Why do we sleep?

>>8727829
To give predators a chance to keep us humble.

Nobody knows, but it seems to have to do with how the brain processes memory

>>8727829
Cause I'm tired nigga

>>8727829
Allow your brain to remove the waste it accumulates from processing things while awake.

Because of the reason why humans sleep

>>8727908
What waste? What happens to the "waste" from processing things in your sleep? Because the brain certainly processes things while asleep, almost to the same extent as awake.

>>8727868
Best answer so far.

>>8727829
It helps us train our neural networks using backprop

>>8727829
because we have nervous systems our organisms relies on

Uninterrupted experience of the material world would be hellish.

>>8727829
Homeostasis. Just apply it to the whole of the human body and brain and you'll see sleep is beneficial.

>>8727829
Brain is like a sponge.
Absorbs info all day.
When you sleep it releases its plasma and expands.

Like a muscle relaxing after its hard work.

When you sleep you are activating this unified process.

Then why the fuck do I have ups and downs with concentration and alertness during the day ? Why do I spontaneously get sleepy ? Its always worse after meals.

There's evidence for several, not mutually exclusive, hypotheses.

1) Sleep provides the context for memory consolidation:
>http://www.nature.com/nrn/journal/v11/n2/abs/nrn2762.html
Basically, the idea is that during sleep, memories are transferred from hippocampus-dependent neural pathways to more robust storage in the neocortex. Quite a lot of work has been done on this particular hypothesis.

2) Sleep allows the brain to wash away accumulated debris from the intracellular space:
>http://science.sciencemag.org/content/342/6156/373
This is a relatively new hypothesis, but there is some evidence to support it. In the above paper, the authors show that during sleep, the cerebrospinal fluid increases in flow rate, and takes with it a number of proteins that are associated with neural degradation.

3) Sleep serves as a means of synaptic scaling:
>http://science.sciencemag.org/content/355/6324/507
During wakefulness, we learn all types of new things, and that means that some synapses get strengthened (i.e. more effectively transduct signals). To prevent our brain from entering a state of runaway excitation that would occur if synapses are allowed to be strengthened indefinitely, during sleep synapses are collectively weakened, so that the overall level of synaptic excitation is reduced, but regional differences in synaptic strength are maintained. This is the most recent hypothesis.

TL;DR: there's probably not just one reason why we sleep, but a few different reasons.

>>8728760
blood sugar, the brain needs glucose to do its job

just after you eat there's a peak in insulin release and an associated dip in blood sugar levels.

File: banana with glasses.jpg (20KB, 400x400px) Image search: [Google]

20KB, 400x400px

>>8727856
>>8727908
>>8728509
>>8728561
All just symptoms of ther real reason. We sleep cuz it's a very good way to minimize energy use during night. Humans a very bad at hunting and gathering during night, so we sleep to dont use up calories during night. The reason the brain needs "rest" is beacuse of something that happend after we first started sleeping.

>>8727856
>Nobody knows
>>8727930
>It helps us train our neural networks using backprop
>>8728536
>Uninterrupted experience of the material world would be hellish.
>>8728561
> you'll see sleep is beneficial.
>>8728595
>Brain is like a sponge.
fucking brainlets, honestly

at least google that shit before you decide to spray verbal diarrhea all over this board

>>8728871
>it's a very good way to minimize energy use during night
During sleep the metabolic rate of our brain is hardly any different from the waking state. It also doesn't account for vastly different times spend in the sleeping state across species. This hypothesis is out dated.

>>8727928
Toxic ideas you get from browsing 4chan. Sleep flushes it out for the most part, but it does bioaccumulate to a certain degree

File: disdain for plebs.jpg (9KB, 336x221px) Image search: [Google]

9KB, 336x221px

>>8728886
>mfw you just implied Geoffrey Hinton is a brainlet

>>8728905
What a spectacular way to demonstrate that you misunderstood his work.

File: tp-hinton-6830-1.jpg (126KB, 940x529px) Image search: [Google]

126KB, 940x529px

>>8728910
Wanna explain the misunderstanding, champ?

>>8728924
The algorithm was never intended to describe actual cortical dynamics.

File: 1440428498045.jpg (99KB, 1200x872px) Image search: [Google]

99KB, 1200x872px

Gotta flush those incorrectly folded proteinz out of your cells or they explode bruh...

>>8728932
That must be why Hinton keeps trying to convince everyone that the brain might do backprop: https://www.youtube.com/watch?v=VIRCybGgHts

>>8728945
kek, fucking computer scientists man

Backpropagation in neural dynamics is trivially present, we've known that for decades. He's not trying to 'convince' anyone, he's teaching a fucking class.

And if you're suggesting that sleep is a direct implementation of the "wake-sleep" algorithm, then you'd be hard pressed to find evidence that isn't circumstantial. That's also not what Hinton claims. This is your interpretation, and it's nonsensical.

>>8728956
>"wake-sleep" algorithm
That has nothing to do with backprop.

Put some more effort into your trolling, friend. Read that Wikipedia page properly!

>>8728966
>nothing to do with backprop
kek, are you fucking kidding?

>>8728942
does she have 3 legs?

>>8728974
no

>>8728993
Then the discussion is moot because you're obviously retarded. Backpropagation is a property that is unrelated to sleep, unless you forcefully relate it by bringing in the sleep-wake algorithm. Fuck off.

File: Relax sonny, no need to be mad, u mad.jpg (42KB, 500x415px) Image search: [Google]

42KB, 500x415px

>>8729003

>>8729022
I'll interpret that as you finally admitting to what we know all along.

>>8727829
>>8727856
I had a professor who said that we run a kind of data compression in our sleep. All day, we're taking in raw information (visual, audio, etc) and the brain has to store it quickly in our memory. During sleep, the brain has time to move it around and store it more compactly.

The evidence he gave was that the more you learn in a given day, the more vivid your dreams are. He was 100% right about that, in my experience. His little joke was "I hope you all have nightmares about this class" because that would mean we were learning a lot. And when you do have nightmares, they're almost always some weird mashup of things you saw/heard/learned that day.

File: Untitled.jpg (115KB, 1200x872px) Image search: [Google]

115KB, 1200x872px

>>8728942
>>8728982
I solved it

>>8729051
fucking hell, how could I ever have seen a leg in that?

>>8729056
Ha, I think the line of the bottom of her skirt lines up with the dark part on the top of the vase just right

>>8728805
this

>>8729022
>got taken to school
>hurr durr u mad
I thought /sci/ was 18+

God made it that way