5 ms·

Neuroscientists roll out first comprehensive atlas of brain cells

102 points by conse_lad 5y ago

mrweasel 5y ago
Neuroscience is fascinating, well at least the part about image processing, but it's also completely inaccessible. From computer science and programming we are use to having introduction available in the form of articles, videos and cheap training. For Neuroscience it's just straight up hard-mode and you don't know any of the words. I still not even completely sure what a brain atlas is. I've seen a few free university courses, but they fill up almost immediately.
- mdp2021 5y ago
  YT, for example, contains many resources on brain anatomy and more. Plus, for example, the lessons from Robert Sapolsky at Stanford. Plus, there exist the primer books from Rita Carter (first of all, Mapping the Mind). There are loads of good material around. Note though, that CS is "forward" engineering, while Brain Science is backward engineering. So, if you do not find material that build from the ground up... it is easily because we are not at that level.
- Frazmatazz 5y ago
  This is decent. https://www.edx.org/xseries/harvardx-fundamentals-of-neuroscience https://www.edx.org/xseries/harvardx-fundamentals-of-neurosc... Comparable to the intro biological psychology course I took in college. It's free if you don't care about certificate.
vanderZwan 5y ago
So I expected this to be about the human brain since sites like http://mousebrain.org/ http://mousebrain.org/ already existed, but it looks like it's research that has been done on mice brains as well? Can someone explain the differences here?
- stenl 5y ago
  Hey vanderZwan, we have a paper in the collection (there’s 17 brain atlas papers in this issue of Nature) on human, marmoset and mouse motor cortex (my group did the human). I think it’s cool that pretty much all the cell types could be aligned between those species, which diverged 90 million years ago. Our full human brain cell type atlas is work in progress but hopefully we’ll post the preprint early next year. (Edit) Direct link to the papers, which are all open access: https://www.nature.com/nature/volumes/598/issues/7879 https://www.nature.com/nature/volumes/598/issues/7879
- pamplemoose 5y ago
  the headline is attention seeking since there are numerous pre-existing brain atlases that describe the anatomy/cellular/molecular parts in different ways. You can check https://portal.brain-map.org/ https://portal.brain-map.org/ for many examples. I think a good comparison to this kind of work is what has happened previously in astronomy, where scientists work to answer their specific questions using large open datasets acquired using shared infrastructure (fancy telescopes for astronomy, microscopes for biology). Here is some more ongoing work from the Allen institute where you can make a scientific proposal for their openscope (https://alleninstitute.org/what-we-do/brain-science/research/products-tools/openscope/ https://alleninstitute.org/what-we-do/brain-science/research...)
swashbuck1r 5y ago
Are there any examples of "Hello World!" with simulated basic brain cells? I'm sure this is an entirely naive question given the complexity of a brain, but I'm imagining a rudimentary program to help understand brain-style processing with some kind of brain cell struct unit that represents real-ish input/output mechanisms that can be connected to other brain cell struct units...leading to some minimal brain-style processing outcomes.
- andyxor 5y ago
  If you want to understand how the brain works this is a good intro with some realistic neuronal network models ( spoiler: these have nothing to do with “artificial neural nets” as we know them) https://www.amazon.com/Brain-Computations-Edmund-T-Rolls/dp/0198871104/ref=nodl_ https://www.amazon.com/Brain-Computations-Edmund-T-Rolls/dp/...
- mdp2021 5y ago
  > some kind of brain cell struct unit Ray Kurzweil presents one theory about the basic neuronal modules in How to Create a Mind (2012). The text is of course controversial - criticism may be as important as the tentative idea presented.
- markenqualitaet 5y ago
  I think you vastly overestimate our current understanding of the brain/neuronal processing, even singular (neuron) cell's workings. The fMRI enthusiasm turned out wrong a long time ago and there hasn't been much progress since, for all I know. And knowing a gene sequence wasn't enough either. We are still lacking theory big time. For most all of biology, we do not understand anything fully really. Not in neuro science, not in evolution theory, and god forbid not in genetics. Complex != complicated. Complexity is a hard barrier. We do not have abstraction for emergence, merely still recording thing's reactions when you poke them in different ways.
  - akomtu 5y ago
    Do we even understand how DNA replication works? It's a relatively simple mechanical process, but it's performed by molecular mechanisms that exhibit oddly intelligent behavior.
SeanFerree 5y ago
Amazing!

102 points by conse_lad 5y ago

Neuroscience is fascinating, well at least the part about image processing, but it's also completely inaccessible. From computer science and programming we are use to having introduction available in the form of articles, videos and cheap training. For Neuroscience it's just straight up hard-mode and you don't know any of the words. I still not even completely sure what a brain atlas is. I've seen a few free university courses, but they fill up almost immediately.

mdp2021 5y ago
YT, for example, contains many resources on brain anatomy and more. Plus, for example, the lessons from Robert Sapolsky at Stanford. Plus, there exist the primer books from Rita Carter (first of all, Mapping the Mind). There are loads of good material around. Note though, that CS is "forward" engineering, while Brain Science is backward engineering. So, if you do not find material that build from the ground up... it is easily because we are not at that level.
Frazmatazz 5y ago
This is decent. https://www.edx.org/xseries/harvardx-fundamentals-of-neuroscience https://www.edx.org/xseries/harvardx-fundamentals-of-neurosc... Comparable to the intro biological psychology course I took in college. It's free if you don't care about certificate.

vanderZwan 5y ago

So I expected this to be about the human brain since sites like http://mousebrain.org/ http://mousebrain.org/ already existed, but it looks like it's research that has been done on mice brains as well? Can someone explain the differences here?

stenl 5y ago
Hey vanderZwan, we have a paper in the collection (there’s 17 brain atlas papers in this issue of Nature) on human, marmoset and mouse motor cortex (my group did the human). I think it’s cool that pretty much all the cell types could be aligned between those species, which diverged 90 million years ago. Our full human brain cell type atlas is work in progress but hopefully we’ll post the preprint early next year. (Edit) Direct link to the papers, which are all open access: https://www.nature.com/nature/volumes/598/issues/7879 https://www.nature.com/nature/volumes/598/issues/7879
pamplemoose 5y ago
the headline is attention seeking since there are numerous pre-existing brain atlases that describe the anatomy/cellular/molecular parts in different ways. You can check https://portal.brain-map.org/ https://portal.brain-map.org/ for many examples. I think a good comparison to this kind of work is what has happened previously in astronomy, where scientists work to answer their specific questions using large open datasets acquired using shared infrastructure (fancy telescopes for astronomy, microscopes for biology). Here is some more ongoing work from the Allen institute where you can make a scientific proposal for their openscope (https://alleninstitute.org/what-we-do/brain-science/research/products-tools/openscope/ https://alleninstitute.org/what-we-do/brain-science/research...)

swashbuck1r 5y ago

Are there any examples of "Hello World!" with simulated basic brain cells? I'm sure this is an entirely naive question given the complexity of a brain, but I'm imagining a rudimentary program to help understand brain-style processing with some kind of brain cell struct unit that represents real-ish input/output mechanisms that can be connected to other brain cell struct units...leading to some minimal brain-style processing outcomes.

andyxor 5y ago
If you want to understand how the brain works this is a good intro with some realistic neuronal network models ( spoiler: these have nothing to do with “artificial neural nets” as we know them) https://www.amazon.com/Brain-Computations-Edmund-T-Rolls/dp/0198871104/ref=nodl_ https://www.amazon.com/Brain-Computations-Edmund-T-Rolls/dp/...
mdp2021 5y ago
> some kind of brain cell struct unit Ray Kurzweil presents one theory about the basic neuronal modules in How to Create a Mind (2012). The text is of course controversial - criticism may be as important as the tentative idea presented.
markenqualitaet 5y ago
I think you vastly overestimate our current understanding of the brain/neuronal processing, even singular (neuron) cell's workings. The fMRI enthusiasm turned out wrong a long time ago and there hasn't been much progress since, for all I know. And knowing a gene sequence wasn't enough either. We are still lacking theory big time. For most all of biology, we do not understand anything fully really. Not in neuro science, not in evolution theory, and god forbid not in genetics. Complex != complicated. Complexity is a hard barrier. We do not have abstraction for emergence, merely still recording thing's reactions when you poke them in different ways.
- akomtu 5y ago
  Do we even understand how DNA replication works? It's a relatively simple mechanical process, but it's performed by molecular mechanisms that exhibit oddly intelligent behavior.

SeanFerree 5y ago

Amazing!