and injecting epoxy into it. So a watch has all this — a fine clockwork watch with, like, a bunch of gears and stuff in it. So, on the one hand, none of those gears are moving anymore. But all the stuff that made the watch work is still there, and it’s still preserved, and it’s still basically exactly where it was when it was functional. And so while the epoxy disrupts the function of the clock, it doesn’t really destroy information about the watch.
So that’s where fixation is. We have a thing that can preserve microscopic details, that can preserve molecular details. And the question is, is that good enough? Right? And again, a priori, we don’t know until you bring in neuroscience evidence. So the question that’s really important at this stage is, what physical differences does the nervous system actually use to encode information in the first place? It could be that it’s incredibly subtle, right? So subtle that it’s beyond the level of proteins, right? And so we think we’re preserving structure, but really the real signal is something that we can’t even see. And so the preservation — we don’t have evidence whether it works or not. It could be that way.
What does neuroscience have to say about it? Well, I would say that we don’t fully understand how the brain works, of course, but we understand a lot about how memories are formed. We know a lot more than most people appreciate about the biomolecular process of memory creation, to the point where we can do some really cool stuff. Like, we can selectively erase memories in mice. We can create false memories by manipulating the brains of mice. And the way we do this is by synapses.
So synapses are these femtoliter-sized connections between neurons in the brain. And they take about a third of the volume of the brain — like, a third of your brain is connecting neurons together with these types of things called synapses. And these are molecular machines that sure do seem to encode information, because when you learn new information, the synapses change. This is something that probably the most mainstream neuroscience says: in order to create a long-term memory, synapses have got to change. They install more ion channels, they physically change in size. We find that if you tag the synapses that change when learning is taking place, and then you reset them back to their original size, it erases that one memory, and not other memories that were learned alongside it. Synapses, I would say, are the smoking gun for this.
And more to the point, if there was a much more subtle thing going on, what I would expect when we look at brains is that we wouldn’t see any real changes occurring at all when memories are learned. So, for example, if we looked inside people’s heads, right — who knows what we’re going to see before we start learning things about it? It could have been that there’s not really a brain at all, it’s just a homogeneous liquid, right? And yet we still learn things, we still do things. I would say, well, I don’t know how to preserve that. But when we look inside people’s heads, we see a quadrillion little switches, called synapses, that clearly change size in response to memories being learned, and that, like, do sensible things when you mess with them, right? Like with mice and other experiments.
This, to me, is highly indicative that the brain is using sort of femtoliter-sized assemblies of about a million proteins, changing size, to encode information. And not just one synapse — you have thousands of synapses changing to reliably encode a memory. And this makes sense, okay? Like, here’s a kind of cool thing that I think is amazing. Let me ask you a question — let’s see how fast it takes to answer. You ready? What is your name?
Nezir Alic
Nezir?
Aurelia Song
Okay. So, like, you did that in, like, a second, which is kind of crazy, because the neurons in your brain work at 60 hertz. So you have, like, 50-ish clock cycles to retrieve that information — and not all the other bits of information that you know — and, like, figure out how to move your mouth in a way to say that, and say it to me. We don’t know how to build this. Like, we cannot make computers that can do that in 50 clock cycles. We can do that, but doing it like that? Not really.
So, if you think about what it would take to do that, it’s really astounding. You have got hundreds of billions of neurons in your brain, right? And to get that, that quickly — it would be like one person in America saying, all right, the whole country needs to do this thing, and not these other things we could be doing. And the whole country, on a dime, just being like, yeah, okay, well, I’ll do that — in lockstep. Really amazing coordination. So it makes sense that you need a lot of amplification to be able to respond like that quickly. And synapses have got the right sort of reaction speed, right, and the right sort of properties to — out of a quadrillion synapses — have a few thousand actually drive the behavior of the animal that they’re part of, quickly.
So, yeah, that’s the whole of the argument, right? We start through information theory. We say, you keep different things different. We reject the need to preserve dynamics, because of the clinical evidence. And then we look at what we’re able to preserve. We look at what is likely the shape of the physical trace of memory, and we see that we can preserve more detail than we expect we need, given what we know about kind of what the physical trace of memory looks like. And so we can conclude the preservation probably works. And that’s really how I think about it.
Nezir Alic
So it seems probable that you’re able to preserve the information. You mentioned, though, that the fixative stage kind of glues the proteins together — isn’t that problematic for if and when people try to reverse the process in the future, or are you banking mostly on uploading the mind?
Aurelia Song
I’m agnostic as to how we get people back, right? What I want to do is I want to do right by the future. I want to give them all the information that they need. And I have faith that the future will figure out how to access that information, even if it seems hard from our perspective today.
The analogy I would use here is with many of the successful historical preservation projects that have come before me. So in the 1970s, there was this project called the San Diego Frozen Zoo. Have you ever heard of it?
Nezir Alic
No.
Aurelia Song
So these guys — they had discovered DNA in the late 1940s, early 1950s. And in the early 1970s, I think 1972, these guys were like, hey, what if we started preserving DNA? Okay, we don’t know how to read it. We don’t know exactly what it means. But it’s obviously the molecule of heredity, right? And we know what its molecular shape is. It’s a polymer. And we can preserve that with a variety of methods, but cold is a particularly convenient way to do it.
And so long before we sequenced the first genome, these guys just started preserving animal DNA. And they didn’t have to say, here’s exactly how you’re going to scan this DNA. It was enough for them to say: DNA is a polymer, and we know that our preservation technique is not going to scramble that information. And so we choose to preserve it now, because these species are going extinct now. And we know that if we don’t preserve them, the future is not going to be able to access it.
So I think today we’re in the same position that we were in in the 1970s with regards to DNA. Like, neuroscience in relation to preserving people today is like how we were with DNA in the 1970s. Do we understand everything about how memories work? No. But do we understand enough to be quite confident we can preserve the important parts of a person? I would say yes. So we just preserve it, with whatever technique we have that we can validate actually preserves it. And the future is pretty smart. They’re going to figure out a variety of ways to access the information, whether that be a type of destructive scanning process, or whether that be some type of mechanism of undoing the cross-links and biologically restoring what’s preserved.
The important thing is you’ve got to have good inputs. That’s what I’ve really learned doing electron microscopy. It’s garbage in, garbage out. And I want to give the future something that is good.
Another analogy I’d use is these things called the Herculaneum scrolls. Have you heard of those, by any chance? So there were these scrolls a long time ago, and this volcano erupted, and it buried these scrolls in the pyroclastic flow of the volcano, and it carbonized them. It basically turned these things — it was a library — into charcoal. Importantly, it didn’t incinerate them. They were charcoalized, right? And then they were buried. So for hundreds and hundreds of years they lay buried. Eventually—
Nezir Alic
Yeah, yeah, I’ve heard of it. They found some way to look through the scrolls, some kind of imaging — I forget exactly how.
Aurelia Song
So a long time ago, these Catholic monks dug some of them up. All right? And they were excited, because you could see the writing, right? It was black on black, but it was a slightly different color of black. The charcoalized ink looked slightly different than the charcoalized paper, right? And so it was tantalizing, because you could tell these were ancient books. These are ancient scrolls that actually had a lot of cool knowledge on them. And you can sort of maybe read the outer parts of them, but how on earth do you read the whole scroll?
So these Catholic monks, who were very smart, using the best technology of their day, tried to do it. They took animal hides, and they laid them out, and they put light glue on them, and they took the scroll and they tried to, like, roll it on the glue to unfurl it. You know, it would break the outer layers, but if you did it just right, then maybe you could get something. And they destroyed a few of these scrolls trying to make this work, and they found that they couldn’t do it. They just could not get the pages to line up properly, so it was all scrambled.
And they had a choice then about what to do. They tried with the best technology of the day and it failed. And so do you, A, destroy the remaining scrolls, because they’re taking up space and you’re confident that the future will never be able to do anything with them because you couldn’t think of a way to do it? Or do you persevere, and do you preserve these scrolls and wait, and have faith that the future will be able to do something?
And fortunately for us, these Catholic monks decided that, in their humility, they would wait. And they died not knowing what the scrolls were, and the next ones after them died not knowing what the scrolls were, and so on and so on for many, many generations, until finally, this last decade, we’ve been able to use X-rays to look at that ink and see the different layers and digitally unfurl the page and look at it.
Now, it would have been a mistake for the Catholic monks to say, well, I can’t think of a way to do it, and so therefore it can’t be done — this is impossible, therefore just don’t do it. And so that’s how I think about it too. Like, yeah — although I can think of ways it can be done, so I’m in an even better position. It has partially been done on animal brains and insect brains — not completely, of course, but we’ve made great progress in scanning connectomes, and even doing very crude simulations of such connectomes.
So I would turn it around. I’d say, like, we’ve got this preservation technique that’s so comprehensive — according to all of our mainstream theories of neuroscience, this works. And so the question is, do you want to bet your life that no one will ever figure out how to access this information? Or another way of putting it is, nobody controls the future. Like, you can’t tell the future what to do. You can’t invade the future. When you choose to preserve yourself, it is simply a matter of whether you choose to deny the future the access to the information, or whether you choose to retain the option for the future to access that information. So, yeah, that’s my thought on that.
Nezir Alic
Yeah, makes sense. It’s definitely better to save it just in case, and then later on, who knows what can happen?
Aurelia Song
It’s surprising, man. You know, I wish there had been a tradition back in the day of, like, if you made a new library, you would take some of the books and you would just pour beeswax on them and then toss them in a lead box and just throw it into the ocean or something, right? As a kind of gift to Poseidon or whatever. I wish they would have done this, right? Because if they’d done that — you know, yeah, that library burned down, but we could read that stuff today. No problem.
Nezir Alic
One of the biggest problems with cryonics is this 12-minute limit that you mentioned, where you have to act really quickly. Therefore, this isn’t really applicable to the majority of deaths, unfortunately, because most people die unexpectedly and in messy situations where you can’t just have Nectome or whatever other cryonics company come in within 12 minutes and preserve them. Could you tell us why that limit is so fundamental, I guess — if it is fundamental — and if there might be any way to extend it?
Aurelia Song
Yeah, okay, so what’s up with this 12-minute window? When I started this project, I thought that people would probably have a two-hour window, okay? Meaning that if you die, almost regardless of how you die, as long as you start preservation procedures in about two hours, that would provide adequate preservation — that a group like the Brain Preservation Foundation would say it preserves memory. And I tested it.
So I ran a human brain bank, and I got human donation cases that died under various conditions with various postmortem times, and I perfused their brains. And then I looked at those brains comprehensively. And what I saw is that those brains failed to perfuse, meaning there were huge regions of them that would be visible by eye that did not receive any preservation chemicals. And so this was obviously a huge problem. If you’re not even getting chemicals to part of the brain, then you’re not going to preserve that part of the brain well.
So then I studied this on animals for many years, working out: how long of a postmortem interval can you have and still get good brain preservation? This means you get preservation chemicals everywhere, and when you look at synapses, they look good. And I found that that limit is about 12 minutes. If you do your preservation up to 12 minutes, it basically works perfectly. And then after 12 minutes, you start to get progressively worse preservation, and it’s pretty catastrophic. It’s not like it’s slightly worse everywhere. It’s more like a whole regional failure, where a whole hemisphere will fail.
And I tried a bunch of interventions to try and make this work better. Most of them don’t work. I suspect that the nature of this problem is such that there’s probably 30 things that are happening around that 12-minute mark. And so fixing just one or two of them doesn’t really move it very much. And that’s very unfortunate. If we had a different biology — like, it could have been the case that my original guess of two hours was true, in which case this would be a lot easier thing to do. But that’s not the world we live in. So, yeah, you gotta work with the biology that exists. And that dictates how the procedure’s got to look. And so it’s got to be a planned type of thing. It’s got to be something that people are terminally ill, and then they do this process.
The other thing I learned from doing these experiments is that the manner in which a person dies matters a lot. It’s not uncommon during end of life for someone to have a coma before they die. And this coma can last for multiple days. And almost definitionally, if you have a coma, you are not getting adequate perfusion to your brain to maintain consciousness. And so this can be damaging. I’ve seen situations where people have actual necrosis in their brain after about a week-long coma, right? And the whole hemisphere is really very heavily damaged, but they still have a heart that’s beating. They’re still kind of alive. And even if you could instantly and magically preserve someone the moment after their heart stopped, in that type of case, the brain’s already been so severely damaged, it wouldn’t be recoverable. So that’s the current constraints that we live under.
And that dictates a protocol. And that’s where we’ve got to start. But that starting point is still something that I’m pretty happy about. Like, it didn’t have to be 12 minutes. It could have been one minute, in which case there would not be enough time to do the preservation procedure. So we’re lucky that it’s 12 minutes. And with regards to its general applicability, I would say most people have enough advance warning that if they want to choose preservation, they could. Because you get a cancer diagnosis, right? And then it’s like, you’ve got multiple rounds of this, and you can do preservation.
Nezir Alic
But — I’m not sure about this, but don’t most people die sort of in old age, in air quotes, which means, I don’t know, stroke or things like that, aneurysms? So a lot of it is pretty unpredictable. I’m not sure, though.
Aurelia Song
Stroke is pretty unpredictable, but if you look at the actual numbers on this, I would say the majority — not like the supermajority, but the majority — of people would be able to choose preservation if they really wanted to. They would have enough time and enough forewarning to do it.
Ultimately, I hope that as a society we start taking this a lot more seriously, and I could imagine a world where every ambulance is able to do a field preservation, and then you probably would be able to successfully preserve people in a more emergency context. But in order for us to get there, this idea has to be something that is more generally seen as desirable and important.
Nezir Alic
That’s something else I wanted to talk about. So although cryonics has been a thing for over 50 years now, it hasn’t really scaled at all, and there’s been not much growth to it. Only several hundred people have been preserved in all that time, I think. How do you envision this going from just a niche thing for rationalists in the tech community to something more mainstream, maybe something that insurance covers and hospitals offer?
Aurelia Song
So I really think that preservation is for everyone, and that this is going to become something that is just considered the basic thing that we do for the future, right? This is something that becomes obvious — that it’s something the future wants. And it’s just a beautiful thing that we can do to respect the people who are dying, and to provide the future this wonderful gift: they’re going to have the option to welcome these people back. And of course they’re going to want them back, because they’re just our children.
I really think the route to doing this goes through extreme attention to the details, and actual regulation about this. So to elaborate on that a little bit: right now, the question is, does this stuff really work or not? And I think in order to make this become a real thing, it’s important to have regulation around, was this preservation performed well enough that it likely works by our current understanding of neuroscience, or not. I want to make preservation into a rigorous field, where people go to college to be a preservationist, and there’s laws about it, and there’s licenses to do it. And there’s rigorous quality assessment when someone’s preserved, so that the family knows whether they are preserved well and are likely to still be there, or whether something went wrong. And then people would make informed decisions based on the track record of companies doing this.
So we’re currently working with this organization called the Brain Preservation Foundation to launch actual regulation around this stuff. In order to do that, the BPF is going to analyze our protocol and confirm that it works on animals. And then every person we preserve is going to have a small biopsy that the BPF will independently analyze, and they’ll determine whether it’s traceable or not — meaning the connections are still there. So we haven’t tried a truly professionalized version of this yet, and I think that’s part of the pathway towards getting this to work.
Separately, I think there is going to be a lot of progress in the coming years around simulation of animal brains, and recreation of animal brains from preserved brains. And I think as this progresses, it’s going to be a clear existence proof that it’s possible to get people back from this type of preserved state, even if it’s too complicated to do now. The analogy would be when we scanned the first genome from an E. coli, right? In the 1990s. I think people mostly realized, yeah, someday we’re going to be able to scan a whole human genome, even though it’s a lot bigger than an E. coli genome. And we’re right on the edge of having, I think, some truly impressive results with uploading — recreation of the mind of animals from their preserved brains. And you’re seeing some early versions of this, like with the FlyWire dataset, where they mapped a whole fruit fly connectome. Well, that fruit fly was preserved with glutaraldehyde. So that’s really strong indication that that stuff is still there. And our simulation of this is still very crude, but we get some pretty compelling results, I would say, already.
Nezir Alic
If you go the uploading route, though, something about that kind of bothers me — like, would it really be you? Like the whole sense-of-self idea, as opposed to biological revival. I feel like whatever I am is the physical brain. If you were to just upload it, that would be helpful for, say, my family and friends perhaps, but I wouldn’t be continuing to experience. Who knows, though?
Aurelia Song
Well, what sort of evidence would you count for or against the truth of that?
Nezir Alic
It’s a tough one. It’s possibly not even an empirical question, because then we’re asking whether something is conscious or not.
Aurelia Song
Isn’t that interesting? For something so important, why is it so hard to make it into a well-founded scientific question? Because for a lot of other things that are really important in your life, that’s doable, right? I’d really encourage you — think about it. Like, what would you think would constitute evidence one way or the other for this type of thing? Like if it’s a hundred years from now, and people are like, oh, we got the answer to that, right? What does that look like?
Nezir Alic
I don’t know. I feel like you can’t rely on anything other than self-reports of people who have, I guess, been uploaded, and then perhaps they would say, yes, I still feel the same way. But we can’t really be sure, you know? Just like when Claude tells me that it’s conscious.
Aurelia Song
So some people I know worry about whether anesthesia preserves consciousness, right? And some people worried when they first did organ transplants — like, if you got someone else’s heart, then are you still really you, right? And so, over time, we’ve come to realize, yeah, that is the case. And I would say the type of evidence that we de facto actually use is just what you said: do they still seem like they’re the same guy afterwards?
With deep hypothermic circulatory arrest, right, you’ve got a person who’s been cooled down to the point where they are mostly just an inert physical object. Then they’re warmed back up, and they come alive again. They’re able to talk, and they’re empirically able to remember all their memories. And so, at least clinically, I would say we accept that that’s the same person before and afterwards.
Nezir Alic
There’s just a difference of locality, though, you know? Like in those situations, nothing has been moved — it’s in the same location, it’s the same set of atoms. Of course, I could be wrong, but it’s just an intuition that I personally have. And I would certainly take the opportunity, if I were dying, to upload my brain, because it could be the case that there is continuity of experience.
Aurelia Song
I think for something that important, it’s probably worth thinking about what would even convince you one way or the other, right? Because at least for me, if I find a belief in my head where I do that calculation and I say, how would the world have to look in order for this thing that I believe — for me to believe in it more, or for me to believe in it less, right? Like, for example, the synaptic basis of memory. If brains could encode memories and they weren’t physically changing synapses, right? That would be evidence against it. Or like the durability of memory, right? Like if DHCA reliably erased some types of memory and not others, that would reduce my confidence in it, right?
I find that when I have a belief that doesn’t actually admit to any change regardless of the empirical evidence that I see, that that is actually evidence that my entire way of thinking about the problem is not serving me well — that there’s a different framework I ought to be using, where it would admit to traction with reality. So it’s that feeling, right? Where it’s like, I already know in advance that nothing will convince me one way or the other. I like it when I have that feeling, because I can take that as a signal that my framework’s wrong. Like, despite the framework seeming reasonable, when it gets into that mode, that’s, to me, a signal to take a step back and consider how I need to tweak the framework until it stops doing that, right? And, you know, frameworks can seem really reasonable, but depending on exactly how they’re used, you can get into this problem.
You can make up all kinds of interesting problems along this vein. Like, how do you measure the value of a dollar bill, right? If you just have the dollar bill, and you’re trying to physically measure it to find where the value is physically in the bill — okay, you might find that it’s very challenging to do, to the point where it maybe makes sense to think about value differently than physically existing inside the actual dollar. And perhaps we could have a different view of identity.
But that being said, people who are preserved, I think, if they’re preserved in a way that preserves information — that’s the best chance for whatever type of revival technology is available in the future. So at Nectome, when people want to be preserved, we interview them, and we ask them very carefully what sorts of things they want us to do as we are custodians of their preserved body. And so the two most important questions that come up — which the people who’ve signed up so far are about evenly split on — is, first of all, chain of custody. If it looks like we’re going to lose control of your preserved body, because there’s a big war or because of some other unforeseen situation, do you want us to cremate you, or do you want us to try to preserve you at all costs, even if we’re going to lose control? Half of people say, no, I want the chain of custody respected. Half of people say, under no circumstances do I want to be destroyed — I’m happy to enter a totally unknown future.
And the other most important question is, if there’s a method of revival that’s like a destructive scan or some type of copy operation, should we use that? Or do you want to have some type of restrictions about using that? Some people say, I wouldn’t want that being used under any circumstance. Some people say, well, if most people in the future feel like the philosophy has been adequately addressed, then I’ll get what they got. Some people say, I don’t want it, but if it’s been a hundred years and nobody can figure out how to do it other than that, then sure, whatever.
Nezir Alic
Yeah, it would be the second or third camps. Like, I would like people to wait a long time, and only when they’re really sure that there’s no better way than to do it that way.
Aurelia Song
So, in any of these cases, our job is to understand what you actually want, and record that really carefully, and then faithfully do that in the future.
Nezir Alic
Very cool. Do you think cryonics could become less important pretty soon because of AGI — it’s expected, like, anytime now — and also longevity escape velocity, so people won’t be facing death nearly as often as we have in previous generations? Like, it’s likely that most terminal illnesses will be solvable by the end of the century. So maybe it’s more of a short-term solution, or it won’t be needed as much in the future.
Aurelia Song
Well, hey, for people your age, for people my age, maybe it’s a situation where we never need preservation, maybe not. Maybe we get really significantly powerful AI and it solves a lot of these issues. Maybe we don’t. But everybody’s got parents, and there’s a lot of people that definitely are not going to make escape velocity. If you’re 85 today and you’ve got cancer, you’re not able to wait for any of these things.
The question is, do you want to deny the future the ability to access your preserved body, or do you want to give that to the future? So for now, I think preservation is clearly the best thing of the bad options that you have when you’re terminally ill. And I would be delighted if preservation wasn’t necessary anymore. And I think eventually it probably won’t be, but that could be a few years or that could be 30 years or who knows? So we’ll see.
Nezir Alic
I was thinking it’s kind of analogous to transplants, where I also think that that field is kind of a temporary band-aid — where, say, in 20 or 30 years, they won’t really be doing transplants anymore. They’ll be taken over by just regrowing organs and regenerative medicine.
Aurelia Song
Oh, yeah, I mean, one day we’ll be able to do a lot of stuff, but sure, it’s good to have it now. I got into preservation largely because it seemed really tractable. It seemed like something that could be done relatively quickly, with a lot less effort than many of the other alternatives, like figuring out some method to solve aging or figuring out some method to upload someone or figuring out full artificial intelligence. Preservation seemed really tractable. It’s like, let’s just lock all this stuff down and do a really good job and confirm that it works. And then that’s a valuable technology.
Nezir Alic
Yeah, there weren’t enough people working on it for a long time. I’m not sure why, but it was a pretty, like, empty field.
What should someone do now if they wanted to get preserved? I think I noticed you’re starting to make it available for $225,000.
Aurelia Song
So we are currently, until the end of the month, offering a special kind of early deal for people. This is before we have our preservation center fully online. And so, you know, there’s risk on your part, and that’s why it’s a fairly low price. But it’s a discount card.
So you pay $20,000, and if you hold onto it for 10 years, it’s full preservation. If you use it sometime intermediate, it’s linear cost savings — so use it in 5 years, it’s 50 percent off. We’ve had a lot of interest in that. Many people have gotten one. It’s something that’s available if you go to our Nectome Substack, which we can link. And I think it’s a pretty good deal. I definitely encourage anybody who’s interested in preservation — this is probably the best deal we’ll ever offer. So, you know, something that they might consider.
Nezir Alic
[unclear — question about an early supporter in Scotland]
Aurelia Song
He didn’t get the card. He was a really super early supporter, from really quite a long time ago, and we definitely are happy about that.
Nezir Alic
How did you arrive at that number, the $225,000? Like, what portion goes to the actual procedure versus long-term storage?
Aurelia Song
Yeah, so for this type of thing, we set aside money to maintain someone in a preserved state so that they would still be preserved even if Nectome went out of business. The entity that keeps preserved people is separate from the startup, and it’s exposed to a lot less risk. And we put enough money aside to do 100 years of storage, plus some margin. And that’s money set aside that goes to an endowment that would pay for 100 years of storage with the interest. And then the rest of it’s kind of split between Nectome and legal defense and advertising — you know, the stuff you’d probably imagine.
Nezir Alic
I see. Do you expect it to go down over time with technology?
Aurelia Song
The cost itself, I think, does go down with a lot of scale. What I would ultimately hope to be the case is that preservation would eventually be considered basically a human right, and that it’d be covered by medical insurance or national healthcare, depending on the country that you’re in.
And that it’d be something that ambulances would be able to do in the field. Or if you have a really risky surgery, you could crash into a preservation instead of just dying. I think all this is possible. But right now, we have a thing that will work for people, which is: a preservation center, and you’re terminally ill, you come, we do a whole-body preservation.
Nezir Alic
Yeah, I would like to see it become a norm for sure. I mean, it makes sense. Like, even if there were only like a one percent chance of working, you would still want to throw all of your money at it, everything you own, because it’s a chance to live for much, much longer.
Aurelia Song
I actually feel a little differently about it, weirdly enough. When I first got into cryonics, I was like, this is one of the coolest ideas I’ve ever seen. But I feel like if I do the math on how likely it is to work, I got about 0.1 percent. And to me, that wasn’t interesting enough to actually put money into it. But it was interesting enough that I was willing to dedicate my life to making that number higher. So for me—
Nezir Alic
Yeah, I’m not saying it is one percent, by the way, I was just saying, like, hypothetically, even if it were one percent. I’m sure it’s higher.
Aurelia Song
For me, it’s the type of thing where it’s like, what do we want the standard of care to be for people we’re preserving? And I think the standard of care should be: our current understanding of neuroscience says they’re still there, and we can verify it with some type of imaging modality. I think we can and should meet that standard. And if you do meet that standard, I think it’s a lot more than one percent likely to work. I think it’s something that is more up to the future, but we did a good enough job preserving that we gave the future the option.
Nezir Alic
How do you compare it to some other companies in the space, like Alcor and the Cryonics Institute, and I think there’s Cradle and Sparks Brain Preservation?
Aurelia Song
So the differences are: we are exclusively whole body, and we’re exclusively scheduled. I haven’t found a protocol that will work as an emergency procedure, and so in the absence of something that we can confirm works, we don’t offer it. So Nectome only does scheduled preservations that are done through medical aid in dying, where you’re terminally ill, you come to us, and after you do your medical aid in dying medication, within 12 minutes afterwards we do the preservation. And then we get external validation that it works. So those key factors are the core differences. And again, that is: whole body exclusively, scheduled exclusively, and external validation.
Nezir Alic
I feel like Alcor kind of doesn’t make sense because they’re not located in a state where there is medical aid in dying, right? So what do they even do? Do they expect long-term transportation to not be problematic?
Aurelia Song
Well, to be fair, Alcor started when there was nowhere in the world where you had medical aid in dying, so it doesn’t matter where they were. And I believe they moved to Arizona, where they currently are, again before medical aid in dying was legal in any state. So they were all kind of about as good as each other. But I would say, yeah, that’s starting to become an issue now. I wouldn’t do this in Arizona, because I wouldn’t be able to do the protocol that I’ve got.
Nezir Alic
It’s getting more and more common, I saw — medical aid in dying, like, more US states, more countries. So that’s good.
Aurelia Song
The trend does seem to be towards increasing autonomy for a person over their own fate and their own body.
Nezir Alic
And so you’ve been working at Nectome for about 10 years. Over that long period of time, what’s been the biggest challenge or biggest obstacle in progressing the technology?
Aurelia Song
That 12-minute window, I think, had to be the most annoying technical challenge, because it wasn’t clear what it even was starting out. I had to discover that, and it took years of work to figure out what the limits are in mice and rats, and then it took another year or so to get the surgery fast enough that it would work in pigs, which are a reasonable facsimile for a human.
The first version of this — I actually hired a surgeon to do it, and it took 22 minutes. And it didn’t work. It was exactly the same sort of situation as with the rats when you go 22 minutes: you have catastrophic perfusion failure. And we spent a year and really just shaved time off of every little part of the protocol that we could and optimized everything. And we got it down to 4.5 minutes. And that does it with a reasonable margin of safety. So that was something I was really proud of. And I was able to hire a team at Nectome that was able to do that.
Nezir Alic
It’s all very impressive. It’s good that you’re very, like, rigorous with the science. I hope that cryonics and Nectome sees a lot of growth in the next few years, next decade or so.
Aurelia Song
I think it’ll be a lot of fun. And I think, when people realize preservation is possible, I think it’ll open up people’s planning horizon. Even if you’re young and you don’t need preservation anytime soon, the fact that you’ve got an actual sensible answer for, like, what do you do as you approach the end of your life — that leads you to think, yeah, I might still be around in a couple of hundred years. Whether you’re religious, whether you’re not religious, no one really seriously plans for being around here 200 years from now. And I think if people started to do that and take that idea seriously, we would have a world that just feels a lot better — where people do long-term projects.
Nezir Alic
Yeah, I would say I kind of do, with AGI and longevity stuff, but good cryonics is another option for it.
Aurelia Song
Well, we’re starting to get a couple of people. Yeah.
Well, it’s been a pleasure speaking with you, and, yeah, I’d love to take a look at the podcast once you get it put together.
Nezir Alic
Yes, sounds good.
