Action Potentials for January
1: A “representation” refers to the way in which a type of sensory information or an internal state is encoded by the brain. A representation of stored memory information specifically is known as an engram. It turns out that representations don’t always stay the same over time. For example, neurons can join in or drop out of them. This phenomenon is known as “representational drift” and it is one of the hottest topics in neuroscience.
A new study uses longitudinal calcium imaging in the dentate gyrus of mice trained on a contextual fear conditioning task to measure the kinetics of representational drift in engrams. At different time points after training — 0 hours, 1 hour, etc, up to 24 hours — they presented the training stimulus and identified the engram cells active at that time point. They found that the activated cells were quite different at the different time points, suggesting a large amount of representational drift over the first day of memory consolidation:
To get a better sense of the long-term stability of the cells involved in engrams, I would be interested to see data on time periods beyond 24 hours. Based on previous evidence, representational drift seems to occur more during the initial consolidation of a memory, as opposed to during long-term storage of a memory over months or years.
2: A new study on long-term potentiation (LTP) — a model of memory formation — finds that LTP increases the uniformity of the distribution of dendritic spine head volumes, so that there are both more small and large spine head volumes compared to controls. Since dendritic spine volume is a correlate of synaptic strength, I wonder if there are potential parallels to weight normalization strategies during artificial neural network training.
3: Identifying cell types in electron microscopy data sets based on their perisomatic features:
4: In November, the first report of synapse identification using synchrotron X-ray ptychography of metal-stained brain tissue was published. It required the Coherent Small-Angle X-ray Scattering beamline at the Swiss Light Source:
It also requires the brain tissue to be at low temperature (~-183°C) and to be embedded in a highly radiation-resistant epoxy resin. How much more detailed and reliable can this imaging modality get?
5: A new study on synaptic connections in the visual system of Drosophila, focusing on the giant fiber neuron (GF) and its synaptic partners, the visual projection neurons (VPNs). Finds that the sequential arrival of VPN axons on the dendrites of the GF neuron help to establish their respective dendritic territories during development. The physical presence of these axons, rather than developmental neural activity, is the key to maintaining the dendritic territories.
6: How many neurons did T. rex have in their forebrains? It depends on which species is used as the reference class. A new study estimates 245-360M neurons if reptiles are used as the reference class and ~1-2B if birds are. Also points out that the idea that there is close to a perfect correlation between having more neurons and being smarter is questionable. Mostly I just think this paper has some lovely diagrams. People need to put more pictures of dinosaurs on their scatterplots.
7: Mapping calbindin-positive neurons across the cerebral cortex of a marmoset brain with immunohistochemistry. Used a convolutional neural network for segmentation. The analysis is still limited by the time required for human annotation.
8: Assigning neurotransmitter identities to C. elegans neurons using molecular stains and single cell RNA sequencing data of neurotransmitters and the enzymes known to be involved in their synthesis. The prospect of C. elegans emulation has long been limited based on insufficient functional annotation, but that is slowly changing, mostly due to the development of new technologies that can be applied across all fields of biology. For example, the decrease in DNA sequencing costs over time has contributed a lot.
9: “Goal-progress cells” in the medial frontal cortex apparently encode the distance of task with a sequence of goals, similar to how place cells create a map of physical space. Next up, procrastination neurons that light up when you start thinking about what you could get done tomorrow?
10: The nasopharyngeal lymph plexus is a major site for CSF drainage. Might help explain those NSFL videos of nasal CSF leaks.
11: Single glutamatergic axons have neural–arteriolar smooth muscle cell junctions (NsMJs) that use synaptic-like transmission to dilate nearby arterioles. Helps to explain neurovascular coupling.
12: Alternative splicing of the postsynaptic adhesion receptor Latrophilin-3 promotes synapse formation by assembling postsynaptic protein condensates. Good example of a molecular marker of a neural structure.
13: The role of microglia in regulating sleep. Microglia intracellular calcium activity decreases during sleep → wake transitions and increases during wake → sleep transitions. I can’t help but associate this finding with the role of sleep disturbance in bipolar disorder.
14: A new study of monkeys implanted with a brain computer interface in their motor cortex were tasked to move a computer cursor with their neural activity patterns, which they were obviously able to do because our technology is amazing. The researchers then provided them with visual feedback of their neural activity and asked them to perform a time-reversed version of their naturally occurring neural activity patterns. The monkeys were completely unable to do this.
The results suggest that your brain has characteristic activity patterns, which are constrained to occur as a result of the underlying network connectivity. Making changes to the connectivity is not possible in the short term. Even if you could directly look at your neural activity, you still wouldn't be able to easily change it.
15: New study using neuroimaging-guided TMS finds that non-invasive stimulation of the left DLPFC leads to a slight increase in hypnotizability. Good study design. However, the effect size is not very large. I would not assign a high probability to this finding replicating. (HT: Owen Scott Muir)
16: Steve Hyman writes a letter to Awais Aftab on the genetic complexity of psychiatric disorders. One point is that wondering whether to call a disorder “genetic” or not is basically semantics and not very useful. Another point is that our language around genetics is confused by framing around Mendelian causal logic instead of quantitative genetics. Recommended. Kudos to Awais for stimulating this interesting conversation.
17: Professor Michael Dickson writes an interesting first person account of living with schizophrenia. He talks about the difference between definite belief and certainty.
18: The Foresight Institute is hiring a remote project manager for a 3-4 month role in creating a whole brain emulation roadmap. Also they are hosting a workshop for whole brain emulation from May 21-22 in Berkeley.
19: New brain banking method uses formaldehyde fixation, cryoprotection using a sucrose gradient, and freezing in a bath of isopentane and dry ice (average temperature of -73°C). Leads to high-quality preservation with well-defined neuron morphology on Nissl staining:
20: Ben Best reports that 21st Century Medicine is working on testing the function of vitrified rabbit brains with EEG. I definitely support this research. I’m curious to see how they would assess what the recovered EEG signals mean in terms of brain preservation quality, such as the potential for preserving engrams. Although I prefer to measure structural brain preservation quality, I realize that many others prefer functional metrics. However, I often struggle to figure out exactly how these would be judged.
This month I tried to come up with a related prediction market: “When will we be able to reversibly preserve isolated primate brains at least 50 yrs with function and memories intact?” Also: “When, if ever before 2060, will a long-term suspended animation procedure on humans be developed and shown reliable?” Both of these milestones seem very far off to me.
21: Jordan Sparks on how to think about the scientific basis of brain preservation and on traditional cryonics vs aldehyde-based preservation.
22: CryoDAO raises more than 2 million dollars in Ether with the goal of incubating cryopreservation research. Can’t claim that I understand how this works, but feels like very good news if there is another significant source of cryonics research funding.
23: Kai Micah Mills’s cryonics manifesto.
24: Good point by Rebecca Ziegler:
