(Cross-posted at the Effective Altruism Forum)
Previous EA discussions of this topic: here, here, here, and here. Note that these primarily focus on cryonics, although I prefer the term brain preservation because it is also compatible with non-cryogenic methods and anchors the discussion around the preservation quality of the brain. See here for more discussion of terminology.
This post is split up into two sections:
(a) Technical aspects, which discusses why I think preserving brains with methods available today may allow for revival in the future with long-term memories and personality traits intact.
(b) Ethical aspects, which discusses why I think the field may be among the most cost-effective ways to convert money into long-term QALYs, given certain beliefs and values.
In this post, I’m not discussing whether individuals should sign themselves up for brain preservation, but rather whether it is a good use of altruistic resources to preserve people and perform research about brain preservation.
Technical aspects of brain preservation
1. What is the idea behind brain preservation?
a. Brain preservation is the process of carefully preserving and protecting the information in someone’s brain for an indefinite length of time, with the goal of reviving them if technologic and civilizational capacity ever progresses to the point where it is feasible and humane to do so.
b. Our society’s definition of death has shifted over time. It depends upon the available medical technology, such as CPR and artificial respiration. In the future, the definition of death will almost certainly be different than it is today. One possible improved definition of death would be when the information in the person’s brain that they value is irreversibly lost, which is known as information-theoretic death.
c. Pausing life without causing information-theoretic death could be done with a long-term preservation method that is not yet known to be reversible today, but which has the goal of preserving enough information in the brain so that it could potentially become reversible in the future with improvements in technology. This is brain preservation.
d. With current methods, we can potentially preserve enough structure in the brain over the long term to retain the information for valued cognitive functions like long-term memories. There are multiple possible methods to attempt to accomplish this, each with upsides and downsides.
2. What in the brain is necessary to try to preserve?
a. It is already possible to stop electrochemical neuronal activity (in humans) and biological time (in other animals) without loss of long-term memories. The cognitive functions that most people care about seem to be encoded by static structures in the brain. More on this here.
b. Adequately preserving the brain alone would be enough to retain the information for long-term memories and core personality traits, because it is the only part of the body that is known to be irreplaceable without massive effects on this information. More on this here.
c. Despite currently lacking complete models, we can use existing knowledge in neuroscience to evaluate the hypothetical process by which structural information for valued cognitive functions could be mapped in the future. More on this here.
d. A wealth of evidence suggests engrams are encoded in neural structures distributed across the brain. More specifically, it seems to be the distributed activity of neuronal ensembles communicating through the biomolecule-annotated connectome that instantiates long-term memory recall. More on this here.
e. Because of the correlated nature of structural information in the brain, it is likely that there are numerous topological maps of the biomolecule-annotated connectome that could retain the information needed for long-term memories. Even if many of these maps were damaged or destroyed by aspects of the brain preservation procedure, if at least one could still be inferred, then the information content would still be present. In the future, our inference capacities are likely to improve dramatically as a result of AI, making the inference of sufficient structural information in the preserved brain much more plausible. More on this here. (This is assuming that humanity survives the transition to transformational AI.)
3. What are the methods we can use to try to preserve structural information in the brain?
There are numerous possible ways to do so. Broad categories include:
a. Unprotected cryopreservation (also known as “straight freeze”). In the absence of cryoprotective agents, ice will certainly form during low subzero cooling and will mechanically crush cellular structures, leading to damage to the connectome. The question is whether that damage is enough to cause information-theoretic death. There is some evidence that structural preservation quality following unprotected cryopreservation may not actually be that bad (for example, Aschwin de Wolf discusses this here), although this is one of the most controversial questions in the field, and others vehemently disagree. Neural Archives Foundation offers this option in Australia, which costs AUS$30,000 (~$22,000 USD).
b. Cryopreservation with cryoprotective agents. By perfusing cryoprotective agents into the brain, ice formation and associated damage can potentially be prevented. Unfortunately, getting cryoprotective agents into the brain is a difficult problem. Also, this causes dehydration and associated damage to brain structure. However, Greg Fahy’s team at 21st Century medicine has reported that they can achieve good preservation quality with their cryoprotective agent perfusion protocol; for the most recent details on this, see page 496 of Robert Freitas’s book. Cryonics Institute offers this option for ~$30,000 for people who live locally in the Clinton Township, Michigan area.
c. Aldehyde-stabilized cryopreservation. Aldehyde stabilized cryopreservation was introduced as a whole-brain preservation method in 2015. It has two main steps (to oversimplify quite horribly): aldehyde fixation, followed by cryopreservation. The fixation step seems to help with the distribution of cryoprotectants to the brain. The reasons for this aren't quite clear to me but likely involve quickly stabilizing the structure of blood vessels and/or cell membranes.
For this method, 21st Century Medicine (where Robert McIntyre and Greg Fahy worked) won the Brain Preservation Foundation prizes -- the small mammal prize in 2016 and the large mammal prize in 2018. These prizes were for showing that the method could preserve the microstructural anatomy (i.e. the connectome) of whole brains using a method that would be capable of storage for the long term -- at least 100 years. Clearly, aldehyde stabilized cryopreservation is an extremely promising approach for people who prefer brain structure preservation. Oregon Cryonics is currently offering an option similar to this for $28,000.
d. Fixation followed by long-term liquid preservation in aldehyde. Aldehyde fixatives act to preserve tissue primarily by covalently cross-linking biomolecules, mainly proteins. This dramatically strengthens natural gel-like structures in the brain such as the cytoskeleton and extracellular matrix and has been shown to cause gel formation in brain tissue. Gel formation is a way for a liquid to have solid-like properties and stability. Vitrification and gel formation actually share many biophysical properties.
Might gel formation alone be sufficient to preserve brain structure for the long term? There are some studies suggesting that neuronal morphology is maintained for decades when preserved even at room temperature in formalin. One study found that neuronal morphology was well-preserved in brain tissue stored in formalin for 50 years. Another study found that there was “excellent preservation of fine and even cellular details in the tissue” in brain tissue preserved in formalin for ~80 years. Formalin only came into use in the 1890s, and there isn’t that much brain tissue preserved this way for such a long time, so it’s unclear how long this tissue would maintain its structure. There are also a lot of biomolecular alterations resulting from formalin storage, although they may not be sufficient to cause information-theoretic death.
Oregon Cryonics is currently offering liquid preservation in aldehyde for $1000. When combined with the fees of a local funeral director or pathology specialist to do the initial preservation, the total cost for this is likely around $3000-10,000. Note that the average funeral cost in the US is around $7000.
e. Fixation followed by plastic embedding. For example, brain plastination. This offers the potential for more stable long-term preservation at ambient temperatures. As far as I know, this is not currently offered anywhere in the world.
Fixation-based options (#3-5) are potentially more scalable because they offer the possibility of tapping into the existing brain banking infrastructure, which exists in much of the world. However, doing so also depends on how long one can wait following legal death but prior to starting the preservation procedure, which is another unknown question.
I want to be clear that I am not endorsing any of these organizations. I am just listing these prices for the purposes of discussing how much brain preservation currently costs because this is clearly relevant to its potential as an effective altruism cause area. Hopefully, it is apparent that there is a range of possible approaches and that room temperature methods in particular have the potential to be quite cheap.
- Which structural aspects of the brain are necessary for preserving the information contained in long-term memories and personality?
- Which brain preservation method is the most cost-effective way to preserve a sufficient amount of brain structure for long-term memories and other valued aspects of personal identity?
- Can we improve the expected structural preservation quality of room temperature options using methods that are still cheap?
- How long can one wait after legal death but prior to starting the brain preservation procedure, without causing information-theoretic death?
Ethical aspects of brain preservation as an EA cause area
I'm going to discuss four areas here: importance, tractability, neglectedness, and scalability.
Importance: To the extent that it might work, the ethical importance of brain preservation is very high, from both a QALY improvement and QALY extension perspective.
1. QALY improvement: The badness of involuntary death – for oneself, one’s loved ones, and others (including non-human animals) – is a huge psychological burden on most people. Involuntary death is widely considered one of the great harms of life, even by anti-natalists. An influential psychological hypothesis, Terror Management Theory, shows that people find the idea of death so aversive as to ward off even thinking about related topics. Even if brain preservation was credibly seen as giving people an optional and realistic chance at living longer, this could be a valuable improvement to people’s quality of life, as Tim Urban has discussed.
2. QALY extension: If brain preservation is successful and the person is revived, then they are likely to be revived in a future in which aging and many other forms of death are no longer a problem, either through improved biomedicine or via whole brain emulation. This has the potential to create an enormous number of QALYs.
There is a question of how long humans can live in a way that achieving a QALY would still be possible. Some feel that living a high-quality life for thousands of years is not possible. However, I’m not sure where people are getting this from, especially given the vast and obvious quality of life improvements that one could imagine in the future. For the purposes of these calculations, I will assume 10,000 QALYs is possible, given the potential for playing games, exploring the universe, understanding the secrets of the universe, learning about the past and future, connecting with other people, etc. One can imagine 10,000 QALYs as being like living 100 different lives, each 100 years long. Given the vast diversity in lives imaginable, if anything, 100 seems to me like an underestimate. If one assumes that more QALYs than 10,000 is possible, then subsequent calculations become stronger (and vice versa).
As of 2016, one estimate is that via bed nets for malaria prevention, $3500 buys 35 QALYs, or $100 per QALY. On the other hand, if a brain preservation procedure costs ~$10,000 and yields 10,000 QALYs, it has the potential to have a cost of $1 per QALY. Naturally, a key question is what is the probability of successful revival from brain preservation. If the probability of success is 1%, then brain preservation would be at $100 per probabilistic QALY, on par with bed nets for malaria prevention. Given our uncertainty in probability estimates, then more research to determine the probability of the success of brain preservation also seems very valuable.
One might argue that it is unfair to distribute all of the QALYs to one person and that they should be shared more broadly. However, there is no reason that brain preservation could not be made available to everyone who wants it.
Others might argue that these QALY arguments are misguided because future society will be able to create more people at will. For example, as Nuno Sempere describes:
> One could also argue "that cryonics doesn't create many additional QALYs because by revival time we've probably hit Malthusian limits. So any revived cryonics patients would be traded off against other future lives."
As I discuss above in QALY improvement, I disagree with this viewpoint that humans are fungible and one can just create more people to achieve the same QALYs, because of the psychological and relational harms caused by involuntary death. But, if someone views humans as replaceable, then I can see why they might think that brain preservation doesn’t really create many QALYs over alternatives possible in the future.
3. Brain preservation may also have effects on people’s propensity for long-term thinking. Having more “skin in the game” for humanity’s future may make people’s behavior more pro-social. This is much discussed and seems plausible. But it also could have complicated second-order effects and seems hard to precisely estimate, so I don’t consider it an EA reason to promote brain preservation.
4. Brain preservation doesn’t seem to contribute to S-risks. For example, if an S-risk were to occur via a malevolent AI that wanted to torture humans, then that AI could simply simulate human minds sampled from mindspace, regardless of whether those people had chosen to pursue brain preservation.
Tractability: Uncertain. Structural brain preservation is already possible and already happening. On the other hand, revival is very much not happening. In my opinion, it is very difficult to know what the probability of success of brain preservation is. There are ways to bound it based on societal factors (for example, the probability of social collapse), but the core problem of sufficient preservation quality is very challenging to predict and is correlated with other forms of uncertainty in the project.
I have studied this topic for years and I have no idea how people are making these probability estimates. Aggregating success probability estimates by those individuals who are willing to speculate despite our current uncertainty doesn’t seem very helpful. Revival also seems to be too far away for a working prediction market given today’s prediction market methods. Instead, all we can have today is feelings about how likely it is to work. So for tractability, opinions likely vary significantly.
Neglectedness: High. There are very few people in the world working full-time on structural brain preservation research for the purposes of preventing involuntary death; probably less than five to ten people. There are also few working full-time on preserving people; probably less than twenty-five to fifty people.
The practice is also highly stigmatized. Brain preservation is thought to run contrary to societal mores and break taboos regarding death. For example, Tucker Carlson called cryonics “obviously grotesque and ghoulish and kind of revolting”. This is a win from a neglectedness perspective because it suggests a clear reason why it is neglected that is orthogonal to its likelihood of success.
Scalability: High. One survey suggests that around 6% of people have already decided to pursue cryopreservation, although they currently lack options to do so. With 178,000 people dying per day, a naive extrapolation means that brain preservation has the potential to probabilistically dramatically extend the QALYs of 10,000 people a day if the infrastructure to perform this was expanded.
Another survey suggests that around 20% of people could imagine doing brain preservation. People would probably be more likely to actually do it if it were a more widespread practice and cheaper.
- How much might the possibility of not definitely undergoing involuntary death improve people’s quality of life today?
- How many QALYs are possible for a human to experience assuming an indefinite lifespan?
- Are lives fungible, so that if one person dies we can just create another one with no significant loss to QALYs? If not, how should this affect our QALY calculations for brain preservation?
- How will future society feel about creating more people until reaching Malthusian limits?
- Given low costs (perhaps around the same cost as a typical funeral or less), what percent of people would be interested in brain preservation?
It seems to me that:
(a) Most current technical arguments against brain preservation, to the extent that there are any at all, don't grapple with the possibility of structural inference.
(b) Most extant arguments in the effective altruism community against brain preservation as a cause area don't grapple with QALY improvement/extension, and for unclear reasons treat humans as replaceable units, neglecting relational and psychological factors. If people think humans are replaceable, I think they should justify this, and also consider whether they are being consistent about it.
(c) With today’s methods, brain preservation may already be among the best altruistic investments available from a QALY improvement/extension perspective, given reasonable estimates about the probability of success.
(d) With more research, substantially cheaper methods for structural brain preservation could potentially be developed, which could further improve the cost/benefit calculus. With more research, our uncertainty about different aspects of the brain preservation project could also be better clarified.
As a result of the above, and given its neglectedness, I think brain preservation for the prevention of involuntary death is one of the best areas for people interested in helping others to work in. I also think it is a great place for people who are interested in helping others to donate money. If you disagree, I would love to hear from you why that is. If you agree, I would love to discuss with you practical topics of how to best improve the field. Thanks for reading!
- Francesca Minerva is a leader in the ethics of cryonics. She has a book on the topic here and has a video discussing QALY extension via cryopreservation here; both are recommended if you’re interested in this topic.
- The Brain Preservation Foundation website: https://www.brainpreservation.org/
- John Smart’s essay: “Do we need a noncryogenic brain preservation prize?”
- My in-progress essays on brain preservation (feedback welcome!): https://brainpreservation.github.io/