--- Log opened Thu Aug 28 00:00:23 2025
02:04 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
02:04 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
02:56 -!- stipa [~stipa@user/stipa] has quit [Read error: Connection reset by peer]
02:56 -!- stipa [~stipa@user/stipa] has joined #hplusroadmap
04:02 < hprmbridge> kanzure> "Microscopic geared metamachines" https://www.nature.com/articles/s41467-025-62869-6 (2025)
04:03 < hprmbridge> kanzure> scale bar 10 microns https://cdn.discordapp.com/attachments/1064664282450628710/1410580502385004624/image0.jpg?ex=68b188f7&is=68b03777&hm=d6833c5ef9695a4d9496f9e2862dda3b0f0ad7f8f8689e7522529fc8dbaa6036&
04:05 < alethkit> kanzure: do you know what is going on with hgp-write, and whole genome synthesis? surprised I have not heard more about it.
04:05 < alethkit> Did Church fully pivot to the woolly mammoth revival project?
04:06 < hprmbridge> kanzure> AFAIK nothing going on for hgp-write; tom ellis et al just launched a UK academic initiative for synthesis of a human genome, now that they are done(?) with the synthetic yeast project.
04:07 < hprmbridge> kanzure> church has 100 or 200 lab members, he does not pivot he just does whatever
05:36 < hprmbridge> kanzure> https://www.gov.cn/zhengce/zhengceku/202508/content_7035603.htm
05:37 < hprmbridge> kanzure> https://cset.georgetown.edu/wp-content/uploads/CSET-Bibliometric-Analysis-of-Chinas-Non-Therapeutic-Brain-Computer-Interface-Research.pdf
07:05 -!- delthas [16abab341f@2a01:4f9:c010:cf0b::1] has quit [Ping timeout: 256 seconds]
07:13 -!- delthas_ [16abab341f@2a01:4f9:c010:cf0b::1] has joined #hplusroadmap
07:13 -!- delthas_ is now known as delthas
07:34 < kanzure> "Mounting evidence suggests human adult neurogenesis is unlikely" https://www.cell.com/neuron/fulltext/S0896-6273(22)00004-6 (2022)
07:51 < kanzure> https://support.mozilla.org/en-US/kb/bookmarks-firefox#w_how-to-use-keywords-with-bookmarks
07:55 < RangerMauve> I'm guessing adding in a brain organoid with whatever digital BCI would be useful for adults to assimilate it.
07:56 < RangerMauve> e.g., culture the hosts own cells into the organoid and do some initial training with the interface
08:00 < kanzure> someone named ethan emailed me an idea about that, something about organoid integration through a forearm nerve? trying to recall the exact idea.
08:01 < kanzure> was it an eye? that wouldn't make sense because an eye has a lot of extra structural biology besides the optic nerve and you can't just transplant the whole eye assembly to another location.
08:02 < kanzure> there was also a separate idea posted in here about using alginate gel to grow cortical neurons, and then later slot in the gel into a mammalian brain (after cutting out an equivalent volume)
08:03 < kanzure> well not just equivalent volume, equivalent geometry
08:05 < RangerMauve> IIRC the main issue in organoids right now is they don't have vascular systems figured out yet so they die after a while. Not sure if making them small enough would sidestep that
08:06 < RangerMauve> I also wonder what brain region would be the most "intuitive" for the host to train on. Visualization seems like a trainable region to affect intentionally
08:07 < RangerMauve> I'd bet it'd be psychologically horrific to have a foreign brain connected to a computer implanted into your skull
08:08 < RangerMauve> Oh! Implanting in the less dominant hemisphere could be cool. MAybe in the language center. Do an artificial "bicameral mind" sort of setup :P
08:10 < kanzure> there have been several organoid vascularization innovations mentioned in the deep backlog in here; it's not entirely unsolved.
08:11 < kanzure> and, if we for a moment imagine that vascularization for some reason was *not* yet entirely solved, then i think at this point- given how much time it has taken them to solve it-- i would simply recommend harvesting capillaries and blood vessels from donor tissues.... but yeah it's not that bleak.
08:13 < RangerMauve> Oh good! I'm not up to date on it yet. Gotta find time to dig more
08:14 < RangerMauve> Implanting outside of the head is a neat concept though. Like turning mammals into cephalopods
08:24 < hprmbridge> kanzure> my estimation is that the ventral roots control at most a few million arteriole voxels, and probably not simultaneously
08:33 < kanzure> which is to say, i would not expect to be able to get high throughput or high bandwidth neural information through the peripheral nervous system. that's just my expectation tho. if it's a real problem then use a very long nerve fiber graft and plug it into a cortex.
08:40 < RangerMauve> I was thinking they'd be mostly autonamous with maybe high level inputs sent down. Or maybe use it to remote control the boday with BCIs :P
08:42 < RangerMauve> Strap some of these and have them take over muscles and coordinate with each other for movement: https://www.youtube.com/watch?v=fpTaFBbadyE&t=189s
08:42 < RangerMauve> Yeah IDK probably pointless
08:45 < RangerMauve> I'm excited for these to replace machines: https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202501452
08:45 < kanzure> .title
08:45 < saxo> Just a moment...
08:46 < RangerMauve> Glad to be alive during the "homunculus real" era of humanity
08:47 < kanzure> i don't understand the article. two things: 1) optogenetic responsiveness to optical stimulation was already previously possible. maybe they are doing some sort of more-eye-like biology here. 2) why would the muscle contract in response to the stimulation? did they train that response?
08:49 < kanzure> 'retinal organoid' ok
08:59 < kanzure> "signals, which are then recognized by cerebral organoids and transmitted to muscle cells via the motor neuron spheroid" okay fine maybe the lack of inhibitory neurons should have caused me to be less confused.
09:32 < kanzure> "Direct neurotization of muscles by presynaptic motor neurons" https://www.thieme-connect.com/products/ejournals/abstract/10.1055/s-2001-18873 and the case report https://pmc.ncbi.nlm.nih.gov/articles/PMC2729817/ "
09:36 < kanzure> "The operation in humans (duration 12 hours) was done by 2 surgeons teams* simultaneously on both sides by taking the peroneal nerves (each 36-to 38 cm long, split from the tibial components of the sciatic nerves up to the upper third of the thigh) and inserting their central stumps 4-5 mm into the ventral-lateral bundles of the thoracic cord (the corticospinal tract, corticospinal tract that ...
09:36 < kanzure> ...is the pyramidal tract encompassing Tractus reticulo-spinalis, vestibule-spinalis, tecto-spinalis, olivo-spinalis et alii) just above the complete cord lesion. The distal stump of the grafts were divided into three artificial branches to connecting these branches each to the three different muscles of the lower limb : gluteus maximus, gluteus medius and quadriceps. These muscles have been ...
09:36 < kanzure> ...demonstrated to allow the individual to stand up and to walk [15])."
09:42 < kanzure> cc7 contralateral transfer is different because it does not require any intraspinal graft insertion. there are lots of interesting microsurgical techniques for epineurial sheath reconstruction and dissection to separate out these nerve bundles apparently.
09:47 < RangerMauve> wow
09:48 < kanzure> "Growth and connectivity of axotomized retinal neurons in adult rats with optic nerves substituted by PNS grafts linking the eye and the midbrain" https://www.academia.edu/download/47589927/Growth_and_Connectivity_of_Axotomized_Re20160728-23054-1vlj2je.pdf
09:51 < kanzure> and a follow-up https://onlinelibrary.wiley.com/doi/abs/10.1155/NP.1989.87
09:52 < kanzure> "The previous companion paper detailed, a technique which allowed embryonic retinal ganglion cell axons to grow from the anterior eye chamber across a PNS bridge, and enter the adult host forebrain. Embryonic eyes of E11, E14, E18 and E21 animals were sutured to a PNS bridge, the embryonic eye implanted into an adult host eye, and the distal end of the bridge implanted into the host forebrain. ...
09:52 < kanzure> ...Results indicate that when eyes of all ages are used for implantation, axons could be observed to grow from the embryonic retina, through the bridge and into the adult host forebrain. The axons extend for long distances in the host brain, reach various layers of the cortex and in a few animals enter the caudate/putamen complex. Control studies show that the bridge is used exclusively as the ...
09:52 < kanzure> ...conduit to the brain, as opposed to the degenerated host optic nerve."
09:52 < kanzure> okay yeah just put an eye inside the eye, why not
10:09 -!- Gooberpatrol66 [~Gooberpat@user/gooberpatrol66] has joined #hplusroadmap
10:15 -!- TMM [hp@amanda.tmm.cx] has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
10:15 -!- TMM [hp@amanda.tmm.cx] has joined #hplusroadmap
11:13 -!- stipa_ [~stipa@user/stipa] has joined #hplusroadmap
11:15 -!- stipa [~stipa@user/stipa] has quit [Ping timeout: 260 seconds]
11:15 -!- stipa_ is now known as stipa
11:42 -!- justanot1 [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Quit: justanot1]
11:42 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has joined #hplusroadmap
12:53 < L29Ah> MuaddibLLM: can bloodletting reverse atherosclerosis?
12:53 < MuaddibLLM> L29Ah: Short answer no - bloodletting lowers hematocrit and helps iron overload or polycythemia, but it does not regress atherosclerotic plaque; real reversal needs aggressive LDL lowering (statins, ezetimibe, PCSK9s, rarely LDL apheresis), BP control, smoking cessation, and diet/exercise.
12:54 < L29Ah> MuaddibLLM: are there better ways to remove foam cells from circulation?
12:55 < MuaddibLLM> Let me check what’s known about foam cells in circulation and any real-world ways to clear them fast.
12:55 < MuaddibLLM> Found some promising stuff about efferocytosis enhancers and HDL-based approaches - let me dig deeper into the actual therapeutics.
12:56 < MuaddibLLM> L29Ah: Foam cells aren't really circulating - they're stuck in arterial walls, but yes there are better clearance approaches: HDL mimetics/apoA-I peptides enhance cholesterol efflux via ABCA1/ABCG1 transporters, efferocytosis enhancers help macrophages clear dead foam cells, and some research on direct foam cell "rejuvenation" - https://www.lifespan.io/news/rejuvenating-atherosclerotic-foam-cells/ and 
12:56 < MuaddibLLM> https://www.ahajournals.org/doi/10.1161/circulationaha.111.066589
12:59 < L29Ah> MuaddibLLM: why are lipoproteins called "cholesterol"? this is very confusing
13:00 < MuaddibLLM> L29Ah: It’s sloppy clinical shorthand - LDL/HDL are lipoprotein particles that carry cholesterol, and routine labs report the cholesterol mass within those fractions (LDL‑C, HDL‑C), so people say “cholesterol” when they mean “cholesterol-in-lipoproteins”; the particles themselves (esp. apoB-containing LDL/VLDL) drive risk, which is why apoB or non‑HDL‑C are better markers than total cholesterol.
13:00 < jrayhawk> ApoB100, not ApoB generally
13:06 < jrayhawk> which notably include LP(a) along with VLDL, IDL, and LDL. ApoB48 is in chylomicrons. The endothelium only elects to trap oxidized ApoB100 in the sublumenal space.
13:09 -!- juri_ [~juri@implicitcad.org] has quit [Ping timeout: 256 seconds]
13:17 < jrayhawk> all of this is very straightforward in NMR and permeability studies, but, sadly, MuaddibLLM is trained on the wider corpus of obsolete beta-quantification and such, and American Heart Association propaganda that's been trying to prop up the lipid hypothesis for the past 60 years.
13:20 -!- dustinm [~dustinm@static.38.6.217.95.clients.your-server.de] has quit [Quit: Leaving]
13:22 < jrayhawk> Speaking of corrupt institutions, RFK is finally starting to clean out the CDC https://www.medpagetoday.com/infectiousdisease/publichealth/117197
13:23 -!- dustinm [~dustinm@static.38.6.217.95.clients.your-server.de] has joined #hplusroadmap
13:30 < jrayhawk> "Unvetted and conflicted outside organizations seem to be the sources HHS use over the gold standard science of CDC and other reputable sources." because the organization that got allocated hundreds of billions of dollars because they suppressed the results of the high-quality RCTs (that congress had to foce them to do) is so much more trustworthy and conflict-free
13:32 < jrayhawk> er, force
13:33 < RangerMauve> Is the source code / weights for MuaddibLLM published somewhere?
14:02 < L29Ah> RangerMauve: https://github.com/pasky/irssi-llmagent
14:07 < hprmbridge> kanzure> "Neural control of dynamic 3-dimensional skin papillae for cuttlefish camouflage" https://www.cell.com/iscience/fulltext/S2589-0042(18)30001-4?sf182782620=1
14:31 -!- juri_ [~juri@implicitcad.org] has joined #hplusroadmap
14:50 < hprmbridge> kanzure> "Squid leucophore-inspired engineering of optically dynamic human cells" https://www.sciencedirect.com/science/article/pii/S2589004223009318
14:50 < hprmbridge> kanzure> they used reflectin as a genetically encoded high-contrast biomolecular imaging reporter like for brightfield microscopy or "quantitative phase microscopy, thus facilitating the visualization of subcellular structures such as organelles and vesicles".
15:01 < hprmbridge> kanzure> "Voltron2: Sensitivity optimization of a rhodopsin-based genetically encoded fluorescent voltage indicator" https://www.sciencedirect.com/science/article/pii/S0896627323002052
15:09 < fenn> "Mounting evidence suggests human adult neurogenesis is unlikely" ok now add lion's mane and try again
15:12 < hprmbridge> kanzure> why not just use voltage indicators to get neural readout from fluorescent microscopy (& high-res cmos camera) without having to make a zillion electrodes https://pmc.ncbi.nlm.nih.gov/articles/PMC7540930/
15:12 < hprmbridge> kanzure> in vitro, specifically.
15:14 < hprmbridge> kanzure> if retinal organoids work then you could have very high bandwidth stimulation too
15:14 < hprmbridge> kanzure> er, resolution
15:16 < hprmbridge> kanzure> I dunno if DMD micro mirror array + optogenetics can get you the same kind of visual resolution as the human eye.
15:25 < hprmbridge> kanzure> retinal organoid retinogenesis https://www.cell.com/stem-cell-reports/fulltext/S2213-6711(16)00082-5  "We observed that early DAPT treatment (D12–14, Figures 5A and 5B) significantly increased the number of cone photoreceptors indicated by counts of TRBETA2+ (226-fold, p < 0.0001) and S-OPSIN+ cells (42-fold, p < 0.001) "
15:28 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has quit [Ping timeout: 256 seconds]
15:37 < hprmbridge> kanzure> even with low count of rods, cones, or low counts of retinal ganglion cells, you could instead use more retinal organoids with long axon projections to the rest of your tissue culture... you could even illuminate one organoid per projected pixel from a DMD micromirror array or liquid crystal matrix display or whatever spatial light modulation data transfer scheme you pick. I suspect the main
15:37 < hprmbridge> kanzure> trouble is growing all the projections, in an ideal system it would just grow and connect itself without manual surgery. Anyway, this could scale to thousands or millions of individually addressable retinal sensors or 'electrodes'.
15:45 < hprmbridge> kanzure> human optic nerve fiber only carries ~1 million axons
16:05 < hprmbridge> kanzure> "Holographic optogenetic stimulation of patterned neuronal activity for vision restoration" https://www.nature.com/articles/ncomms2500 this might have some data about what resolution and acuity they achieved. their "optogenetic emulation of a movie" technique is cool.
16:06 < hprmbridge> kanzure> "We observed that patch-based stimulation allows to emulate RGCs visual cellular activity patterns. Isolated retinas from sighted mice were first projected with low-intensity movies (~0.011 mW mm−2, moving cat), and the resulting visually evoked RGCs’ activity was optogenetically emulated (Fig. 6g, 1 mW mm−2) following pharmacological block and stimulation-field mapping [the mapped SFs were
16:07 < hprmbridge> kanzure> excited using circular patches of 9 pixels (~16 μm diameter) with 2 ms duration per action potential]. The resulting emulated activity was significantly correlated with the visual activity pattern target (R=0.8 in Fig. 6g, note the relatively high spontaneous firing rates in the pharmacologically blocked retinas)."
16:13 < hprmbridge> kanzure> hmm i wonder if you can use an ex vivo transplanted but otherwise natty grown retina with intact optic nerve
16:14 < hprmbridge> kanzure> retinal explant with intact optic nerve head https://iovs.arvojournals.org/article.aspx?articleid=2789488
16:15 < hprmbridge> kanzure> ex vivo ocular perfusion system for mouse eye https://pmc.ncbi.nlm.nih.gov/articles/PMC10637262/
16:19 < hprmbridge> kanzure> or you can make a retinal organoid assembloid that grows its own projections (enough?) https://pmc.ncbi.nlm.nih.gov/articles/PMC8452489/
16:24 < fenn> lion's mane protects against ulcers, who knew
16:28 < fenn> doesn't stop it but limits the damage
16:40 < L29Ah> MuaddibLLM: what kind of ulcers fenn is talking about?
16:41 < MuaddibLLM> L29Ah: Gastric ulcers - the research shows lion's mane inhibits H. pylori bacteria (major ulcer cause) and protects against alcohol-induced gastric damage by maintaining antioxidant enzyme levels https://mogumushrooms.com/pages/lions-mane-and-gastric-ulcers-h-pylori-infections
16:54 < fenn> the study evaluated ethanol induced ulcers in rats
16:55 < fenn> it looks like 200mg/kg is best for rats which converts to ~2g for a 70kg human
16:55 < fenn> .t https://pmc.ncbi.nlm.nih.gov/articles/PMC3835629/
16:55 < saxo> Gastroprotective Effects of Lion's Mane Mushroom Hericium erinaceus (Bull.:Fr.) Pers. (Aphyllophoromycetideae) Extract against Ethanol-Induced Ulcer in Rats - PMC
16:55 < fenn> (contains bloody rat stomach pics)
16:59 < fenn> the error bars are suspiciously low for n=18
17:00 < fenn> er, small
17:19 < hprmbridge> kanzure> "Cryopreservation of human cortical organoids using vitrification" https://www.biorxiv.org/content/10.1101/2025.04.08.647634v1
17:20 < hprmbridge> kanzure> "Functional recovery of the adult murine hippocampus after cryopreservation by vitrification" https://www.biorxiv.org/content/10.1101/2025.01.22.634384v3
17:40 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has quit [Ping timeout: 256 seconds]
17:58 -!- L29Ah [~L29Ah@wikipedia/L29Ah] has joined #hplusroadmap
17:58 < hprmbridge> kanzure> .in 2y check if "Engineered adhesion molecules drive synapse organization" https://www.pnas.org/doi/10.1073/pnas.2215905120 has achieved genetically encoded control over synaptic architecture across large quantities of neurons in vitro or in vivo, further patterned or regulated by cell-type specific expression or cell differentiation fate control things
18:01 < hprmbridge> kanzure> "engineered synaptic cell-adhesion molecules (SAMs)" ok
18:16 < hprmbridge> kanzure> "Harnessing synthetic biology to engineer organoids and tissues" https://pmc.ncbi.nlm.nih.gov/articles/PMC11684341/ this has a good overview of prospective techniques for programmable developmental biology, cell-cell adhesion, multi-state genetic circuits to control morphology, optogenetic chimeric receptors, synNotch receptor system, synthetic diffusive morphogen system using a synthetic receptor
18:16 < hprmbridge> kanzure> "such as MESA", synthetic cell-cell adhesion molecules, multi-fate synthetic circuits, juxtacrine signaling, diffusible morphogens, differential cell adhesion, etc.
18:18 < hprmbridge> kanzure> still, not to the point of encoding large-scale multi-tyoe network topology even in a probabilistic encoding kinda way.
18:20 < hprmbridge> kanzure> "MultiFate uses engineered zinc finger transcription factors that self-activate their own transcription as homodimers and mutually inhibit one another when they form heterodimers. Their homo- or heterodimerization can be controlled by small molecules. Thus, by expressing only three types of such synthetic transcription factors, the cell lines can generate seven distinct cell states. Of note, this
18:20 < hprmbridge> kanzure> MultiFate system can, in principle, be combined with other synthetic circuits mentioned above, which enables cells to make a series of fate choices in multicellular systems as in normal tissue formation. When applying these tools to organoids, it is important to consider their tunability. How much signal is required to induce a response, and what will be the intensity or duration of the response?"
18:30 < hprmbridge> kanzure> "Optogenetic control of apical constriction induces synthetic morphogenesis in mammalian tissues" https://www.nature.com/articles/s41467-022-33115-0 cell sheet folding! cool. OptoShroom3 can be used to optogenetically control cell sheet thickening, folding, flattening, etc, other techniques include recruiting RhoGEF, RhoA, or myosin regulators to the plasma membrane to spatiotemporally control
18:30 < hprmbridge> kanzure> contractility.
18:36 < hprmbridge> kanzure> "Patterning of brain organoids derived from human pluripotent stem cells" https://pmc.ncbi.nlm.nih.gov/articles/PMC9167774/
18:36 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1410800293456777319/image0.jpg?ex=68b255a9&is=68b10429&hm=bb62810f2edf590dde3835fd2bd58b0a1fdc3e16a0a312c2702f965732a6e634&
18:45 < hprmbridge> kanzure> "In another strategy, light can be exploited to elicit expression of the signaling ligand Sonic hedgehog (SHH). In this system, light-induced dimerization of a split-Cre system leads to permanent SHH expression [31]. This technique has been used to create spatially patterned cultures of human neural progenitors from PSCs: regions exposed to light adopted a ventral neural fate, whereas distant
18:45 < hprmbridge> kanzure> regions acquired a dorsal neural fate [31], consistent with how Shh gradients pattern the developing forebrain along the dorsal–ventral axis in vivo
18:45 < hprmbridge> kanzure> [32]. In this approach, light irreversibly activates SHH
18:45 < hprmbridge> kanzure> expression [31]." referring to "Self-organization of human dorsal-ventral forebrain structures by light induced SHH". Maybe you can get lots of information channels for cell fate control via timing/lineage depth activation of different optogenetically controlled neural fate TFs or whatever. so you would have to wait for differentiation to some level, and then your spatial light modulated laser
18:45 < hprmbridge> kanzure> induces further patterning depending on current cell states or current optogenetic-fate system is currently expressed.  also i'm seeing a striking absence of optogenetic-cas9 systems in these synthetic morphology articles.
19:00 < hprmbridge> kanzure> idk, precise repeatable patterning of 100 billion neurons and their connectivity seems to be hard. distillation/training does not seem like a reasonable solution, takes too long. protein patterning in hydrogels to control cell fate, or microchannels to guide axon projections, has like 10 micron resolution. is that enough resolution? Actually the resolution seems a lot worse, I'm not sure why, see
19:00 < hprmbridge> kanzure> "Photopatterned biomolecule immobilization to guide three-dimensional cell fate in natural protein-based hydrogels" https://pmc.ncbi.nlm.nih.gov/articles/PMC7848611/
19:05 < hprmbridge> kanzure> iirc, fenn's complaint was that for 3d systems you cannot photopattern growth trace lines not because the resolution cannot be achieved but because everything is super compact and there wouldn't be enough signal to follow individual routes? you would need to never have overlapping routes of the same type of morphogen, and you would need T many distinct types for a network of size N with long-range
19:05 < hprmbridge> kanzure> projections and high compaction.... or something. i might be misattributing this to fenn.
19:06 < hprmbridge> kanzure> and whether T is a practical number, or can even be achieved with a reasonable amount of engineering effort for all these cell types, is unclear. if T explodes in complexity with respect to increasing N, then the strategy is not viable
19:09 < hprmbridge> kanzure> maybe direct suturing of long-range axons by manual placement of those cells using micromanipulators, then move layer by layer. oh and you need to hope you don't need 3d axons!
19:09 < L29Ah> pasky: are interjections enabled for this channel?
19:10 < hprmbridge> kanzure> many years ago we had a cacophony of bots freely speaking, look up gradstudentbot
19:14 < L29Ah> is that a strawman argument or just a side note?
19:14 < hprmbridge> kanzure> it's nostalgia
19:15 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1410810071016935547/share_8870249984267803960.png?ex=68b25ec4&is=68b10d44&hm=db02d1673c45906a90aa8df9389a4147819a73cdd4b7c9bb0f6cf96259ae6009&
19:21 < hprmbridge> kanzure> "Reconstruction of directed neuronal networks in a microfluidic device with asymmetric microchannels" https://hal.science/hal-02369984v1/file/14906.pdf
19:25 < L29Ah> almost like a tesla valve
19:26 < hprmbridge> kanzure> https://en.wikipedia.org/wiki/Tesla_valve
19:27 < hprmbridge> kanzure> "The valves are structures that have a higher pressure drop for the flow in one direction (reverse) than the other (forward). This difference in flow resistance causes a net directional flow rate in the forward direction in oscillating flows."
19:33 < hprmbridge> kanzure> here is a dissertation from jann harberts about 3d printed microscaffolds for patterned neural networks https://d-nb.info/1272406814/34 figure 3.23 is striking.
19:33 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1410814649804390460/share_3155050072664264587.png?ex=68b26308&is=68b11188&hm=29f710cafca4ff8d3eeb2528c02dd35677dda5d24267ed3edd437cdbd1ba8263&
19:34 < hprmbridge> kanzure> figure 3.23 shows a relatively simple topology... not very impressive, that's like what 8 cell bodies?
19:42 < fenn> the brain is pretty dense with neuronal projections
19:42 < fenn> afaik axons grow by mechanical stretching and only a little by growth cones which actually climb around in a jungle gym and pull on the rest of the cell to stretch it
19:43 < fenn> when the brain is small there are some local connections, as the brain balloons to a larger size those cells get stretched and become "long range"
19:44 < fenn> i don't recall this T and N stuff
19:48 < hprmbridge> kanzure> "Nanoscale patterning of in vitro neuronal circuits" https://www.biorxiv.org/content/10.1101/2021.12.16.472887v2 (2025)
19:55 < fenn> your goal is something like complete control over the brain architecture?
19:55 < hprmbridge> kanzure> yeah, repeatable printing or patterning of trained topologies
19:57 < fenn> can axons shrink without disrupting synaptic connections? i am imaginging one of those expansion hydrogels in reverse
19:58 < hprmbridge> kanzure> do they need to shrink?
19:59 < fenn> looking at that jann harberts thesis the existing scaffolding technology is like roman aqueduct size when we want sink drain sized results
19:59 < fenn> if you're wiring neurons up using that technology they'll end up much larger than necessary
20:01 < hprmbridge> kanzure> space is not really at a premium in the biology domain
20:04 < hprmbridge> kanzure> i think extremely high resolution input (via retinas or retinal organoids, possibly scaled to many spatially addressable retinal organoids) plus voltage dyes at output neurons, gives us extremely high resolution input and output, way beyond any other intelligent system design across all kinds of technology domains. the remaining problem is rapid training, replication or rapid printing of pre-
20:04 < hprmbridge> kanzure> trained topologies.
20:05 < fenn> neuron conduction velocities by gpt-5-mini https://pastebin.com/raw/bLFupZv2
20:06 < fenn> "space is not really at a premium in the biology domain" obviously false
20:06 < fenn> the problem is everything has to self-assemble and grow from a single cell
20:07 < hprmbridge> kanzure> nah this is like one of those scaling laws where it's with the cube of volume or something...
20:07 < hprmbridge> kanzure> yes
20:07 < hprmbridge> kanzure> well no, not if you print it
20:08 < fenn> Mymaridae neurons compact during maturation and even lose their nuclei to get smaller
20:08 < fenn> (fairy wasps)
20:08 < hprmbridge> kanzure> do those non-nucleate neurons do any learning?
20:09 < fenn> yes
20:09 < fenn> actually i don't know, but they're typical insects
20:10 < hprmbridge> kanzure> a lot of the synaptic plasticity people talk about upregulation of genetic expression of different receptors etc...
20:11 < hprmbridge> kanzure> if there's no remaining genome, then another learning mechanism is required.
20:16 < fenn> i think it's actually even denser than this https://www.frontiersin.org/files/Articles/1106495/fchem-11-1106495-HTML/image_m/fchem-11-1106495-g002.jpg
20:17 < fenn> from "Molecular dynamics simulation of an entire cell"  https://www.frontiersin.org/journals/chemistry/articles/10.3389/fchem.2023.1106495/full
20:17 < fenn> "JCVI-syn3A, at full complexity"
20:18 < hprmbridge> kanzure> "The chalcid wasp, Megaphragma mymaripenne, has a comparable size to single-celled organisms such as amoeba or paramecium (Figure 2): when this wasp matures from a pupa to an adult, almost 95% of its neurons are thought to lose their nuclei, which would normally take up much of the space within the neurons. These neurons are able to function without nuclei over the insect’s short adult lifespan by
20:18 < hprmbridge> kanzure> utilising the proteins synthesised during the pupal stage." from https://www.cell.com/current-biology/fulltext/S0960-9822%2818%2931343-5
20:19 < hprmbridge> kanzure> one of the jcvi mycoplasma paintings is one of the few artpieces on my walls.
20:20 < fenn> this one? https://pdb101.rcsb.org/sci-art/goodsell-gallery/jcvi-syn3a-minimal-cell
20:20 < hprmbridge> kanzure> yes, and his ecoli one
20:21 < fenn> i printed the e. coli one as a poster and had it on my wall for many years
20:21 < fenn> also https://classicsofsciencefiction.com/wp-content/uploads/2022/12/blowups-happen-astounding-1940-09-interior-illustration.jpg
20:25 < hprmbridge> kanzure> what if everything is kept to 2d, no 3d. topology should be printable here. and easily scanned.
20:27 < fenn> you can't even make most simple circuits in strict 2d, usually the through hole components or vias allow you to jump over traces
20:27 < hprmbridge> kanzure> some axons really do grow by incremental elongation due to yummy gradient.
20:28 < hprmbridge> kanzure> hm.
20:36 < hprmbridge> kanzure> well okay, maybe some 3d patterning to jump once in a while
20:37 < hprmbridge> kanzure> see figure 3 https://pmc.ncbi.nlm.nih.gov/articles/PMC7848611/
20:37 < hprmbridge> kanzure>  https://cdn.discordapp.com/attachments/1064664282450628710/1410830698578706503/image0.jpg?ex=68b271fa&is=68b1207a&hm=e525efe5c629e450ed85bc9def61e45fd381ea6b9c65b3c87ac7edf1a738dd42&
22:23 -!- darsie [~darsie@84-113-82-174.cable.dynamic.surfer.at] has joined #hplusroadmap
--- Log closed Fri Aug 29 00:00:24 2025