[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y ] [Search | Home]
4Archive logo
Talk shit about my field:
If images are not shown try to refresh the page. If you like this website, please disable any AdBlock software!

You are currently reading a thread in /sci/ - Science & Math

Thread replies: 68
Thread images: 3
Talk shit about my field:
- Systems Neuroscience.

I study how the brain processes visual information experimentally and computationally.
>>
>Talk shit about my field
OK

Correlation =/= causation
>>
Sounds fairly interesting.

Sorry I didn't provide the sperg out you were hoping for.
>>
How many animals do you kill a week
>>
>>7764738
Optogenetics, bro. I can cause neurons to spike and measure responses. The causal loop has been closed
>>
>>7764740
Maybe 10 mice
>>
I'm interested in that shit. What do you do exactly
>>
>>7764817
I study a rhythm in the brain called 'gamma oscillations' in the visual system. the stuff I'm working on and about to publish is on the cellular basis of this oscillations, as in these cells connect to each other in this way and therefore the circuit generates this resonant cycle. Hopefully my work will result in understanding how we process visual data into objects and other abstractions.
>>
>>7764735
>you fags still haven't come up with a way to stream HD video direct to the visual field
Fucking get with the program OP.
>>
>>7764955
Listen, I could actually totally do that. No joke. It's just super invasive.
Calcium imaging coupled with holographic laser stimulation optogenetics. It's literally going on in my lab right now.
>>
>>7764958
>I could actually totally do that.
[citation needed]

Also wrap it up in a commercial package please so I can buy it.
>>
>>7764958
>super invasive
Wait, how invasive?
>>
>>7764964
It's a thing going on at a bunch of labs. I work at one of these labs
https://projectreporter.nih.gov/project_info_Like.cfm?aid=8572797&icde=27633684
>>
>>7764967
Fairly invasive. No electrodes in the brain though, which is nice.

1. make a window in your skull so we use a microscope to look into your brain
2. inject a virus in your visual cortex containing dna for a calcium indicating protein so that we can see when your neurons fire.
3. inject a virus in your visual cortex containing dna for optogenetic proteins so that we can fire and silence them when we want using light.
4. train software to know what makes each neuron fire (what features of visual space)
5. make neurons fire to replicate what your brain does when it is seeing things
>>
>>7764978
How big does the window have to be?
Could it be sealed up with hardware left inside or does it need to remain open?
>>
>>7765044
The window to the brain needs to be large enough to contain the field of view of the microscope objective. All things considered, the largest it could be would be 1.5 inches.

No it could not be sealed up. You would be attached to a gigantic rig that costs around $250,000. Your head would be locked in place so your brain doesn't move.
>>
>>7765058
>No it could not be sealed up. You would be attached to a gigantic rig that costs around $250,000. Your head would be locked in place so your brain doesn't move.
That makes it rather impractical don't you think?
I just want to be able to have a HUD and web browser access direct from my visual field day-to-day. Do you guys have anything that might permit that?
>>
>>7765067
google glass.
If you want something that interfaces directly with your brain... you either have to go with the giant expensive immobile rig, OR we have stuff that might work decently today, but is not perfect:
http://www.cell.com/neuron/abstract/S0896-6273(14)01005-8?_returnURL=http%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0896627314010058%3Fshowall%3Dtrue

In case you're wondering, none of this has ever been tested in humans, but based on a priori principles, most people in the field agree that humans would actually be a much better system to work in. The neurons are bigger and code for more explicit visual features. Only problem would be that we couldn't see that many neurons at a time (maybe 5000 or so), which is probably enough to get this to work, but is not perfect. The trade is resolution for greater numbers of neurons.
>>
Judging this thread, is this field /sci/-approved?
>>
>>7765089
Judge me, bro. Anyone remember a synchtube channel called DrRichard? Guess who
>>
>>7765089
Systems neuro is fine. Psychology and fMRI are garbage
>>
>>7765097
seriously... could it be...?
>>
>>7764941
>>7764735
>>7764958


But can you see why kids love cinannon toast crunch?
>>
>>7765080
Google glass is pretty nifty, but it comes up a bit short I think. Really want a direct nervous system interface.

The electrodes, ones I've looked at anyway, seem to have a fairly shitty number of electrodes (in the hundreds) which just isn't good enough. Need at least 800x600 for it to be usable really.
>>
File: 1420947728410.gif (437 KB, 420x420) Image search: [iqdb] [SauceNao] [Google]
1420947728410.gif
437 KB, 420x420
>>7764735
I won't, it looks awesome
>>
So answer a couple of questions I've long wondered about.

If you have few or no neurons connecting the two sides of the brain, you have a problem in which you can only identify things to the right. For example, if you stand on main street in your home town, you can identify the stores to your right but not to your left. Go down a block and turn around and you can identify the stores to your right that you could not identify previously when they were on your left and you cannot identify the stores you identified previously that were on your right but are now on your left.

It seems clear that that it is because of the lack of pathways from the right visual cortex (which processes your left field of view) to the left fusiform gyrus.

So here are the questions:
1) Prosopagnosia (face blindness) is apparently the result of a problem with the right fusiform gyrus. Is someone with the same problem above only able to identify people to their left but not to their right?
2) In the case of albinos (and obviously Siamese cats) the optic nerves apparently do not regroup in the optic chiasm so the left visual cortex processes all of the data from the left eye and the right visual cortex processes all of the data from the right eye. How much is an albino affected if they have few or no connections between the left and right brains?
>>
And another question.

How is the processing in the visual cortex different In tetrochromats such as birds which have an extra set of primary colors?
>>
>>7764978
i like this op nigga
>>
>>7764735
>I study how the brain processes visual information
So what do you know about how it works
>Absolutely nothing
Okay...
>>
>>7766837
>>7766845

Waiting for an answer.
>>
>>7766837
>1) Prosopagnosia (face blindness) is apparently the result of a problem with the right fusiform gyrus. Is someone with the same problem above only able to identify people to their left but not to their right?
Yes, if the assumptions are as simple as you stated. In reality you get a little bit of left-eye info to your right brain and vise-versa, making it possible that you might be able to discern a face. That said, there are beautiful studies done by gazzaniga that show interesting effects of showing split-brain people objects in their non-dominant eye-field. Their responses in identification are often correct, but seemingly sub-conscious.

>2) In the case of albinos (and obviously Siamese cats) the optic nerves apparently do not regroup in the optic chiasm so the left visual cortex processes all of the data from the left eye and the right visual cortex processes all of the data from the right eye. How much is an albino affected if they have few or no connections between the left and right brains?
Nobody knows. Why do sensory nerves switch sides... I don't know. Is there a benefit I'm unaware of? There must be.
That said, I'd guess there would be more poor integration of contrast in the fields of view, due to the bionocular integration that typically takes place at even low visual processes regions in the brain.
>>
>>7766845
Some humans, typically women, are tetrochromats (https://en.wikipedia.org/wiki/Tetrachromacy#Human_tetrachromats).

As far as what the experience would be like... I'm not positive, but if we are talking about a new photoreceptor in a significantly different frequency range than the other ones (say IR or UV), then there are two possible options:
1. We would experience a new color. This is not unthinkable. If we consider how dichromats (color blind people) see the world. In those that are missing the red receptor, they can only see the differences between blue and green. Their green receptor picks up all the red they will see. So, what colors do they experience? They don't associate green with anger and hatred, nor do they associate red with that emotion. (I think this shit if fascinating) In most of these people, that association between RED and RED EMOTIONS is utterly missing. They don't experience it.

Now, this doesn't mean that we would experience new emotions per se, but it might...

2. Normalization of color space; which seems less likely
>>
>>7766983
We know a lot about how it works! We have been able to recompose movies seen by the visual system for years. We are now considerably further along in our understanding.
https://www.youtube.com/watch?v=piyY-UtyDZw
>>
>>7764735
OP. I've been trying to gather neuro people from /sci/ to create a discussion group if you're interested?

We're in the process of finding enough people to make it worthwhile.

As for myself, I'm interested in cellular and molecular neuroscience, particularly neurodegeneration.

If you want to, give me a throw away email on which I can contact you

I use cock.li for my throwaway
>>
>>7765534
Right... this shit is a lot more complicated than you might think. Each electrode would not act as a pixel as you seem to imply. There are really significant nonlinearities that would need to be compensated for.
An approach using parallel stimulating electrodes could would for a simple brain machine interface (https://en.wikipedia.org/wiki/Brain%E2%80%93computer_interface) but wouldn't allow the spatial resolution necessary for you to 'experience' images. For that you need something much more nuanced than stim electrodes (ie optical techniques). Take a glance at the paper I posted, it actually has a chance at success (http://www.cell.com/neuron/abstract/S0896-6273(14)01005-8?_returnURL=http%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0896627314010058%3Fshowall%3Dtrue)
>>
>>7768501
[email protected]
>>
>>7768520
sent you an email
>>
>>7768465
isn't that the sort of thing you could potentially do without any understanding whatsoever? just train a neural network with brain activity as inputs and video frames as outputs.

i'll need to see a paper.
>>
>>7765107
>psychology is garbage
Depends, what field?
>>
>>7768543
yup.
Dario L. Ringach Receptive Fields in Macaque Primary Visual Cortex Spatial Structure and Symmetry of Simple-Cell (2002)
J. H. van Hateren and D. L. Ruderman Independent component analysis of natural image sequences yields spatio-temporal filters similar to simple cells in primary visual cortex (2002)
>>
>>7768553
These fields:
http://www.nature.com/news/over-half-of-psychology-studies-fail-reproducibility-test-1.18248
>>
>Systems Neuroscience.
>Neuroscience
>Biology

NICE MEME DEGREE FAGGOT


GTFO
>>
File: 1167792994290.jpg (74 KB, 800x600) Image search: [iqdb] [SauceNao] [Google]
1167792994290.jpg
74 KB, 800x600
>>7768465

That's not "processing" but just reading in visual information. We don't know shit about how someone looks at a photo of a cat and realizes it's a cat.
>>
>>7768568
>Systems Neuroscience
>Biology, Chemistry, Physics, Computer Science, Math
>>
>>7768574
Right. I was just making a quippy rhetorical response. In reality, what matters are the algorithms that were implemented to carry out the computations necessary for that video to be made.
About coding for explicit representations... we know a good deal about that too. The term given in the field is 'Sparse Coding'
Read this http://www.nature.com/nature/journal/v435/n7045/full/nature03687.html
Also skim this
http://hubel.med.harvard.edu/book/bcontex.htm
>>
>>7768557
how are these papers related to that vid
>>
>>7768596
- They are recording from a population of neurons in the visual cortex
- They play random video to the animal which recording what on the screen made the neuron fire
- They correlate and average what makes each neuron fire
- They play a video to the cat
- They predict what that video looks like based on the known firing preferences of the neurons

in jargon: They map the STRFs, then reconstruct a movie based on the population activity while a video is played to the animal. This is the bread and butter of visual neuroscience
>>
>>7768604
but one of those papers is just doing ica on videos and the other seems like hubel and wiesel v1 stuff but with macaques
>>
>>7768642
Right. Those are the building blocks to getting an experiment like that to work. Here's Yang Dan's relevant papers:
http://www.nature.com/neuro/journal/v10/n6/full/nn1895.html
http://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.0030342
>>
How does it feel knowing that nonsynaptic plasticity may play a significant role in how the brain operates?

How does it feel knowing that we will never be able to properly simulate neurons because computers don't use true real numbers?
>>
>>7768701
>How does it feel knowing that nonsynaptic plasticity may play a significant role in how the brain operates?
Fine... I'm sure it plays a huge role

>How does it feel knowing that we will never be able to properly simulate neurons because computers don't use true real numbers?
'properly'? I'm confident that we'll get close enough. Compete simulations are a long way off, but we'll get there. C. Elegans is about 15 years away, fly is maybe 30, mouse is probably 60, humans maybe 70
>>
>>7768720
Do those estimates assume moore's law continues forever?
>>
>>7768899
The estimates are based on my experience in the field. I'm aware of Moore's law
>>
>>7764735

I hate how people on /sci/ now constantly use the phrasing "meme X" to denote that they are not interested or do not like the discipline of X.
>>
>>7764738
Lol doesn't even know how science works
>>
>>7769264
I haven't bene on sci in at least a year (used to be very active like 4-6 years ago), and yeah, it's weird. I'll assume there's a decent backstory and I'm just a newfag, but it still seems like a weak insult
>>
>>7769271
To be fair, a LOT of biology is correlational only. Neuroscience is also definitely behind the rest of the fields it integrates, in terms of methods and depth of understanding.
>>
your field is overhyped and over competitive. It's in a bubble right now with all the money getting splurged into it and the results its producing aren't worth it.
>>
>>7766983
>implying that because we don't no a step by step to how it works right now, it makes the science questionable

How do you think science operates? Immediate answers? Because if you think anything in science is accomplished quickly, you've never done any science.
>>
>>7768501
I'm trying to get into behavioral neuroscience. Currently a evolution bio/developmental bio student, coupled with mathematics.
>>
>>7768457

In the ice cube model of the visual cortex, are the different photoreceptors handled differently or do they form some kind of "layers" processed separately?
>>
>>7769317
science does not always yield practical results. However, somethings are worth knowing for their own sake. Don't you want to know the solution the hard problem of consciousness?
>>
>>7768457

Actually, I wasn't really asking about the experience, but if there were any significant difference in the processing. I originally came across this in one of Gordon Shepherd's books on neuroscience but he didn't address this issue.

I assume that for humans, the basic processing might not necessarily be different for the few tetrachromats who exist. But in birds, turtles, ... which are all tetrachromats (I guess there could be a color-blind tetrachromat that was really a trichromat) would there be an area of the brain for processing the extra color?
>>
>>7768576
>Biology, Chemistry, Physics, Computer Science, Math

Any biologist who wants to be worth a fraction of a shit studies all of these things
>>
>>7769345
Color blobs brosef
http://cercor.oxfordjournals.org/content/18/3/516.full

They are regions of low orientation selectivity in V1. They are better understood in higher animals like cats and monkeys (I study mice). The ice cube model doesn't really include this, since the blobs go in regions where there is poor orientation selectivity. Think hyper-hyper-icecube model. First hyper is binocular, second for other modalities like color and other esoteric stuff.

Fun fact, in mice, the upper half of their vision is receptive almost entirely to UV, and the lower half to green (like grass and sun). This is actually a really active area of research
>>
>>7769354
In short, yes. I'm not sure how the neuroanatomy differs in other animals, but there have been dooooope ass studies done in mice and monkeys to INDUCE TETROCHROMACY. Yes, new and segregated blobs form in V1 coding just for that new receptor type. This means that there is an activity dependent development mechanism, which is amazing.
>>
>>7769356
Yep. You don't really get into a decent graduate program with less than a proven proficiency is all of these subjects. I didn't take CS as an undergrad, and only took up through calculus in math, but after graduating I took a year off, worked in a lab, and had to teach myself to do some tough signal processing analysis. This was enough for grad schools, as it should it be.
>>
>>7769317
>>7769349
Neuro is not overhyped. Neuro is definitely overcompetitive though. I'm struggling not to get scooped as we speak. I work 60-80hr weeks just to be status quo.
Neuro has has had a boom in funding in the last 15 years. It's also produced incredible results that have already begun trickling into medicine. That said, I don't do neuroscience for the medical benefits, I'm in it for the long game: Total brain simulation, brain-machine-interfaces, undercovering general principles related to perception and experience, and uncovering fundamental truths about our thinking machine.
Thread replies: 68
Thread images: 3
Thread DB ID: 377873



[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

[Boards: 3 / a / aco / adv / an / asp / b / biz / c / cgl / ck / cm / co / d / diy / e / fa / fit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mu / n / news / o / out / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / t / tg / toy / trash / trv / tv / u / v / vg / vip /vp / vr / w / wg / wsg / wsr / x / y] [Search | Home]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the shown content originated from that site. This means that 4Archive shows their content, archived. If you need information for a Poster - contact them.
If a post contains personal/copyrighted/illegal content, then use the post's [Report] link! If a post is not removed within 24h contact me at [email protected] with the post's information.