flashing fish brains
March 19, 2013 3:41 AM   Subscribe

When cutting edge microscopy meets cutting edge genetics: flashing fish brains

Microscopist Phillipp Keller and neurobiologist Misha Ahrens use light-sheet microscopy (wikipedia, Zeiss product trailer) to record neural activity in the brain of a genetically modified zebrafish. The technique, which scanned the entire brain, captured the activity almost every second, detecting 80% of its 100,000 neurons.

Original article for those with access privileges: Whole-brain functional imaging at cellular resolution using light-sheet microscopy

Phillipp Keller made the cover of Science a few years back using the same technology (article).
posted by kisch mokusch (23 comments total) 7 users marked this as a favorite
posted by j03 at 4:37 AM on March 19, 2013

At last, a brain that actually lights up, the way every piece on neuro-anything has always shown them.
posted by Segundus at 5:25 AM on March 19, 2013

Yay zebrafish!!! So, this is what I do. Well, almost.

I use 2-photon confocal microscopy to watch "flashing fish brains" in my amazing little transgenic larvae. Light sheet (which we have been looking in to for years - we actually have someone willing to build one with me!) is much faster than confocal, meaning that it's possible to document neural activity on a very fine scale using that technique. In the paper, they were capturing only once every 1.3 seconds, which is actually slower than my capture rate of once every 0.7 seconds.... but they were capturing WHOLE BRAIN activity!

I'm not watching the whole brain. I'm watching the visual processing center and the telencephalon, which is the larval zebrafish's (rather small) "higher brain" region. And of course, I'm watching only ONE PLANE instead of all of them.

Still, I get to see some pretty amazing stuff every day. Because I'm using an infrared laser for my 2-photon microscopy, my subjects (the larval fish) can't see the laser that I use to scan the brain. That frees me up to present visual stimuli to them and watch, in real-time, how their brains respond - because they aren't being distracted by constant illumination by a scanning laser.

Light-sheet and the labs that use it made a really big splash at the 2012 zebrafish conference this year (June, Madison, WI) and I think a lot of people were inspired. There will be more of this great work to come, I'm sure, especially now that Zeiss is selling a light sheet scope commercially.

Anyway, I'm stuck at home thanks to Boston's latest snow storm, waiting for the roads to be clear enough to get to work, so I can get on the microscope at my scheduled time!!
posted by Cygnet at 5:27 AM on March 19, 2013 [7 favorites]

That's very cool, though I can't help but be reminded of this study using fMRI on salmon (here's an explanation for the layperson).
posted by exogenous at 5:54 AM on March 19, 2013

Please feel free to memail me with an email address I can send a PDF of the original article and a promise not to distribute it further if you would like to see it - for the purposes of this academic discussion that we are currently having of course.
posted by Blasdelb at 6:14 AM on March 19, 2013 [1 favorite]

Now that we have mapped them, the next step will be driving them around like tiny cars! BwaHaHaHaHa! Then, once we have mastered zebra fish, on to zebras!
posted by GenjiandProust at 6:57 AM on March 19, 2013 [1 favorite]

HHMI Janelia Farm Research Campus, where this was done, had an internal conference recently in which every lab head spoke. By far the most jaw dropping two talks were the ones on this work. Everyone here does great work, but at some point one gets accustomed to the wonder and precision of the various techniques we use to understand the brains of fruit flies and mice. Imaging the distinct activity of nearly every cell in a whole brain was just insane. Needless to say, the maggot people (rather, the people trying to understand the neural basis of behavior in the fruit fly larva, of which I am one) are looking forward to having Keller do something similar on our organism.
posted by Schismatic at 7:20 AM on March 19, 2013

Does the original article say if the big flashes of activity (like around 0:19 in the video) are associated with anything? Introduction of yummy fish flakes?
posted by Kabanos at 7:30 AM on March 19, 2013

By far the most jaw dropping two talks were the ones on this work.

Zebrafish larvae agree.
posted by Kabanos at 7:32 AM on March 19, 2013 [1 favorite]

Kabanos, these experiments are performed on embryonic/larval stage fish which are embedded in a thin agarose jelly. They can't move, but even if they could, they wouldn't: larvae aren't free-swimming until 4 days post fertilization, and their mouths don't really work until they are 5 days, and they don't have to eat until 6 days (because they are nourished by their yolk).

I wish I could tell you what the big flashes are! I see them in my own fish periodically and I still don't know what it means!
posted by Cygnet at 8:02 AM on March 19, 2013 [1 favorite]

I read about this research a couple weeks ago, but when I clicked on the FPP I was secretly hoping it was instead about flashing transparent fish brains with new firmware.

Like a port of DOOM, for example.
posted by CynicalKnight at 9:53 AM on March 19, 2013 [2 favorites]

heh, Zebra Danios, I had those in my aquarium when I was a kid.
posted by smoothvirus at 9:53 AM on March 19, 2013

Why are they trying to install new firmware in zebrafish?
posted by oonh at 10:16 AM on March 19, 2013

Can anyone tell me the what and the why of the "genetically modified"? And is that the same thing as "transgenic"?
posted by benito.strauss at 2:53 PM on March 19, 2013

"Can anyone tell me the what and the why of the "genetically modified"? And is that the same thing as "transgenic"?"

In this case 'genetically modified' and 'transgenic' are pretty much synonymous, I've heard people make a distinction that transgenic should only refer to organisms modified with sequences from other critters but that is so trivially meaningless as to be ridiculous.

In this case it refers to the fact that the researchers took 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, and allowing the researchers to see that as they have a clear view through the transparent eyes. What is cool is that these researchers used fancy physics and fancier 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 those very pretty maps.
posted by Blasdelb at 3:58 PM on March 19, 2013 [2 favorites]

Thanks as usual, Blasdelb.
posted by benito.strauss at 4:10 PM on March 19, 2013

This is awesome. FYI, the paper appears to be open access, so you all should be able to read it. I may comment some more when I've read it - I do a lot of work with microscopy, but never with light sheet microscopy - and I'm very curious to see how they got it so fast.
posted by pombe at 7:50 PM on March 19, 2013

The movie in the Nature news piece is a little confusing (from my point of view) in that it makes it look like they're imaging the whole volume a lot faster than they are. It takes them a little bit more than 1 second to image the whole brain, but they get a single slice every 30 milliseconds (30 frames per second). The movie updates each slice at its acquisition time, so the whole brain is always flashing, but any given neuron is only imaged once every 1.3 seconds.

Interestingly, I think they can probably go faster - they don't mention what's rate limiting, but the camera they're using can run at 100 frames per second. It's a very clever setup. It's also worth mentioning that the amount of data they're collecting is enormous - they're recording about 300 MB/sec, and they have recordings for an hour which is about a terabyte of data for one data set.
posted by pombe at 8:01 PM on March 19, 2013

I don't think the camera is the rate-limiting factor and they say they upgraded the hardware to deal with the data load. I think the biggest issue is simply how fast they can scan and illuminate all of the cells. You can't illuminate everything at once as you would gain no spacial information. Thus, you have to scan a beam of (excitation) light across the sample while simultaneously detecting the emitted light. If you reduce the area that you're looking at, you can image much faster, but obviously then you don't get to look at the whole sample. The light-sheet approach is actually the fastest around for large organ imaging (at least, according to the rep that came around to our institute trying to sell us one). But I think you are also correct about being able to go faster. More light sheets, more cameras! Less fields of view (FOVs) per detector with a slight overlap for each FOV and then stitch the whole thing together later. No idea how feasible it is in practice and if possible it would almost certainly require some very ingenious engineering. But when you read the methods section it's apparent that the technology is constantly being pushed forward.

Definitely a "watch this space" technology.
posted by kisch mokusch at 12:00 AM on March 20, 2013

allowing the researchers to see that as they have a clear view through the transparent eyes

Allow me to be a bit of pedant here. Actually, the eyes are the least transparent part of the zebrafish embryo! They are transparent for the first 36 hours or so, but after that, the eyes are very strongly pigmented. (The pigment is opaque to the entire visible spectrum and IR, too.) This means that the most difficult neurons to observe in live zebrafish embryos are those inside the eye (in the retina) and those right between the eyes (because those are only visible at a few angles).

This is really frustrating for zebrafish researchers! It is possible to aim a microscope directly through the lens of the eye and see neurons in the retina (helpfully magnified), but generally speaking it's a roadblock.

Fortunately, the REST of the fish is entirely transparent. Until 5-ish days, they are utterly 100% jelly-like and clear and you can see each of their amazing cells, including individual red-colored blood cells pushing through tiny capillaries... it's amazing and I never get tired of it. However, they do start to develop melanophores, which are their zebra-stripe precursors, on their heads and back. These large black cells make it hard to see what's underneath them.

There are mutant zebrafish strains, most notably caspar, which do not have melanophores. (They are "albino".) They remain totally transparent as embryos and they are unpigmented as adults, although they do get large enough that you can't see through them any more (but you can see all their organs).
posted by Cygnet at 10:01 AM on March 20, 2013 [1 favorite]

I thought that was really strange too, being not a zebrafish person beyond having a GFP mutant pet (before they were cool even), and not being familiar with the albino mutants. The authors though do mention that they started with albino (slc45a2) mutants that they got from their friends,
"We used the albino (slc45a2) mutant, whose lack of pigmentation on the outside of the eye allowed us to scan through the eye and excite tissue also between the eyes."
posted by Blasdelb at 10:58 AM on March 20, 2013 [1 favorite]

Oooh, neat, Blasdelb. Maybe they can send me that line... I could use it!
posted by Cygnet at 12:31 PM on March 20, 2013

Their acknowledgments section also gives more names for who they got the fish from
posted by Blasdelb at 5:16 PM on March 20, 2013

« Older Predicting is hard, especially the future   |   "Urgent… Have possibility of taking photos… Send... Newer »

This thread has been archived and is closed to new comments