Action Potentials for April
1: Study uses unsupervised learning to analyze the dendritic morphologies of over 30,000 excitatory neurons in the mouse visual cortex, built from electron microscopy data. This is one of the largest studies of neuronal morphology to date.
Here are the features they calculated for each of the neuronal skeletons. Includes the basal bias, i.e., the depth in the y-axis of the center of mass of the basal dendrites relative to the soma.
They found that excitatory neurons mostly form a morphological continuum rather than having discrete cell types, with a few exceptions in layers 5 and 6. This is consistent with gene expression data, which has also found that “cell types” can usually be more accurately thought of as different cellular states along several different axes.
Here are the main axes of variation in neuronal morphology they found for each cortical layer and how those axes correlate (layer 1 is not included because it contains very few neuronal cell bodies):
In a baby step towards linking neuronal morphology and function, they also found that the basal bias of layer 4 neurons was correlated with the neurons' predicted functional properties, which were inferred using an algorithm based on matched calcium imaging data.
2: A collection and analysis of more than 4400 preserved human brains from archaeological sites over the last 12,000 years. For more than 1300 of them, the brains are the only remaining soft tissue. They identified five types of brain preservation — dehydration, freezing, saponification, tanning, and an unknown mechanism — and found differences in the geographical, climate, and historical distribution of each. They hypothesize that molecular crosslinking and metal complexation may be the unknown mechanism allowing for very long-term natural preservation of brain tissue for thousands of years.
3: Measuring fiber connectivity, typically studied at the neuroimaging level, using computational approaches and light microscopy data. One of the sections they studied was embedded in celloidin and myelin stained 120 years ago (D-E):
4: Volume electron microscopy of brain tissue immersion fixed at autopsy with relatively short postmortem intervals of less than 4 hours, to study synapses and cell morphology:
5: Mapping mouse brain-wide axonal projections with light sheet microscopy. A good reminder that axons can be extremely long:
6: Inferring synaptic connectivity — the “approximated connectome” — based on neuronal skeletons from light microscopy data:
7: Study (and summaries) on the role of a DNA-associated innate immunity molecular pathway in memory formation. While this finding is interesting, it’s not as directly relevant to brain preservation as one might think when they hear something like “this study found that DNA is responsible for memory”.
The key point is that the experiments in the paper primarily focus on the role of TLR9 in memory formation rather than memory recall. This is because the TLR9 genetic knockout was induced before the contextual fear conditioning (CFC) training. The TLR9 knockout mice showed impaired freezing behavior during these memory tests compared to controls.
This experimental design doesn't directly address the role of TLR9 in memory recall. To make conclusions about memory recall, they would need to perform the knockout after the memory has been formed, and then test the animals' ability to recall it. To be clear, the authors don’t conflate these things, but some of the commentary I have seen online does.
In brain preservation, what people typically care most about is the ability to recall already formed memories. Notably, "memories" in this context refers to more than just specific recollections of life events or memorized facts like names. Instead, it refers to the neural encoding of both our life experiences and innate predispositions that form during neural development, mostly in utero. The neural machinery for creating new memories is distinct from the neural machinery for recalling stored memories.
8: Study finds that dendritic spine head volumes in the hippocampus each encode 4.1-4.6 bits of information, with a nearly uniform distribution of volumes across the 24 distinguishable sizes. Note that this does not include other aspects of the synapse that could affect its properties.
9: An example of how an extreme type of genetic variability can affect neuronal morphology. Loss of Nab2, a Drosophila ortholog of a gene linked to intellectual disability in humans, leads to excessive dendritic arborization:
10: Lifelong persistence of certain nuclear RNA molecules in the mouse brain.
11: New study performs high-density recordings in the hippocampus of food-caching chickadees. They find that each caching event was represented by a unique, sparse pattern of neural activity they called a "barcode". These barcodes are transiently reactivated during retrieval of the specific associated cache and could be a mechanism for encoding episodic memories.
12: Study finds that global astrocyte calcium signaling occurs on a timescale of seconds following neuronal and behavioral events, compared to the millisecond-level communication between neurons. Signals propagate from astrocyte processes to the cell body over several seconds, in what seems to be a complementary form of slow temporal integration.
13: “Scale is all you need” theory of human brain evolution argues that humans became more intelligent than other primates primarily through genetic selection for an extended period of cortical neurogenesis that leads to an increased neural density in the cortex compared to other primates. The idea is that this quantitative expansion led to improved general capabilities for learning, memory, attention, and information sharing, rather than domain-specific adaptations.
14: Study argues that the invention of agriculture around 10,000 years ago occurred because seasons become more irregular due to worldwide changes in Earth's climate. The need for better food storage to smooth consumption during extended periods of scarcity forced hunter-gatherers to adopt sedentary and then agricultural lifestyles. This shift was strongly correlated with the degree of seasonal instability in different regions.
15: Medieval parenting advice was to hang your baby on a hook so they would not be eaten by animals. Not exactly Baby Mozart?
16: Maybe hallucinations should not be just a part of the “perception” part of the mental status exam: “We encourage a revision of the definition of auditory verbal hallucinations to move away from the necessity for auditory perception, and towards beliefs in perception due to the loss of first-person authority.”
17: Retrospective study in a Swedish cohort finds that among people with a diagnosis of ADHD, initiating ADHD medication within 2 years of diagnosis is associated with a lower rate of all-cause mortality (hazard ratio 0.79; 95% CI 0.70 to 0.88). This is driven by a lower risk of unnatural-cause mortality (2-year mortality risk, 25.9 per 10,000 individuals vs 33.3 per 10,000 individuals). Confounding will obviously be present in a study like this, yes. But it fits with what we know about e.g. the effects of ADHD medication in improving safe driving for people with ADHD.
18: Review of people with Highly Superior Autobiographical Memory. These rare individuals can access memories from nearly every day of their life. During memory retrieval, neuroimaging shows activation in brain regions typically associated with autobiographical memory, particularly in posterior visual areas like the precuneus.
19: A claim that the unique mechanism of action of ketamine is desensitizing kappa opioid receptors. Not being an expert in the area, this seems to me to explain the data better, including the effect of naloxone in blocking its antidepressant effects, the efficacy of other opioids as rapidly acting antidepressants, and the failures in attempts to use other NMDA receptor antagonists to treat depression.
20: Arguments that suboxone is no better than buprenorphine at actually treating substance use disorder, while increasing cost and decreasing accessibility.
21: Did you know: Brain tanning was widely used by hunter gatherer tribes on all continents. It involved removing the brain from the skull of animals and using it to tan hides.
22: Eon Systems — a new whole brain emulation company.
23: New paper on dielectric warming by Brian Wowk et al shows that they can rewarm rabbit kidneys with little injury from ice formation. Unlike nanowarming, does not require perfusion with nanoparticles. These kidneys are about 13 grams, as opposed to 1300 grams for an adult human brain. This is what their dielectric warming system looks like:
24: Podcast with Ken Hayworth, the president of Brain Preservation Foundation. Several points I disagree with, but he has a very interesting perspective on the problem of brain preservation. Some key quotes:
"We are at least 100 years away from routine uploading. Maybe 100 years away from the first uploading attempt. And if there was an uploading attempt of a human being it would be an Apollo moonshot scale project 50-75 years from now. Neuroscience has a long long way to go. That doesn't mean though that I am dismissing it, obviously I am enthusiastic. I think that that's what the neuroscience train is hurdling toward. It might be a slow train but it is hurdling toward that. Everything that has been uncovered so far points to the idea that the mind is computational and that the computational level that the mind is written on is the level of neurons and synaptic connections between neurons and ion channels and receptor proteins that are allowing the electrophysiology of those neurons to do the computations."
“My current strategy is I'm probably going to die like everybody else, but I'm going to die with this understanding that it didn't have to happen because there are plenty of people in the neuroscience and medical community that could make quality brain preservation happen essentially overnight… So this is really a sociological problem. If there's neuroscientists out there listening, if there's medical doctors out there listening, if you don't buy the idea that we're computational, that what we really need to do is preserve the computational aspects of the brain so that you can undergo future uploading, hey that's fine. But realize that that's counter to the neuroscience textbooks… if you don't buy that, that's fine. But if you do buy all of that then please realize that you are preserving animal brains in your experiments and measuring the molecular changes that synapses [have] undergone during learning and things like that. You know that brains can be preserved in animals extremely well. That same thing could be applied to human beings.”
"Progress is not steady right now. The environment is not great for brain preservation. If this generation bites the dust — again there's so many things in the world that could be solved that aren't getting solved for sociological issues — this is just one of those. But over the long arc of history I think we're going to get to mind uploading, we're going to get to a great future. And I'm confident in that as well."