Comparing resilience of brain preservation with digital data preservation

[PCmorphy72]

When we talk about long-term preservation, it is notable how similar the challenges are between biological preservation (brains, patients, biological specimens) and digital data preservation (scientific data, archives, individual memories). Both must survive not only technical hurdles, but also environmental and geopolitical risks. These risks cannot be understood solely in terms of earthquakes, floods, or blackouts: they are shaped by the broader trajectory of technological progress and social priorities.

After the decades from the post–World War II surge of innovation to the 1970s, driven by massive public investment and accompanied by widespread optimism and a moral drive toward scientific progress, a slower and more uneven progress has been observed, shaped by social choices and political climates. The narrative of “exponential growth” has often been more apparent on paper — through metrics like Moore’s law — than in genuine breakthroughs such as the discovery of the Higgs boson or high temperature superconductors (HTS). Too often, resources have been diverted toward military competition and geopolitical dominance rather than long-term scientific exploration without immediate utility. This historical context matters, because resilience in preservation — whether of brains or of data — depends not only on engineering but on the willingness of societies to sustain infrastructures through periods of stagnation, conflict, or shifting priorities.

Environmental risks
Brain preservation facilities could face earthquakes, flooding, fires, and infrastructure collapse from various causes. Salem (where many of SBP’s brains will likely remain) is located near the Cascadia Subduction Zone, one of the most dangerous seismic areas in the world. By contrast, Alcor deliberately moved to Arizona to minimize seismic and climatic risks, while CI in Michigan faces only moderate seismic exposure.

Digital archives like Zenodo are hosted at CERN in Switzerland, a relatively low-risk seismic region, but their data centers are not underground bunkers. Their resilience comes from redundancy and distributed backups rather than location alone.

Systemic risks
For brains: continuity of preservation protocols is critical. Today SBP primarily uses chemical preservation at –20 °C, which depends on stable refrigeration and power supply. Only a portion of brains remain cryopreserved in liquid nitrogen, where risks involve continuity of LN2 deliveries and secure containment. In both cases, systemic fragility arises from prolonged outages, supply-chain disruptions, or governance failures.

For data: cyberattacks, prolonged blackouts, or even global conflicts. Redundancy helps, but no system is immune to systemic collapse.

Pragmatic resilience and the “exponential growth” narrative
Jordan Sparks once wrote:
“If things go really badly, we just shut down all services and the company itself goes into a sort of hibernation mode. The building would just sit there with no activity. Volunteers would top off the LN2 every few weeks for the handful of patients who are cryopreserved. But I seriously doubt it will be centuries. If I survive another 50 years, then we're only looking at less than another 50 years after that before revival will likely be possible. Remember that technological progress will probably continue to be exponential.”

This statement captures both pragmatism and optimism: the idea of a “hibernation mode” as a survival strategy, and the reliance on exponential technological progress as a guarantee of eventual revival. Yet here lies a critical issue: such information should not be scattered across Reddit comments or forum threads. It would be reassuring if SBP made public the protective and mitigation criteria it has adopted, to strengthen both member confidence and the credibility of the project. Transparency is not a luxury; it is a prerequisite for trust.

The importance of addressing the unforeseen
The word “unforeseen” should also include risks that go beyond familiar environmental or systemic threats. These risks cannot be reduced to earthquakes, floods, or blackouts; they encompass broader challenges such as:

• Technological stagnation: History shows that progress can stall. After the post–World War II surge of innovation, the decades since the 1970s have often seen slower, more uneven advances. Moore’s law gave the illusion of exponential progress, but genuine breakthroughs in energy, medicine, and fundamental science have been less frequent.
• Geopolitical shifts: The last fifty years illustrate how fragile the balance can be between peace, dignity, and freedom on one side, and expansionist ideologies on the other. Indeed, human rights and the centrality of rationality in law and culture are increasingly being challenged today.
• Social inertia: Future generations may not accelerate progress; they could amplify stagnation through complacency, distraction, or misplaced priorities.
• Global pandemics: As demonstrated by COVID‑19, biological risks can disrupt economies, slow scientific progress, and strain social cohesion.
• Extreme scenarios: Nuclear attacks or similarly catastrophic events. These may seem remote, but resilience planning must include them if credibility is to be maintained.

In this sense, resilience requires planning for scenarios where exponential growth does not materialize, or where it is interrupted by geopolitical crises and social regression.

Constructive proposals to SBP

• Explicitly address unforeseen risks, including systemic outages, supply-chain disruptions, environmental events, geopolitical instability, and even extreme scenarios such as nuclear attacks.
• Clarify contingency strategies such as “hibernation mode,” but in formal documents rather than scattered forum posts.
• Acknowledge the uncertainty of technological trajectories, and explain how resilience is ensured even if exponential progress slows or halts.
• Reassure members whose brains are or will be already preserved by committing that their preservation will benefit from future scientific and technological advances. (For example, if future studies demonstrate that lowering storage temperatures below –20 °C can be combined with methods that prevent cracking or avoid instability of the aldehyde-based perfusion compound, then SBP would progressively lower storage temperatures, even if this requires purchasing more powerful refrigeration units. This would extend durability beyond the ~100 years currently anticipated for uploading, while minimizing long-term damage from aldehyde fixation.)

[jordansparks]

The Cascadia earthquake is a non-issue. Risk = Severity x Frequency. With a frequency of once every 400 years, the risk is therefore extremely low. If we had that earthquake right now, I estimate tens of thousands of deaths along the Oregon Coast from the tsunami. But here in the valley, 50 miles away, the damage would be more nuanced. The main considerations would be liquefaction amplifying the waves and also unreinforced masonry buildings. Our facility is built on a rock bluff sixty feet above the floodplain where there cannot be liquefaction. Our facility is also built from steel with massive cross bracing as well as heavily reinforced concrete. Everything is engineered to easily withstand the G forces. So no. It's simply not an issue.

The refrigeration is not required. It probably prevents slight damage when used constantly for 100 years, but that's very different than claiming that a week without electricity would cause damage. It just wouldn't. Alcor's containers are very tall. It sort of makes sense for them to be somewhere with zero earthquake risk, especially since they started in LA. But that's not an issue for our small plastic containers or for our short dewars. Even in the very worst scenarios, we can still get liquid nitrogen. There might be a delay of a couple of days and it might be more expensive, but it will be available. We could go for about a month and a half before any of the cryopreserved patients would be at risk.

This is already covered in https://www.sparksbrain.org/riskManagement.html, although that page could use a refresh.

I wouldn't normally argue this point, but it might be fun. I disagree that progress has stagnated. It's been pretty consistently exponential. We had a few golden decades in the US between about 1950 and 1970 because we were the only country that hadn't been bombed in WWII, but that was a bit of an illusion. It was localized and temporary. Yes, war could interrupt that exponential growth, but it hasn't happened yet. We will need a LOT more exponential growth to get to a high enough technological level to be able to scan all the molecules of a brain or to use any nanobots. I'm estimating another 100 years of exponential growth, but it's so hard to predict.

[PCmorphy72]

Jordan, thanks for pointing to the Risk Management page: I agree it’s useful, and as you noted it could use a refresh.

My quake example wasn’t only about “infrastructure collapse”. It was also about mechanical stress transmission. With the older ASC protocol, vitrified tissue is fragile and susceptible to micro‑cracks under vibration. Even with the current formalin protocol (less fragile), earthquakes can still cause tipping or displacement of the small plastic containers (oscillations of the liquid are indeed a much more subtle issue, but still not a “non-issue”: this should be a fundamental principle of resilience, which is the focus of my thread).

Two practical points that seem under‑addressed:

• Seismic damping vs tipping: Your reply focused on tipping risk for tall dewars (as Alcor had to consider during relocation). But short dewars without adequate seismic damping can still transmit vibrational energy through the liquid. A fragile object immersed in fluid during shaking isn’t protected like a crystal chandelier packed in polystyrene; the fluid itself carries forces to the tissue.
• Container retention: In Figure 11 of Charles Platt’s article, the storage vault arrangement of small plastic containers doesn’t appear fully secured. Although rare, such an event could displace containers or cause falls (regardless of whether the tissue is vulnerable to vibration).

For context, Alcor’s move to Arizona involved road transport of very tall dewars, where the main concern was tipping and logistics: truck vibrations were a non‑trivial, largely unmitigated factor. By analogy, a seismic scenario shares the same class of subtle vibrational risks, not just the obvious tipping hazard.

Earthquakes were only one example in my post. I also mentioned other scenarios, such as the possibility that uploading may require storage well beyond the ~100-year estimate, which I also agree with.

As a further example, I noticed that the Risk Management page already covers vandalism quite well, but sometimes risks come not from malice, but from simple negligence (an employee forgetting to perform a check, or mishandling routine maintenance). This “human factor” is equally important, because it can cause damage just as surely as an earthquake or vandalism. I learned this the hard way in 2023, when my main PC, through a small oversight during a move, ended up in the wrong place: buried in a box in the cellar, instead of safely upstairs where I thought it was, lost among a jumble of boxes. Shortly afterwards I left for a year, 1,300 km away, unable to intervene. During that same year, a flood struck: an event whose magnitude was comparable, yet far superior, to Italy’s flood of 1966, with parallels found only in medieval chronicles. The overlap of these circumstances — the misplacement, my absence, and the extraordinary flood — shows how resilience must account for chains of unlikely events. Of course, my personal negligence was far greater than anything one would expect from professionally qualified SBP staff, but the principle remains: even highly improbable negligence can have serious consequences when combined with external hazards. That is the essence of resilience: “very unlikely” does not mean “non‑issue”, a principle otherwise known as Murphy's law.

So, if you don’t consider adding the “hibernation mode” idea I suggested in this thread, perhaps at least the Risk Management page could clarify the flooding section: instead of just “Flooding: Essentially impossible in our facility” it could add the explanation you gave here in the forum: “[because] Our facility is built on a rock bluff sixty feet above the floodplain”. That kind of detail strengthens confidence and shows that management goes beyond listing hazards.

My aim wasn’t to push a single hazard, but to invite adding further transparency, e.g. by public protocols showing how SBP manages a broader range of risks in practice.

I’m not criticizing your work; I’m advocating for clarity. If future updates addressed seismic damping for dewars, container retention under shaking, and longer storage horizons, that would strengthen confidence that “management” truly covers both obvious and subtle risks (e.g., in the design of the new “huge freezer”).

[jordansparks]

Let's start with the brains that are in liquid nitrogen. Those are the ones I've been thinking about for the longest. They will be fine. They are all individually wrapped in padding. But I have not spent as many years thinking about storage in fixative, and it seems that I overlooked some details. I was sort of imagining that the liquid would pad them just like it does inside our skulls, but that's not quite right. They will slosh. The solution is to add foam padding in the liquid, which we will be doing shortly. We need to research which kinds of foam will be completely inert in formalin. If there was an earthquake tomorrow, it would not be catastrophic, but the brains might get "bruised".

As for securing the containers, we could also do better there. We will work on padding between the containers. The patients are all in a smaller refrigerator where they cannot fall off the shelves. The ones in the big walk-in are all research cases, so those are not important. The way they are currently sitting, some of the research cases would indeed fall off the shelves, although the containers would not crack. We've been meaning to get around to that, but it's not very high priority.

For the most part, fixed tissue is resistant to many physical risks including those that might result from employee mismanagement. Flooding is literally impossible when on top of a hill. This location was very intentionally chosen to avoid flooding. I'll keep working on that page to further clarify.

[jordansparks]

After thinking about it for a few more hours, I'm less concerned about the brains stored in fixative. The density of the brain would be very similar to the density of the liquid, so it would be well protected. But we will still look into improvements of course.

[PCmorphy72]

I really appreciate that you took the time and care to protect the vitrified brains with padding: as far as I know, neither Alcor nor CI have ever documented such measures. What I’d like to ask, though, is whether your concern was focused mainly on shocks, impacts, and sudden jolts, or if you have also considered the issue of vibrations.

Vibrations are not just another form of “impact”, but are technically different from the other aspects of an earthquake: while shocks and jolts are short, high-energy single impulses, vibrations are prolonged oscillations at low or medium frequency, which can transmit energy in a very different way, also building up resonance in the dewars. This is why shock-absorbing padding or foams are not necessarily effective against vibrations and often are not the same as damping or anti-vibration materials, which can require damping systems or anti-vibration mounts specifically tuned to reduce oscillatory energy.

I think this may be relevant from the perspective of the fragile vitrified brains, which are susceptible to micro-cracking through such vibrations (and, under significant thermal gradients or high-energy shocks, potentially even to macro-cracking). If you think it is relevant, one might also imagine scenarios such as a nearby explosion (an accidental one, a planned demolition, or even a terrorist act), which actually would involve both types of mechanical stress: the initial shock wave behaves like a sudden impact, while the subsequent pressure waves can induce vibrations in the dewars.

I make no secret of a certain inclination toward cryopreservation. I would actually like SBP to maintain, in the long term, the relevant know-how, technology, and equipment needed to implement an updated version of the ASC protocol, subject to a specific hypothesis: “if future studies demonstrate that ASC can be combined with methods that prevent cracking or avoid further damage”.

Note that in this hypothesis I’ve simply recycled the wording of the final proposal opening this thread, replacing “lowering storage temperatures below –20 °C” with “ASC” and “instability of the aldehyde-based perfusion compound” with “further damage”.

I will now go slightly off‑topic by talking about adaptability (though it might be considered complementary to resilience).

The validity of cryopreservation has shifted over the years, as your own words illustrate:

• 2019
  Q: “… why even bother with all the expense of cryopreservation? Why not just chemically fix the brain?”
  R: “No, high quality chemical fixation alone is not good enough. Damage over time is very significant due to molecular motion. There is no sharp deadline on how fast fixation must be followed by cryopreservation, but it's in the range of hours, not years.”
• 2023
  In favor of simple preservation in formalin you wrote: “The arguments that I've heard against long term fixation are that damage would happen by just sitting there over the course of decades. But I have not seen evidence for this yet.” (Although you also noted: “… lipids. These are actually the molecules that I worry about because they are not locked in place by the fixative crosslinks. Instead, they are trapped in a web of proteins, but they very well might still migrate. The brain has a lot of lipids in it.”)
  By contrast, when concluding about cryopreservation, you wrote: “Cooling adds nothing except possibly better preservation over very long periods of time. But I'm increasingly skeptical that it even does that at all. […] The downside is that each case would need to be ramped through cryoprotectant, which itself could be damaging. It's known that osmotic pressure can cause damage, it might be complex to come up with a protocol that could be used safely and reliably. Complexity has a certain risk associated with it. Because we can always transition a patient from fridge to freezer, and from freezer to LN2, …”

On this last point, my question is: would such a transition ever provide a “sharp” benefit even “in the range of years”?

I suppose I will have to wait decades for an update of your pages on that…

Anyway, you must have a good pragmatic adaptability to change the validity of cryopreservation while remaining grounded in scientific evidence.

This validity also appeared in the topic “glutaraldehyde”, which seems to have disappeared for almost three years now — a circumstance I do not particularly mind, since formaldehyde preserves molecular information in a far more inferable form; but unfortunately, this topic seems to be leading in the opposite direction from the path favoring cryopreservation:

• 2022: “But if we instead use glutaraldehyde, then it's plausible that the preservation over time could be every bit as good as with liquid nitrogen. This is speculation. We will need to show evidence for this, but I think it's very much worth exploring. Liquid nitrogen does still have some advantages: It's guaranteed to lock all molecules in place, […] So cryopreservation is always going to remain the gold standard in cases where perfusion and immediate cooling is possible.”
• 2023: “The Brain Preservation group advocates for very aggressive fixation with glutaraldehyde. I would tend to agree that we probably want something stronger than formalin for our brains, but I'm unclear if the glutaraldehyde would distort the tissue. I'm guessing not since glutaraldehyde is always used when taking electron micrographs. We're looking into those issues as well.”

Now, as an analogy “by adaptability”, I might say “I fear I will have to expect months for an update of your pages on that…”

By the way, just out of curiosity (I've gone off-topic for too long…), how your facility is geographically protected against fires? (I was a bit worried by the news from just a few months ago, as well as those of 2020)

[jordansparks]

Alcor and CI both use padding for storage in liquid nitrogen, just like we do. Yes, I've always been willing to shift my position when presented with new evidence. In this case, the evidence has come largely from Andy reviewing huge numbers of scientific papers and pointing out to me that my previous assumptions about liquid nitrogen being protective weren't actually backed up by any evidence as I had assumed. Would changing the temperature ever provide a sharp benefit in the range of years? There seems to be diminishing returns. In other words, refrigerator is great, freezer is marginally better, liquid nitrogen probably does nothing. As for glutaraldehyde, Andy was again instrumental in that view shift. But it's very nuanced. We're coming back around to that, and glutaraldehyde will probably be part of the long term protocol. The reason we moved away from it slightly was that it can form barrier that slows formaldehyde from penetrating quickly. So there are many situations where it probably shouldn't be part of the initial perfusion. Again, this came from a large number of published scientific papers. Finally, I've come to appreciate how difficult perfusion is after many many cases. That means that there's no such thing as good perfusion, so cryo cases have stopped making sense to me because of that. Poor perfusion means ice. Our facility is nearly impervious to fire. It uses metal cladding, no wood, no combustibles nearby, fully sprinklered, and then the concrete vault inside is inherently fireproof.

[PCmorphy72]

Ah, ok, you meant to those “mummifying” bands Alcor used to show in their videos years ago. I had always thought they were mainly meant to protect the fragile vitrified tissue during placement in the dewars, together with the fixing straps. As far as I understand, those bands are about 20 cm wide and create a padding layer of no more than 3 cm. Personally, I’ve always assumed that the “sleeping bag style” padding used by KrioRus (probably copying CI) might be more effective. For isolated heads, Alcor seems to use more targeted padding, but again the purpose is to prevent jolts during insertion in the dewars and, more generally, to stabilize the body during handling and movement. Perhaps for brains you were thinking of a different thickness or a more refined distribution of padding, since there is no natural cranial structure (even if vitrified) to protect the tissue. But if you consider cracking to be “inevitable” (I’m not so sure…), then vibrations and thermal gradients become almost irrelevant (but not for me…) — which is one of the reasons you prefer the simplicity of preservation in fixative, trusting in limited degradation over ~100 years, rather than risking the complications of cryopreservation (“in fact probably slightly lower in quality due to inevitable cracking”, as I read on https://sparksbrain.org/services.html ).

Changing topic, but not entirely: when I asked “would such a transition ever provide a sharp benefit even in the range of years?” I wasn’t so much asking whether the transition from freezer to LN2 is “diminishing returns”, but whether it is feasible after a decade or more. To recycle the question with some changes in wording and scope: would a transition from formalin to glutaraldehyde still be possible after years, without particular damage? You yourself mentioned that glutaraldehyde is not ideal for initial perfusion because of poor capillary penetration, and I had assumed a hybrid perfusion (formalin + glutaraldehyde) “in the range of hours”. So why not “in the range of years”? Glutaraldehyde has become more interesting to me lately: you convinced me of the molecular inferability of formalin fixation with the image https://sparksbrain.org/images/crosslinks400.jpg . I would be surprised if a similar image of glutaraldehyde fixation were equally convincing. In the meantime, I asked for a comparison with the paper Andy McKenzy linked in his thread.

[jordansparks]

I thought Alcor was hearing cracks in all of their cases when they were using the crackphone. I could be misremembering. With such a large block of tissue, I would not be surprised, but I also don't think it would cause very much damage. A crack is generally one of the least significant kinds of damage. Well we have already used glutaraldehyde in most cases, and I think that will continue to be true. So the scenario of conversion to glutaraldehyde in 10 years is moot, but yes, it would be possible if the initial preservation was only formaldehyde. Yes, I love glutaraldehyde as well. They used to use it to make leather, so I like to think of it as leatherizing the tissue.

[PCmorphy72]

You are referring to this 2003 news: “Despite multiple large acoustic fracturing events recorded during cooling, the brain remains a cohesive whole with no grossly apparent fracturing or freezing damage. The consequences of fracturing seem to remain microscopic as long as tissue remains at cryogenic temperature.” ( https://www.cryonicsarchive.org/library ... racturing/ )

They thought to solve the problem with intermediate temperature storage (a method you also experimented with, right?): “Fracturing is a universal problem in cryonics patients stored at −196 °C. Long term inspection of vitrified patients has revealed extensive cracking damage, motivating the development of intermediate temperature storage systems.” ( https://www.cryonicsarchive.org/library ... avoidance/ )

This is because they — and perhaps you as well — did not go in the direction (too expensive…) of methods that could theoretically eliminate thermal gradients completely in the cooling phase down to LN2. Such approaches would include ultra-controlled cooling curves (extremely slow and uniform temperature descent), intermediate conductive fluids or gases to distribute heat more evenly, thermal buffer materials (phase change gels or interfaces to absorb and redistribute stress), or even multi-zone active cooling systems that regulate temperature from both outside and inside simultaneously. In principle, these strategies could minimize or eliminate gradients without changing the final storage temperature, but they are technically complex and economically prohibitive for whole-body cryopreservation.

In any case, if you move from “Don't forget about cracking, which can include areas of pulverization. […] Incidentally, cracking might be worse if you use aldehyde prior to cryopreservation.” (May 13, 2025) to “A crack is generally one of the least significant kinds of damage” (Dec 09, 2025), this may create some confusion.

That is why I worked out a set of percentages, which I carefully calibrated throughout the whole day with the help of AI (in the end I rounded them a bit, since they are obviously very approximate). Their estimates take into account the maximum theoretical inferability that could be achieved with future technology.

Thanks to this exercise I also got closer to understanding, for example, the weight of glutaraldehyde compared to formalin, and realized that cracking, in a sense, is the “last” item on the list.

Comparison of long term memory inferability loss for different damages

Warm ischemia (without reperfusion or resuscitation interventions)

• 5 min (legal death): 5–10% (early metabolic failure; onset of mild edema; initial dendritic spine retraction; inferability largely preserved, with only subtle microdomain compromise detectable)
• 30 min: 10–20% (early metabolic failure; mild edema; initial spine retraction; cumulative burden increases; most synaptic architecture inferable, non-structural molecular states partly degraded)
• 1 hour: 20–30% (energy failure; edema; pronounced spine retraction; beginning no reflow; macro connectome preserved; fine scale synaptic parameters degraded but inferable in part; non-structural molecular information substantially compromised)
• 4 hours: 50–75% (widespread synaptic compromise; cytoskeletal collapse; microdomain disorganization; severe no-reflow; macro connectome topology preserved; fine-scale synaptic parameters inferability reduced drastically; molecular/epigenetic signals and DNA severely degraded, with only residual fragments and indirect chemical signatures inferable in part)

Cold ischemia during SBP perfusion (add percentages to prior warm ischemia)

• 5 min: 2–4% (substantial fixation coverage; metabolism contained; damage limited by early no-reflow and partial diffusion constraints, largely contained by fixation onset; macro-connectome topology preserved; synaptic compromise limited or partly stabilized depending on prior warm burden; cytoskeletal stabilization begins; microdomains largely preserved with mild stress; fine-scale parameters partly inferable; molecular information modestly compromised)
• 10 min: 3–5% (near-complete fixation onset; residual microvascular obstruction; cytoskeletal stabilization more advanced; autolysis further contained; microdomains preserved with minor degradation; fine-scale parameters degraded but inferable in part; molecular information modestly to moderately compromised, consistent with incremental exposure)
  
  Note: Just for comparison, after 4 hours of Alcor/CI protocol cooling phase, reaching −100°C, the inferability loss due only to cold ischemia would be 30–45% (metabolism essentially arrested; damage dominated by no-reflow and diffusion limits rather than autolysis; macro connectome topology preserved; partial synaptic compromise; micro-connectome parameters degraded; cytoskeletal stress evident; moderate microdomain degradation; partial DNA fragments; residual molecular patterns compromised).

Formalin-based SBP preservation at −20°C

• 10 years: 5–10% (slow hydrolysis/oxidation; moderate crosslink aging; synaptic architecture still broadly readable; membranes and PSDs intact but slightly softened; cytoskeletal stabilization; partial preservation of extra-connectome molecular/epigenetic signals)
• 100 years: 10–20% (progressive crosslink aging; gradual microdomain drift; vesicle distribution less discriminable; gross connectome stable; extra-connectome signals degraded but partly inferable)
• 200 years: 15–25% (cumulative hydrolysis/oxidation; increased brittleness and microfracture susceptibility; ultrastructure attenuated; macro-connectome topology largely preserved; macro connectome topology largely preserved; fine-scale synaptic parameters degraded but partly inferable; extra-connectome readability substantially compromised yet not null)

Glutaraldehyde preservation at −20°C:

• 10 years: 4–9% (tight crosslinking; minimal hydrolysis; superior ultrastructural fidelity; crisp membranes, spine apparatus, vesicle pools, PSDs; cytoskeletal architecture highly stable; extra-connectome states largely inaccessible but trace-level inference possible)
• 100 years: 8–18% (slow crosslink aging; fine features remain legible; modest brittleness; reconstruction dominated by structural topology; extra-connectome information effectively locked, with only residual inferibility)
• 200 years: 17–27% (gradual fine detail attrition; rare microfractures; ultrastructure still superior to formalin; macro-connectome topology preserved; fine-scale synaptic parameters locked/inaccessible due to dense crosslinking; extra-connectome inference extremely close to zero, but trace patterns or indirect fragments could be inferable)
  
  Note: The “extra-connectome” layer (synaptic plasticity, molecular states, epigenetic signals) accounts for about 8–12% of the overall inferability loss values (see the table in a complementary post).

Cryoprotectant osmotic damage

• Alcor CPA: 15–25% (shrink/swell of neurites and spines; dehydration; volume shocks from loading/unloading gradients; micro-connectome parameters distorted but partly inferable; molecular states compromised)
• low-DMSO CPA: 8–15% (reduced shrink/swell; gentler gradients; better preservation of spine and microdomain morphology; micro-connectome parameters degraded but inferable in part; molecular states minimally compromised)

Cryoprotectant toxic damage

• Alcor CPA: 4–10% (chemical effects on proteins and lipids; oxidative stress secondary to osmotic injury; synaptic architecture largely inferable; residual molecular degradation detectable but not total)
• low-DMSO CPA: 2–6% (lower chemical burden; residual non-osmotic effects remain; synaptic and cytoskeletal structures largely preserved; molecular patterns slightly compromised)
  
  Note: Recent Chinese research has introduced low-DMSO cryoprotectants that reduce osmotic shrinkage and minimize chemical toxicity while maintaining vitrification capacity. These formulations aim to balance permeability with reduced stress on synaptic microdomains, improving long-term inferability compared to traditional CPA mixes. (see JoVE, 2025).

Cracking

• Localized microcracking: 2–6% (few discrete fracture points; continuity punctuated — caused by mild thermal gradients during cooling to LN2 or minor impacts during storage, e.g. handling shocks)
• Diffuse microcracking: 10–22% (distributed fractures; global reconstruction harder — caused by poorly controlled thermal gradients during cooling to LN2, or accidental vibrations during storage, e.g. earthquakes, explosions, resonant oscillations of the dewar)
• Macrocracking: 30–55% (large fractures separating entire areas; global continuity heavily compromised; possibly even localized areas of pulverization — caused by accidental severe thermal gradients or strong impacts, even with padding, e.g. drops, collisions, major mechanical shocks)

[jordansparks]

Your percentages for warm ischemia seem way off. There is variation between patients, and the death process itself can cause significant damage, but I would tend to have ballpark numbers more like this:
3 hrs warm ischemia: 2% loss
6 hrs warm ischemia: 4% loss
18 hrs warm ischemia: 15% loss
The reason for this is that brain banks get brains all the time with hours of warm ischemia. If they had 75% percent loss of memory that would also mean 75% loss of structure. Those brains would be useless to the brain banks. Instead, they generally look just fine.

I also think your numbers for fixation degradation over time are way off. I think there is essentially no degradation of the memories. Also, I wouldn't worry at all about cryoprotectant osmotic or toxic damage. I think that damage is essentially zero. I would also put cracking as essentially zero loss in spite of my guess about pulverization. The real concern with cryo is not those things. The concern is ice. I looked into it more, and all cryo patients have ice, not just some of them. I do know that a little bit of ice might not be a big deal, but there's going to be a point where the ice is causing real damage. I cannot estimate that. I'll leave it to a cryobiologist.

I'm going to expand on cracking. If you take a cryopreserved brain and crack it cleanly into two halves with no pulverization, then that is zero damage. I'm very surprised that you would peg that at 50% loss. Its rudimentary to fit those two pieces back together again. You know exactly where everything belongs and it can easily be stitched up. I'm starting to think you are completely misunderstanding what's meant by "damage". It's damage to the information of where each molecule belongs. Damage means not knowing how to put it back together again. When you have full knowledge of the original state, that is zero damage. You should re-evaluate all of your percentages with the definition of damage as relates to "known state" rather than the more naive "physical damage" that you seem to have used. Because this same exact definition issue applies to all the smaller molecular changes as well. If you know where the molecule should go, that's zero damage. I really thought you already knew that because you were talking about inference.