Action Potentials for June
1: Paper describes a new method for cryogenic electron microscopy of mouse hippocampus tissue. They first isolate the region and section it into 100-200 μm thick slices. They then incubate these in combinations of at least 10% low molecular weight cryoprotectants (sucrose or ethylene glycol) plus high molecular weight cryoprotectants (e.g., 40 kDa dextran), for at least 15 minutes. Finally, they preserve the slices using high-pressure freezing.
Using fluorescence microscopy to target specific hippocampal layers, they then perform focused ion beam-based cryo-lift-out to extract tissue chunks in serial. They section these into 5 μm slices, and thin them to 150-300 nm. This enables them to collect electron tomography tilt series that reveal nanometer-scale structural details like synapses. It’s a complicated process but at the end they can visualize synapses across volumes without the use of chemical fixatives:
This is not the first study to show this, but their good synapse visualization is another example that synapses can clearly be preserved in brain tissue that is well vitrified. Here’s a 3d projection of one of the synapses they were able to see:
It’s also interesting to note that the process leads to ischemia for 1.5 to 4 hours prior to preservation via high pressure freezing, but they found that this duration of ischemia did not affect synaptic vesicle diameters.
They had some trouble finding a protocol that did not lead to any ice formation, requiring them to try out 18 different combinations of cryoprotectants. It may be useful to learn from their failures, even on such small samples with the use of high pressure freezing.
I guess that the ice formation could have been more easily prevented by using a higher concentration of cryoprotectants, but possibly using higher concentrations would have led to more dehydration, ischemia, osmotic damage, or all of these, which would have then made the synapse visualization less effective.
2: New paper uses the Drosophila connectome to provide the first comprehensive map of how touch-sensing neurons in a fly’s head connect to its brain circuits. They suggest that this explains the anatomical basis for how flies automatically groom specific body parts when they are touched there.
Note that these are innate behaviors, not learned memories. But Tomás Ryan et al. have previously argued that innate behaviors and learned memories likely have similar neural underpinnings, calling the structures underlying innate behaviors “ingrams” by analogy to the “engrams” that underlie learned memories.
3: New study measures hippocampal neurons using calcium imaging in mice while they learn reward locations over the course of a week. They found that place cells are reformed daily through behavioral timescale synaptic plasticity (BTSP), which is a phenomenon where calcium plateaus in dendrites strengthen synapses that fired seconds earlier.
Most place cells are unstable and disappear after 1-2 days. But some cells become increasingly likely to reactivate at the same location each day, though the mechanism driving this increased probability is unclear. This growing stable population of place cells correlates with better task performance over time.
What this suggests is that hippocampal place cell memories require reconstitution through plasticity occurring during each experience, though some unknown structural trace must persist to guide the cells and synaptic connections in which this plasticity is biased to occur.
It seems that there is a trade-off between plasticity and stability, allowing neural systems to maintain enough flexibility for new learning while preserving enough structure to reliably reconstruct previous spatial memories when they are needed.
Another note is that it seems to me it wouldn’t be possible to preserve the behavioral timescale synaptic plasticity in a static structure. It’s much too ephemeral. But the structural trace biasing these representations to form in particular ways seems like it could potentially be captured in a static state.
4: Neuromodulator-annotated electron microscopy reconstruction from larval zebrafish:
5: All-synthetic memory inception in the vinegar fly via optogenetic activation of both sensory and dopaminergic neurons:
6: Proteomic profiling of a brain that was excavated from a former cemetery in Bristol, UK. As far as I can tell from the article, this brain is from someone buried several decades to around 200 years ago. Given the circumstances, it was surprisingly well-preserved, with identifiable cerebral hemispheres and intact sulci and gyri. They took a biopsy sample of it from the anterior right frontal lobe.
They developed a protocol to measure the proteome of the brain, using urea lysis. Their data suggests that that hydrophobic proteins preferentially survive decomposition, perhaps because they precipitate and aggregate in the acidic conditions of decaying tissue. On the other hand, their extraction methods recovered mainly acidic peptides. These peptides especially came from cytoskeletal proteins, which seemed to be relatively well-preserved over the long term.
7: We can learn some details about the operations of brain banks from this article. For example, regarding the Mount Sinai Brain Bank:
Unsure whether it’d be able to pay staff after April, Mount Sinai Brain Bank sent layoff notices to its nearly two dozen employees, then revoked them when it learned of the funding extension less than 18 hours before the deadline, said program director Harry Haroutunian.
In the month before the contract expiration, the program, which stores 2,640 brains in 43 freezers set at minus 80 degrees Celsius, stopped taking [new brain donations]. Mount Sinai typically takes two brain donations a week.
One takeaway here is that brain banks require a large number of staff and funding to bank brains from new donors, which in this case is provided by the federal government. But long-term storage of the already banked brains, especially the ones in fluid preservation, is a much less resource-intensive process.
8: Profile of neuroscientist Davi Bock at the University of Vermont. Discusses connectomics and its implications. Interesting ending:
As I head out of the building, I think about going back in to ask him more about uploading our brains. In our conversation, he said that, even if we could, he wouldn’t want to, afraid that the digital copy of his brain would be low-fidelity. “It’s very likely to be some weird, altered thing that just doesn’t have much to do with me,” he said. I want to ask him if he’d like to make a copy of his brain if he knew that it would be a perfect copy—but then I think the better of it and keep walking out into the sunshine. His reasoning seems clear, and he’s made up his mind.
9: New study does exome sequencing on 500 people with obesity. Finds that “Monogenic obesity was diagnosed in 5.8% of patients, while 7.1% carried a potentially obesogenic variant” and “Surprisingly, diagnostic yield was lower in severe obesity cases. 40% of patients with monogenic obesity carried variants in genes not included in current obesity panels.” The most common mutation they detected was MC4R, followed by PHIP and SRRM2. Here are a few examples:
10: New method for clot removal does spinning to densify the clot’s fibrin network and allow red blood cells to be released.
11: In older mice, facial massage is found to help get lymphatics flowing and restore CSF outflow to youthful levels. 🧐
12: Review of potential treatments for intellectual disability. Most success in the space so far has come from early treatment of metabolic disorders. They argue that different treatments may be possible for a broader set of disorders.
13: New GWAS for sleep duration. The strongest association they found was with rs2863957 on chromosome 2, which is within the PAX8 gene, explaining ±2.4 minutes in sleep duration. Together, the variants explain ± 220.5 minutes of sleep duration across all genome-wide significant loci. The variants are most associated with genes known to function in lung, liver, immune, brain, and eye cell types, in that order. Although I kind of don’t totally trust this type of enrichment analysis, to the extent that it is true, this is an example of the importance of whole body function for what I would personally think of more as a brain phenotype.
14: Paradromics does its first testing of its brain computer interface implant in humans, just for 10 minutes in tissue that was going to be removed for epilepsy surgery regardless. They’re using a device with 420 electrodes that record from individual neurons.
15: Comparison of the contrast achieved by different heavy metal stains in electron microscopy.
16: I created a Manifold market on when there will be a full mouse connectome. Not many predictions yet! Criteria:
Requires synapse and cellular connectivity mapping across one entire brain (i.e. not combining data across multiple specimens). Does not require a particular technology, e.g. could be electron microscopy or expansion microscopy based, but does require synapse-level resolution and accurate tracing of cell processes. Some degree of potential inaccuracy or very small gaps is accepted, to the same level as the adult Drosophila connectome, e.g. as described here: https://www.nature.com/articles/s41586-024-07968-y. Resolves upon publication or preprint or public announcement of completion, does not require peer reviewed publication to be accepted.
17: Aschwin de Wolf on how funeral directors can assist with brain preservation/cryonics. I am most interested in the extent to which embalming protocols could be adapted to perform aldehyde perfusion for structural brain preservation.
18: New 5 million euro funding round by Tomorrow Biostasis.
19: Interview with Laura Deming about preservation and personal identity.
A lot of good material here. First, I totally agree with this take: “I think because cryo is so weird, it’s not worked on. There's also actually a bunch of other reasons, but there’s a lot of baggage around it that makes it an idea that has like a force field around it. But then once you’re inside it you’re like: this is so beautiful and so impactful. It’s so underworked on for what it could be.” She goes on to call it “anti-memetic,” which I also agree with.
Second, she has some interesting thoughts on personal identity (some editing for clarity and brevity, feel free to check out the video/transcript):
Interviewer: So say I reject the ego view of personal identity and I accept Derek Parfit’s bundle theory [that] each of us is just a web of experiences shifting through time. How should that affect how I think about longevity?
Deming: I'm very obsessed with this question right now… I feel like not enough people are thinking about it even though some people are thinking about it… I think the question is, what do you want to preserve? And you can actually make arguments for continuity, you can reconstruct reasons that you might care about physical continuity over time. But I think they’re very different from what you might be born with, which is kind of this feeling of to survive you need to make it to the next moment. I think that becomes hard to defend unless you just want to pick it intuitively as a thing to hold on to, which you can also make the choice to do that.
Interviewer: How close are we to the point at which these debates start influencing how capital is allocated?
Deming: Today.
Interviewer: Seriously?
Deming: 100%.
Interviewer: Can you say more about that?
Deming: I think that today they’re influencing how capital flows in a very subconscious way. Most people are born with certain beliefs around these topics that are not that well examined. And they’re guiding intuitions about what is correct and what is not correct to invest in today. This is interesting because the investments today then determine what might be most available in the future and that might determine what people have access to. …
Interviewer: Which view do you think is winning at the moment, at least in Silicon Valley?
Deming: What I see personally, although it might be very biased, there’s a lot of starting out with what we’re born to be most adapted to, which is I just want my physical self to continue for as long as possible. And just the most conservative possible perspective on preserving yourself. Like Ship of Theseus kind of a way.
I think when people really think about it often there’s like kind of a one-way door — or maybe it's a two-way door — but a door you go through where it's like oh actually like it’s pretty hard to defend that. In Buddhism this might be like analogous to doing like a no self meditation, just repeatedly asking “what is the I,” “what is the I,” “what is the I.” I'm not that knowledgeable, but I think in Buddhism when people just do this enough, eventually they’re kind of like well actually there's like — maybe they come closer to the Derek Parfit view, which is interesting that both types of philosophical traditions kind of get to a similar place. But then I think you should still be doing things to keep your body healthy and alive and so you kind of have to reconstruct notions of what’s meaningful about that.
I agree that people tend to start out with — “are born with” — views on personal identity that require maintenance of the physical continuity of one’s atoms over time. But that with further examination, they often come to views on personal identity such as branching identity, and if they do so, they tend to stick with those.
20: Another perspective on preservation and personal identity by a new Substack, “This Body is Beta.” Links to a couple of my favorite articles on the topic. The author on the goals of the newsletter:
“If you’re curious about the real science of aging, what it means to preserve identity, excited about breakthrough technologies that both have an immediate impact on health right now and advance longevity research…or simply love someone mortal, “This Body Is Beta” is for you 🖤
Last year I thought I was dying. This is my love letter to staying alive 💌”
21: Article by Aurelia Song and Charlie Dever in Asterisk Magazine about brain preservation. I agree with many of the points made and I disagree with a few others. Disagreements are more interesting, so I will focus on those.
One of my disagreements is regarding the 12 minute window they posit is required for high-quality ultrastructural preservation, due in turn to a requirement for high-quality perfusion. What is nice is that this is an empirical question, so the disagreement can theoretically be settled via observation (either with new data or references to existing data in the literature).
For now what I will offer is a couple of questions. First, how do we know that this window is the same in humans as it is in mice and rats? It seems to me that it could be different, especially if the cause of death is different (euthanasia vs respiratory depression).
Second, how do we know that there is no biological variability across brains in the extent of perfusion possible after a given period of ischemia? For example, in their study on the no-reflow phenomenon in cat brains, Ames et al 1968 found that after the same 10 minutes of ischemia, 0% of the brain of rabbit #33 was white and non-perfused, while 23% of the brain of rabbit #48 was. My guess is that, rather than a threshold, there is more likely a spectrum in how perfusable a brain is after a given period of ischemia.
I’m also curious about this:
I immediately notified Alcor myself, even though he had been dead for two hours and counting. Something is better than nothing, right? In the end, it took multiple days for Alcor to recover him, and he was straight-frozen.
Around that time, as part of my research, I happened to be working with a human brain preserved via immersion fixation after a similar time delay. I wanted to see for myself how well the synapses were preserved, so I took samples from the outermost, best preserved tissue in that brain, which was likely to be an upper-bound for the quality of my mentor's preserved brain.
Like Darwin's early work in cats, the preliminary results were encouraging: Using an ordinary light microscope, I could see all of the cells and nerve bundles, and there was no obvious physical damage. But the important information in the brain is really in the microstructures between cells — the synapses. And these tiny details are too small to see under a light microscope.
So I prepared a sample for an electron micrograph. What I saw devastated me. The brain was shredded. Not only were the synapses ruined, the entire structure was riddled with holes. That was what my professor's brain must look like.
When we did electron microscopy on the cortex of immersion fixed human brains after around 1 day after death, we did not identify any “holes.” Here’s a link to the raw data. And I believe that our findings are consistent with the literature, as discussed in the paper. However, we might be missing something.
It’s possible they are referring to what we have called ambiguous interstitial zones, which are probably due to a local expansion of extracellular space. These are quite concerning, but not necessarily indicative of complete information loss — in my view, it’s still unknown. I’m interested to hear more details.
Most importantly, I appreciate that Aurelia is pushing for higher standards in this field, which I agree is urgently needed.
22: New interview with musician Sufjan Stevens. I agree with him that there’s a greater power in survival than in strength. And also that sometimes survival requires being open to transformation.
You know, strength is funny because it suggests a kind of power and authority and vigor. But I also think there's greater power in survival. And sometimes survival requires sensitivity and openness, and even subservience. I think I've just become a lot more subordinate to the chaos of the world around me and less inclined to fight it, because I'm starting to learn that you cannot create change by force: You just have to move through it, and be open to transformation.