Join 3,439 readers in helping fund MetaFilter (Hide)


Brains have never looked so pretty
April 10, 2013 1:35 PM   Subscribe

Karl Deisseroth and his team at Stanford University [previously] have developed a completely new technique to make a brain perfectly see-through. They call it CLARITY, and the result has to be seen to be believed.

The paper, Structural and molecular interrogation of intact biological systems, was published in Nature (paywall) today.
posted by harujion (43 comments total) 43 users marked this as a favorite

 
This is incredibly cool. I'm getting the paywalled article emailed to me, so haven't been able to struggle through it yet, but the publicly accessible text proposes the possibility of "creat[ing] a library of brains that different researchers check out, study and then return" and if that is really where this leads us, this is truly a discovery worth celebrating...
posted by Chipmazing at 2:14 PM on April 10, 2013


Unbelievable and gorgeous. Not surprised it's from the Deisseroth lab - as a relatively neuro-ignorant genomics guy they are one of the few neuro labs I know by name.

The ability to label proteins is the most exciting part to me, since it should allow you to see not only macro connections between neurons but also very specific and fine-grained distinctions between cells in the brain that might otherwise look almost identical.
posted by en forme de poire at 2:19 PM on April 10, 2013 [2 favorites]


We've known how to label proteins for a long time. What's novel about this method is that you can visualize those proteins without cutting the brain into micrometer-thin slices and analyzing them individually.

There are techniques that create a horizontal plane of light that can be moved vertically through the z-dimension of the brain, this can be used on this gel even though it's so thick since light transmission is so good. This allows us to make a 3D representation with everything in its original location (AKA not sliced up).
posted by Strass at 2:23 PM on April 10, 2013


(I should note that I only read the New Scientist article on the paper, haven't had time to read the full article yet)
posted by Strass at 2:29 PM on April 10, 2013 [1 favorite]


Strass, right, sorry, I meant labeling proteins in 3D as opposed to in slices - from the article it seems like there were a couple of previous methods for making brains transparent, but those dissolved protein so you couldn't do IF.
posted by en forme de poire at 2:32 PM on April 10, 2013


Absolutely jaw dropping. I was fully convinced that the "see through brain" was just a headline friendly way of describing a new computer visualisation, but they truly do have a whole, transparent brain sitting there like a glass sculpture.

The level of detail in the resulting scans is amazing. I imagine it must be incredibly exciting too for people who actually work in the field.
posted by lucidium at 2:38 PM on April 10, 2013


I'm dubious about their ability to preserve microstructure, but this is huge for macro level connectomics and just generally some very, very nice work. If nothing else, it will cut out so much of the labor involved in painstakingly cryosectioning a brain into hundreds of slices and imaging and then trying to figure out where the hell a given connection goes.
posted by angst at 2:39 PM on April 10, 2013


If Santiago Ramon y Cajal is in heaven, they're letting him read this paper right now.
posted by maryr at 2:46 PM on April 10, 2013 [3 favorites]


angst: "I'm dubious about their ability to preserve microstructure"

They use formaldehyde to preserve protein microstructure, which is generally considered a very effective way of stabilizing proteins. Apparently it also cross-links the hydrogel monomers to biomolecules including proteins, nucleic acids and small molecules, which will greatly enhance stability.

Of course, since we can't see any aspects of microstructure before we are doing the cross-linking it's hard to empirically demonstrate that the end result recapitulates the in vivo organization and structure.
posted by Strass at 2:47 PM on April 10, 2013 [4 favorites]


I can see this having some arcane uses in science, but will consumers really go for them? They tried Crystal Pepsi, and that didn't fly. And does the process affect the taste?
posted by Philosopher Dirtbike at 2:54 PM on April 10, 2013 [1 favorite]


I'm going to be a little mean here and say it's not really "intact" if you had to kill it and put it in a plastic tub beforehand.
posted by polymodus at 2:59 PM on April 10, 2013 [3 favorites]


And does the process affect the taste?

The consumers don't care.
posted by hat_eater at 3:08 PM on April 10, 2013


My field is genomics, not imaging, but aren't imaging artifacts from fixation with formaldehyde generally understood by now (or at least understood on some level)? Presumably that background effect could be subtracted (if it isn't already) or it would have been an issue during the review process, no?
posted by Blazecock Pileon at 3:10 PM on April 10, 2013


Nice way to visualize whole pickled & plastinated brains...... Not really the same as looking at a living brain tho'.
posted by lalochezia at 3:11 PM on April 10, 2013


Interesting.
posted by brain_drain at 3:13 PM on April 10, 2013


Of course, since we can't see any aspects of microstructure before we are doing the cross-linking it's hard to empirically demonstrate that the end result recapitulates the in vivo organization and structure.

Well, histological sections, experience from cortical imaging using two-photon in vivo, etc. make a good comparison point. There are also other solvent-based clarifying techniques, which lack the hydrogel component, though they haven't been demonstrated so spectacularly.

Presumably that background effect could be subtracted (if it isn't already) or it would have been an issue during the review process, no?

There's additional concern in that the protein contents are being fixed onto an acrylamide matrix, which may alter epitope structure in some way. Probably you'd want to retest any antibody you'd be using for effective binding under those conditions. But they go so far as to visualize sections of the resultant product by TEM and find that even at that level ultrastructure is (mostly) preserved.
posted by monocyte at 3:16 PM on April 10, 2013 [3 favorites]


If anyone would like a copy of the paper itself, thanks for linking to it, please don't hesitate to memail me with an email address I can send a PDF to and a promise not to distribute it further, for the purposes of this academic discussion we are currently having of course.
posted by Blasdelb at 3:19 PM on April 10, 2013 [1 favorite]


Right at the beginning of the video there were some rendering glitches where some stuff was changing brightness as it rotated. For a moment I had the jaw-dropping impression that they had not only managed to image the brain in breathtaking 3D detail, but could do so while it was still alive and I was watching it think. When they explained the process, it became obvious that I was mistaken. But it sure was weird to feel a some degree of momentary disappointment at something that is so astoundingly cool and advanced. Like someone from 150 years ago learning about e-readers and then finding out they're only in black and white.

They say this could be done with other organs too, which sort of surprised me - is it really true that fat is the only thing keeping most internal organs from being optically transparent?
posted by aubilenon at 3:26 PM on April 10, 2013 [2 favorites]


"I can see this having some arcane uses in science, but will consumers really go for them? They tried Crystal Pepsi, and that didn't fly. And does the process affect the taste?"
Acute formaldehyde toxicity literally plasticizes you, like it did those brains, by binding your macromolecules together into ever increasingly impossible to break apart single massive molecules. This is why formaldehyde makes such a fantastic preservative. Eyes and gonads are the most sensitive to this, and so go first, which is where the trope of blind sterile mountain men who accidentally make methanol, which converts into formaldehyde in the liver, with their moonshine comes from.

If you think that is terrifying, the chronic toxicity is something else entirely. Aldehydes are inherently pretty reactive with biological molecules and bind to them pretty easily, what makes formaldehyde do what it does is that it is a double aldehyde on the same carbon, which means that when it reacts to something, it forms a reactive aldehyde wherever it lands. Thus, it will take one biological molecule and covalently bond it to another in a process known as cross-linking, and it does this best with nitrogenous compounds like proteins and DNA. The big problem is that if you end up with proteins irreversibly bound to DNA in one of your cells, that will seriously fuck up mitosis, here are some awesome helpful videos, in a very specific way. A chromosome with a formaldehyde induced lesion will likely fail to replicate and stop the replication complex in its tracks, but not stop the overall mitosis process. This means that one of the two daughter cells will get no copies of the affected chromosome while the other daughter gets somewhere between one and two copies. This can be a massive problem.

In almost all cancers, if you take one of the cancer cells and break it apart to produce a karyotype, you will find a bizarre assortment of missing and extra chromosomes. There are proto-oncogenes, or genes that have the potential to cause uncontrolled growth (cancer), in all of us if they become misregulated, say from being copied a few times. There are also tumor suppressor genes that cause programed cell death in the event of uncontrolled growth that can fail, when say they are missing because the chromosome they were on is gone. Formaldehyde is profoundly efficient at causing this specific defect in mitosis, so we are incredibly sensitive to even the smallest amounts of it, especially over time.

So yes, of course consumers will probably go for it in between smoke breaks, but no.

No they shouldn't.
posted by Blasdelb at 3:27 PM on April 10, 2013 [7 favorites]


"Right at the beginning of the video there were some rendering glitches where some stuff was changing brightness as it rotated. For a moment I had the jaw-dropping impression that they had not only managed to image the brain in breathtaking 3D detail, but could do so while it was still alive and I was watching it think. When they explained the process, it became obvious that I was mistaken. But it sure was weird to feel a some degree of momentary disappointment at something that is so astoundingly cool and advanced. Like someone from 150 years ago learning about e-readers and then finding out they're only in black and white."

There was a paper, actually featured on metafilter, recently where the authors used a mutant strain of zebra fish with transparent eyes and managed to modify their eggs by adding in "GCaMP5G", a gene that encodes for a protein that fluoresces in the presence of calcium. Because part of the mechanism by which nurons electrically 'fire' is that calcium is transported into them, intracellular GCaMP5G will fluoresce when nurons are activated, allowing the researchers to see that as they have a clear view through the transparent eyes. What is cool is that they used the same fancy physics and differently fancy cameras with their fancy fish to be able to localize where the fluorescence was coming from on a cellular scale in three dimensional space and on the scale of seconds - creating some very pretty maps as video. Their maps did not have nearly the level of astoundingly fine detail of neuron structure that these maps do, but do show the live brains over time you want.
posted by Blasdelb at 3:37 PM on April 10, 2013 [2 favorites]


But they go so far as to visualize sections of the resultant product by TEM and find that even at that level ultrastructure is (mostly) preserved.

I did an EM class in university. Mostly remember safety issues with handling osmium — these substances will fix you for good, just as fast as whatever you're staining. But I was under the impression that artifacts are generally known quantities with long-used fixation techniques, along with ways to avoid them. If one gets more or less the same answer with different imaging methods, it seems reasonable to say an image is probably close to the real thing.
posted by Blazecock Pileon at 3:40 PM on April 10, 2013 [1 favorite]


I didn't quite follow - it sounded like they added a mesh to keep structures in place and then dissolved the fat, but how do you apply a mesh to a fully 3D object that folds in on itself in multiple ways?
posted by ymgve at 4:15 PM on April 10, 2013


ymgve: "I didn't quite follow - it sounded like they added a mesh to keep structures in place and then dissolved the fat, but how do you apply a mesh to a fully 3D object that folds in on itself in multiple ways?"

Here's a schematic showing some of the interactions that allow this
posted by Strass at 4:19 PM on April 10, 2013


Also keep in mind that you're not applying a mesh. You're flooding the system with monomers, which settle into place and begin to gel into a polymer (this is the mesh).
posted by Strass at 4:24 PM on April 10, 2013 [1 favorite]


wow. I don't know how well neuroanatomists could visualize these structures before, but this seems transformational to me: when you can understand structure this deeply, it must inform your understanding of function.

When Watson & Crick elucidated DNA's structure, it led a huge leap in the understanding of function, right? Could this Clarity thing lead to the same kind of leap?
posted by superelastic at 4:30 PM on April 10, 2013


Wow. So cool. So gorgeous. Just... wow.

Q: Where do hippopotamus brain surgeons get their training?
A: At the hippocampus

bonus pun - Q: What do you call the tallest hippopotamus in a threesome? A: The hippopotenuse

posted by It's Raining Florence Henderson at 4:36 PM on April 10, 2013 [3 favorites]


I already achieved a similar effect with a pint of Everclear in college.
posted by kirkaracha at 4:39 PM on April 10, 2013


superelastic: "wow. I don't know how well neuroanatomists could visualize these structures before, but this seems transformational to me: when you can understand structure this deeply, it must inform your understanding of function."

If only :(

The connections between neurons are super important to know, they're literally the wiring of the brain. However, this doesn't actually tell us much about function. There's a huge gap between being able to trace the circuits on a motherboard and understanding what each wire does (let alone how computations are performed). Digital logic is easy... 0's and 1's 'combine' to perform discrete computations. The brain isn't digital though. While neurons can be on (firing) and off (not firing), there's an immense amount of discrete processes dictating this, on a scale far beyond any transistor.

Structure != function.

When Watson & Crick elucidated DNA's structure, it led a huge leap in the understanding of function, right? Could this Clarity thing lead to the same kind of leap?"

I want to say yes, but I suspect the answer won't be nearly as clear cut. We still have a lot to learn about what goes on inside neurons (rather than just their output) to understand their function. For example, hormones which change the responsiveness of the brain aren't well characterized, but obviously hormonal changes have huge effects on how we think.
posted by Strass at 4:39 PM on April 10, 2013


To clarify my earlier post: Strass is correct in that formaldehyde is generally effective at stabilizing proteins. However, (and I’m going off memory here, as I’m in the middle of something and don’t have time to read the paper in depth) I saw the first author’s poster presentation at SfN last year and he also recently gave a job talk at my university. iirc, the brain swells substantially after electrophoretic clearing, which they think is due to removal of structural proteins. It then reshrinks during the subsequent incubation. Are these size changes truly isotropic? Who knows.

I’m nitpicking, of course, as there aren’t other current alternatives, and I don’t see any technical comments about this from a quick paper skim, but it’s worth thinking about when talking about applications to microstructure connectomics.

(On another note, the Deisseroth lab has made their protocol publicly accessible, which other scientifically-minded Mefites may find of interest.)
posted by angst at 4:45 PM on April 10, 2013


As an entirely nonscience person, I see something like this and think, Gee, if we as a species weren't quite so interested in the never-ending accumulation of surplus wealth and the subjugation/extermination of those whom we, for whatever reason, don't like, we might actually get somewhere!
posted by the sobsister at 4:49 PM on April 10, 2013


So can we make OUR cortex completely translucent also?

Jokes aside, this is amazing. Combine this with the type of infracstructure the LHC has for storing and processing massive amounts of data and all of a sudden mapping a human brain seems positively doable within the foreseeable future.
posted by Hairy Lobster at 5:08 PM on April 10, 2013


That's very cool.

Almost as cool as the army of sleepwalking rats I will soon command.

Fuck yeah neuroscience!
posted by homunculus at 5:44 PM on April 10, 2013


If Santiago Ramon y Cajal is in heaven, they're letting him read this paper right now.

Unfortunately, due to prohibitive increases in annual subscription costs, heaven was forced to suspend its contract with Nature Publishing Group back in 2010.

The two organizations "have agreed to work together in the coming months to address our mutual short- and long-term challenges, including an exploration of potential new approaches and evolving publishing models."
posted by dephlogisticated at 6:36 PM on April 10, 2013 [3 favorites]


I guess if nobody else is going to say it, I will.

BRAAAIIINS!
Pretty, pretty brains.


I understand that there are limitations and certain drawbacks with this technique, but I still think it's cooler than shitz.

Blazdelb, you have scared me straight on formaldehyde toxicity. There's a hoof dressing used for horses with frog (soft tissue) infections that contains iodine and formaldehyde. I've always been pretty casual about getting this on my hands--not any more!
posted by BlueHorse at 6:40 PM on April 10, 2013


"So can we make OUR cortex completely translucent also?"

Dude seems pretty pale, so actually it looks like pretty much yeah; that is if cortex, congress, and our friendly local IRB, allowed us to make GFP mutants of him. If a mutant cortex clone were submerged in an especially tall and fantastically dangerous vat of formaldehyde and this treatment applied to him it sounds like you would be able to map his peripheral nervous system more or less just as precisely. After washing away all of mutant cortex's lipids there would still be bones, some pigments unaffected by formaldehyde, and things like fingernails that would still be kinda opaque, which would interfere with light transmission. However I don't think there is much they couldn't map by just changing the camera angle. The technique could also be applied to mapping other things too like veins, or the lymphatic system, or individual proteins in model systems that would be hard to slice - like a Mutant Millard.
posted by Blasdelb at 12:29 AM on April 11, 2013 [2 favorites]


You're the virus expert Blasdelb. Do we have any viruses 'ready' to be tested for gene therapy? My understanding is that while we can re-encode viruses by replacing their endogenous DNA/RNA with our own sequences, we can't a) target which cells will be infected (although I'm sure you could get around this with some clever promoters) and b) we can't target where in the genome an insertion occurs. Insertion will be random and can disrupt important genes (hello cancer!).
posted by Strass at 6:21 AM on April 11, 2013


The wikipedia article is pretty good.

There are a variety of viral vectors with different advantages and disadvantages competing for gene therapy standard. One of the aspects of Adeno-Associated Viruses, the front runners currently being used clinically in a couple of limited ways that could be an advantage or disadvantage depending on context is that they can infect all sorts of different human cell types. Expression of the payload though can be targetted with clever promotor use.

Nonspecific Integration is a big problem with many viral vectors from a clinical perspective, like lentiviruses, but not all. Adenoviruses do not integrate into their host's chromosomes and instead remain freestanding.
posted by Blasdelb at 7:12 AM on April 11, 2013 [1 favorite]


Blasdelb, have I told you lately that I love you, even though your username is impossible to remember accurately? Because I do.
posted by maryr at 7:13 AM on April 11, 2013 [1 favorite]


Blasdelb has replaced The Whelk as my (current) Metafilter crush
posted by Strass at 8:36 AM on April 11, 2013 [2 favorites]


You have excellent taste in users, Strass.
posted by maryr at 9:45 AM on April 11, 2013 [1 favorite]


Blasdelb is a fun one to love.
posted by Leucistic Cuttlefish at 10:42 AM on April 11, 2013 [1 favorite]


I just checked my recent history and am blushing furiously.
posted by Blasdelb at 12:22 PM on April 11, 2013 [2 favorites]


I want to see the transparent human brain on a slab. I know by its nature it would be hard to photograph, but the mouse brain pictured on top of the letters isn't as exciting an image.
posted by juniper at 5:06 PM on April 11, 2013


« Older ...Although as he self-styles himself "King Koopa,...  |  Sebastião Salgado, whose past ... Newer »


This thread has been archived and is closed to new comments