(Cross-posted from here.)

If you read even a tiny bit about brain preservation, you will pretty quickly find people saying things along the lines of “they can’t even freeze a mouse and bring it back to life” and thereby dismissing people in the field as hopelessly deluded.

It will not surprise you to hear that I think that they are wrong. The whole point of brain preservation is that while we are not able to reverse the process of long-term preservation today (trust me, if we could, you would know), we can still attempt to preserve the information in the brain so that if powerful technologies do arrive in the future (which many expect), people in the future might be able to use those technologies to revive people preserved today. An actual argument against this being technically possible requires an argument that (a) the important information cannot be preserved by current techniques and/or (b) such future technologies cannot or will not ever be developed. I welcome any such arguments.

But there is also a related question: Is research towards reviving a preserved mouse the best way to help make brain preservation successful, so that people today have the best chance of being revived in the future? In other words, should we pour most of our resources into trying to cryopreserve a mouse and rewarm it?

My answer to that is also no.

I will try to explain why. But first, I need to give you a bunch of context.

What do other people in the field think about this topic (and others)?

We recently posted a new preprint, in which we collected and analyzed predictions about the future of biostasis/brain preservation from 22 people in the field, who are mostly researchers.

Participants were gathered based on asking the speakers at Vitalist Bay Biostasis Week and their professional networks. So we should expect that they are not exactly pessimistic about the field. Otherwise, they probably wouldn’t be so involved in it.

I will use a few of the results from this preprint to frame the discussion. Of course, my opinions in this post may be different than others in the field, including the other co-authors. This is good to keep in mind, in no small part because I might be wrong. Please consider reading the full preprint.

What are the main methods used in brain preservation?

There are two major types of methods used for preservation today: (a) aldehyde preservation and (b) pure cryopreservation (not using aldehydes).

Aldehyde preservation is the gold standard for structural preservation of the whole brain, such as in immunostaining or connectomics research. On the other hand, pure cryopreservation is the gold standard in IVF, tissue banking, and organ banking for transplant.

(Very infrequently, pure cryopreservation methods are used for preserving large pieces of brain tissue for microscopy studies, for example when researchers also need to do some sort of molecular assay that is currently incompatible with aldehyde preservation. And also very infrequently, aldehyde preservation can be used in tissue banking, such as for bioprosthetic heart valves. But both of these use cases are much less common.)

Aldehyde preservation could not be used as the initial step in a preservation procedure without the use of very advanced future technology for revival. On the other hand, some believe that a cryopreservation procedure theoretically could reach this level at some point in the future, potentially requiring a relatively simple rewarming procedure.

To put my cards on the table, I’m partial to aldehyde preservation, because of the validated structural preservation that this method provides.

How far away are we from being able to do provably reversible cryopreservation of whole mammalian organisms?

Opinions on this vary widely. In our preprint, we asked participants the timeframe by which they thought it would be possible to reversibly cryopreserve a whole laboratory animal and recover it to a long-lived healthy state. On the y-axis, you can see the cumulative percentage of respondents who estimated that this milestone would be achieved by that time frame or earlier:

Participants, by necessity, will make these timeline estimates while having wildly different worldviews on other topics. For example, based on their other responses, some of the participants seem to expect with high probability what is effectively a near-term technological singularity. Among this subset, it’s reasonable to see why their timelines would be accelerated. But most respondents don’t believe that this milestone is likely to occur until the 2040-2050 time frame, and most don’t put high confidence on it until 2070-2100.

I myself find even this distribution of estimates to be sooner than I expect. The problem just seems really difficult to me. First, a whole body cryopreservation procedure would need to be able to cryopreserve every tissue and organ in the mouse body with a single mix of cryoprotectants and rewarm them all to a viable state. But for many of these tissues and organs, we can’t yet reliably cryopreserve and warm them in a viable state even individually, with the cryopreservation procedure optimized for just that tissue or organ. Second, such a procedure would also need to deal with the fact that some parts of the body are vascularized poorly or not at all, and therefore cryoprotectants cannot be easily distributed to them by perfusion. Seems hard. There’s a reason people have been trying this for decades without much success.

Here are the predictions for when people participants thought it would possible to reversibly cryopreserve a human patient with recovery to a long-lived healthy state:

These timeline estimates are considerably further away. They are still faster than I’d expect. But again I’m not factoring in truly transformative AI, which might be an Everest regression — “controlling for altitude, Everest is room temperature” — when it comes to predictions about the late 21st century.

What do people in the field think about other future revival technologies?

Even though provably reversible cryopreservation and rewarming appears to be decades away, that doesn’t mean that the field of brain preservation is necessarily hopeless. The whole idea is that there could be an entirely different technology tree that could eventually allow for the revival of preserved people without requiring the preservation to be provably reversible with rewarming possible at the time of preservation.

What might those future technologies be? Well, we obviously don’t know for sure. In our preprint, we focus on the three classes of methods that have been most commonly discussed in the field: (1) in situ repair, such as Robert Freitas’s proposal of conventional cell repair, (2) molecular nanotechnology (MNT)-based reconstruction, such as Ralph Merkle’s proposal of off-board repair and (3) whole brain emulation, such as Ken Hayworth’s proposal of slicing a preserved brain, scanning the structures with electron microscopy, and building an emulation based on that.

Ignoring all of the other ways that preservation could not work, we asked participants what their probability estimates were that each of these three proposed revival technologies (conventional cell repair, molecular nanotechnology, and whole brain emulation) would ever be technically able to revive any people who were preserved in 2025 and earlier. Here were their probability estimates:

Whole brain emulation and molecular nanotechnology are each estimated at around a 50-60% probability of technical feasibility, and conventional cell repair at around a 25% probability of technical feasibility.

Of course, it’s hard to bring up whole brain emulation in this context without some people immediately pointing out that they don’t consider it to be revival. Which is understandable. This is a common viewpoint and it makes sense that people have strong feelings about it. But it matters, because whole brain emulation was rated in our survey as both the revival method most likely to be developed first and the most technically feasible method for people preserved today. To better understand where our participants stood on this question, we asked them whether they believe that whole brain emulation could allow for genuine continuity of experience:

As you can see, although most said yes, a significant minority in our survey thinks that whole brain emulation could not constitute genuine revival. This percentage is larger in other surveys, such as Max Marty’s survey of cryonicists, which found that around 50% of people would consider it to be survival. (Of course, even if people think whole brain emulation could constitute survival and allow for continuity of subjective experience, they still might not necessarily want it, for other reasons. These questions can and should be separated.)

Of the two main preservation methods, aldehyde preservation is sometimes strictly associated with whole brain emulation, meaning that people claim that it could only work with that method of revival. This is, in my view, a highly unfortunate association, because (a) aldehyde preservation is the gold standard for structural preservation in neuroscience research, which means it makes sense to focus on it if you think (like I do) that provably reversible cryopreservation and rewarming is very unlikely to be developed for humans anytime in the next few decades, and (b) aldehyde preservation really does seem to be compatible with molecular nanotechnology-based reconstruction, which is probably why key proponents of molecular nanotechnology, such as Eric Drexler, Robert Freitas, and Ralph Merkle, have said or implied as much. It seems to me that the molecular crosslinks formed by aldehydes could be reversed in the same ways that molecular damage from ischemia or cryoprotectant toxicity would need to be for molecular nanotechnology to ever be able to revive people preserved today.

So the next question we asked participants is whether they thought that preservation methods that use aldehydes would be compatible with molecular nanotechnology, if such technology were ever developed. The options were “Very likely”, “Likely”, “Unsure”, “Unlikely”, or “Very unlikely”. Here’s how they answered:

As you can see, nearly all participants thought that it was likely that molecular nanotechnology, if ever developed, would be compatible with a type of aldehyde-based preservation. And they also thought that molecular nanotechnology was no more likely to be compatible with pure cryopreservation approaches than with aldehyde-based ones. Perhaps you think I am belaboring this point, but the memeplex has been so contaminated on this question that I feel I must make this point crystal clear, or else some people will inevitably respond to this post by saying that they think that aldehyde preservation is a bad idea, because they don’t want whole brain emulation.

Here are a few of my predictions for the future

So far a lot of this post is me arguing against other positions. So to be more straightforward, here is my current thinking about what I do think is likely to happen with the field of brain preservation in the future, if we humans don’t blow ourselves up first.

First, I don’t think that scaling up provably reversible cryopreservation and rewarming methods that work for tiny animals or small organs is the method that is most likely to lead to revival for most individuals. Especially not for those who are legally die and choose to be preserved in the upcoming few decades.

Instead, I think we will need to rely on the development of powerful future technologies, like molecular nanotechnology or whole brain emulation (I consider in situ repair methods like conventional cell repair less likely to be feasible, because of their inability to perform detailed molecular repairs, although I wouldn’t rule it out).

How long will it take to develop these future technologies, if they are ever developed? I was part of a team that did a different survey of a few hundred neuroscientists about memory. One of the questions was about when they estimated that whole brain emulation would be possible for different species, if ever. The median estimate of these neuroscientists was that human whole brain emulation would be achieved in 2025 — i.e., 100 years from now:

This seems like a reasonable estimate to me. Right now one of the main bottlenecks for brain emulation efforts is proofreading of electron microscopy data, but I think that advances in computation will render this much faster in the upcoming 5-10 years. Instead, it seems like the main bottleneck is going to be understanding how the brain works on a mechanistic level. This is a hard problem. It depends crucially on the level of cellular detail and complexity that the brain uses for different types of information processing, which is unknown. Progress here will be accelerated to the extent that understanding how the brain works can be turned into a computational problem of how to predict electrophysiology from molecular and cellular data. It will be slowed to the extent that it requires a ton of wet lab work.

Like the majority of the participants in our survey, I think that if any revival technology is ever developed, whole brain emulation will likely be the first. But as we have discussed, some people do not and will not desire whole brain emulation, and these preferences should of course be strictly respected. I think that molecular nanotechnology will most likely be the next method to be developed, which could be used to revive preserved people who do not desire whole brain emulation. By the time that whole brain emulation is developed, the pace of technology development seems likely to be very fast. As a result, my guess is that even molecular nanotechnology approaches that seem magical to me today, like Ralph Merkle’s off board repair, will probably be able to be developed.

To be clear, I’m not claiming that these potential methods would work perfectly for everyone who has been preserved. I think that revival would still only be possible with either of these methods to the extent that the information in the person’s brain is actually preserved well enough.

If you agree with my prediction that this is likely to happen in the future, what should you do today? Well, I think one reasonable approach is to mostly ignore this, perhaps aside from signing up for brain preservation services, and do something else entirely. There are a bunch of other important problems in the world as well.

Personally, for whatever reason, I have long had enough conviction about these predictions for the future that I am really motivated to work on this problem. It seems to me like the most useful thing for me to do is to focus my efforts on structural brain preservation, to attempt to provide the best preservation of the information in the brain that we can perform today, to increase people’s chances of revival in the future.

I also want to point out that even though I think structural brain preservation is the most urgent area where work is needed, I still think that provably reversible cryopreservation and rewarming is an interesting research topic. There are a lot of synergies between these areas. Researchers in both fields use similar methods, such as the perfusion of chemicals through the vascular system. Provably reversible cryopreservation experiments in laboratory animals could also test our theories about how memories work, which could help us learn which structures we need to preserve in order to retain long-term memories. Finally, this research might help get more people interested in and excited about the field.

Does it make sense to work on developing future revival technologies today?

Some have asked: shouldn’t we be directly working on making these revival technologies? I disagree with that too. At least beyond basic sketches of how they might work.

To me, researching the application of these potential revival technologies today seems like trying to write a video game before the development of the transistor. It’s just not possible to even get meaningfully started without access to powerful underlying technologies that we don’t have yet. It’s also not possible to make a difference in the rate of development of these fundamental technologies, which are being pursued independently for other reasons that are more immediately economically valuable, unless you have billions of dollars. If someone in the 1930s could imagine the development of video games and really wanted to play them (which, I mean, fair), their best bet would have been to focus on their personal longevity and try to survive 40 or so years. Trying to make Duke Nukem with vacuum tubes would not be a good use of their time.

What exactly do I think we should be working on, then? Well, there are plenty of problems with our current options for structural brain preservation. In one of the other parts of the questionnaire, we asked participants about how likely they thought that different possible failure modes would lead to their information-theoretic death, if for some reason they were to legally die and needed to be preserved within the next month. Here are the probabilities that they predicted for each of these potential failure modes:
 

The three consensus most likely failure modes were inadequate preservation quality even under ideal conditions, geographic barriers preventing timely enough preservation, and poor procedural execution.

To me, most of these seem like worthwhile problems to work on. For example, while I think that the methods we have available today are pretty good, they are far from ideal. No brain preservation method has yet been demonstrated to preserve the entire connectome in a human brain. It would be fantastic to get our demonstrated ultrastructural preservation quality across the entire brain to the next level of validation. A next step would be to do so very reliably, and under less-than-ideal conditions.

I also think that most of the options in this field are too expensive. When I was thinking about this whole topic a few years ago, I really didn’t like that I couldn’t afford to sign up myself, my family, and my close friends who wanted to do it as well. Which to me is one of the points of this whole thing… wanting to spend more time with the people I love. Also, historically people have sometimes lost money near the very end of their lives and therefore not been able to afford preservation at all. This even happened recently to a former president of Alcor! I don’t want to have to stress out about this possibility my whole life, especially when something unpredictable and outside of my control like medical bills could cause bankruptcy.

I think that if it was done locally, aldehyde-based brain preservation could be done for thousands of dollars — and ideally be subsidized by philanthropy for those who can’t afford it — rather than tens or hundreds of thousands of dollars. There’s a reason that people in the US can donate their brains to a brain bank for free. Everything involved is relatively cheap, and there is enough philanthropic funding of the topic by wealthy and powerful people who care about it to cover these costs. A lower cost would make brain preservation more accessible and maybe even help with some of the social stigma around it (it won’t solve the social stigma entirely, but that’s another discussion for another time).

Third, I think it is unfortunate that the gold standard method for preserving the structure of the whole brain, perfusion-based aldehyde preservation, is currently only offered by one organization in the world, which is Sparks Brain Preservation (and which, to disclose an obviously major conflict of interest of mine here, is where I work). It would clearly be better if there were other such options as well, for example for people living far away from us. For this, we need the field of structural brain preservation to grow. I am grateful that Tomorrow Biostasis is open to this method and performs aldehyde preservation in certain cases with high ischemia, and that Hiber and Nectome are planning to start offering aldehyde preservation services in the future. Our own organization has ambitious plans for growth. We also plan to make our methods publicly accessible for other organizations to adopt if they are interested in doing so.

All of these are practical areas where I believe we can make progress in the next few years, if there is sufficient effort and funding.

Anyway, thanks for coming to my Ted talk. Whether you’re skeptical or supportive, I invite you to consider also reading the full preprint. And if you’re interested in helping, please reach out to me.