I don't know why, but this shit has been on my mind since I was like 10

Burnblaze
Burnblaze

I don't know why, but this shit has been on my mind since I was like 10:
What is stopping people from creating prosthetics that are equal (or at the very least close to) to how human limbs work?
From what I understand, making "arms" that move with the precision of a human arm isn't a problem, since we can create a lot more precise shit already. And the people who lose arms and shit still have the neurons that would normally control the missing limbs, so is the problem in making an interface between the arm and the neurons?
I'm in no way an expert on either biology or robotics(?), so the answer might be really simple, but I'm still very curious.

All urls found in this thread:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3497935/
WebTool
WebTool

Do adults grow enough to have to get their prosthetic replaced?

BlogWobbles
BlogWobbles

Making artificial muscle fibers which can be controlled neurally ain't exactly easy.

BunnyJinx
BunnyJinx

@BlogWobbles
right, what I had in mind was that if we can create "limbs" that can have really high precision, but that aren't controlled neurally, then wouldn't it be possible to create some sort of interface between the neurons in the remaining part of the arm/leg (or maybe right at the spine?) that would basically read the way that the neuron is firing, and translate it to commands for the artificial limb.
And like I said, not in any way an expert in this shit, so idk if that kind of interface is possible, but if it isn't, I would like to know why.

@WebTool
No idea, but I don't think it would stop people from trying to build cool shit.

askme
askme

I NEED MY ANSWERS, Veeky Forums!

Sharpcharm
Sharpcharm

@Burnblaze
The mechanical side of prosthetics is easy compared to "getting a signal and processing it". If you use electrodes on skin you get a lot of noise, but it's good enough for basic movement. Connecting to actual neurons is hella hard and invasive. Also don't forget that you need processors or microcontrollers, actuators and batteries in one small package.

TechHater
TechHater

@Sharpcharm
Right, kind of what I assumed. I wonder just how hard and invasive connecting to actual neurons would be, because I feel like if a proper connection could be made, then creating really precise prosthetics would become possible. Also I wonder if such a connection would be able to send signals back to simulate the sense of touch.
Also wouldn't the limb itself be big enough to hold most of the shit?

Soft_member
Soft_member

@Burnblaze
You're right. Why haven't you done it yet?

farquit
farquit

@Soft_member
I never mentioned that this was easy, just that I didn't know why it was hard, and that I was interested in finding out why.

WebTool
WebTool

@TechHater
There are no neural interfaces. Neurology is a witch doctor tier science at the moment. How or why neurons do anything is a fucking mystery. Everything is trial and error. That's why ever medicine that affects the brain has insane possible side effects.

So to calibrate a neural interface, you're going to need to do it by trial and error on a healthy volunteer. And even then it isn't guaranteed to work for anyone else. You could easily be feeding pure pain to someone's brain instead of a touch sensation. Good luck finding volunteers willing to potentially suffer permanent brain damage.

FastChef
FastChef

@WebTool
huh, this is actually pretty surprising to me.
I kinda assumed that we had the basics of what and how and why neurons do shit figured out already. Not sure why I assumed that, but it's nice to know the answer.
Though I still what the next step would be.
Like even if you ignore the sending information back to the brain part, just simply reading the output from the neurons would be interesting, but I guess that's also a big problem.

Nojokur
Nojokur

@FastChef
There's no reading the output. It's an analog signal level through nerves that don't necessarily take a sensible route to activate a muscle. The most advanced read only tech we have right now senses muscle twitch, which isn't terribly useful when the muscle is gone.

Crazy_Nice
Crazy_Nice

@Nojokur
hmm, what I assumed would be possible to do would be to connect some sort of "reader" to the neurons, to basically read the signal, but I guess they're too small for us to properly build anything of the sort?

CouchChiller
CouchChiller

@Crazy_Nice
The best versions of this we have are called genetically encoded calcium indicators. If they neurons could be selectively tagged, they would emit light when they fire which could be detected. The ethics of introducing transgenics in humans is still up for debate.

Neurology is a witch doctor tier science at the moment
Pretty much what this guy said. Biologists/Doctors are idiots who can't figure out manmade computers let alone ones that are a product of evolution. The people who know enough about math/computers to understand the brain are too busy jerking off to models or too autistic/pedantic over small details that they completely frame the problem wrong.

Supergrass
Supergrass

@Crazy_Nice
The tricky thing is for it to be small enough to fit, the right material to not have the body actually fight its presence, and the sensitivity to obtain data from one spot and only one spot. And even then you're talking surgery to get the thing in there and you won't know what that neuron even does until you start reading it.

Here's what we CAN do with current tech, we could basically wire an "exo" skeleton inside you. We can add extra bones made of titanium connected to servos that monitor muscle twitch and move accordingly. It'd all be controlled by your normal movements and it would significantly add to your weight but it'd dramatically increase the upward limits of your strength.

What it can't do is make you faster or let you safely jump higher because remember, you're still wrapped in meat. So if we move a servo faster than your muscles can follow along, they'll rip.

But really, why not just have the exoskeleton on the outside where it belongs. It's a lot easier to armor too and won't cause serious damage if its battery leaks.

Dreamworx
Dreamworx

@CouchChiller
genetically encoded calcium indicators
pretty neat.
Also yeah, I don't think you'd even need to understand the brain to be able to do this kind of shit, no?
Like, if you could get the neural output from the motor neurons, you'd be able to convert that output into sensible commands to artificial limb.
I guess the problem would be getting that info fast and accurately enough, and with something that won't take up a whole bunch of space. Oh and I guess the amount of motor neurons needed for fine motor skills would be pretty insane?

Booteefool
Booteefool

@Supergrass
ooo... Now that you mention it, exoskeletons do pop into my head sometimes too.
So if we kind of have the current tech already, why aren't we seeing exo-skeleton workers and/or combatants? Just too expensive for what they'd bring to the table?

iluvmen
iluvmen

@Dreamworx
Also yeah, I don't think you'd even need to understand the brain to be able to do this kind of shit, no?
How do you arrive at this conclusion? The entire issue is interfacing to someones nervous system well enough to make the prosthetic as good or better than the original. To do that you need a complete understanding off how and why the human nervous system works, and as other anons have said neuroscience is still black magic tier.

FastChef
FastChef

@iluvmen
hmm, well my reasoning was that you wouldn't need to understand the whole of the brain if you could just read the output from the motor neurons.
The brain would do what the brain does with normal limbs, we would just use the output of that to figure out what our artificial limb would have to do. Mind you that would probably be a pretty big problem in and of itself.

Nojokur
Nojokur

@Burnblaze
They can be hacked remotely, meaning someone could literally make you kill yourself anytime they want.

When they create prosthestics that cannot be hacked by signals then sign me up until then fuck no.

Spamalot
Spamalot

@Nojokur
They can be hacked remotely
what?
Unless someone specifically designed them so that they could be connected to remotely, then you can't hack them remotely, and I don't really see a reason why anyone would design them in that way.

eGremlin
eGremlin

To make this possible you would need a neural net processor - a learning computer

Playboyize
Playboyize

@Burnblaze
What is stopping people from creating prosthetics that are equal (or at the very least close to) to how human limbs work?

The firing system interface with the nervous system.

TreeEater
TreeEater

@Nojokur
The most advanced read only tech we have right now senses muscle twitch

I find that hard to believe. It should be rather easy detecting the potential difference travelling along a single strand of neurons, we have equipment that can ready minute fields.

Skullbone
Skullbone

@eGremlin

Not even.

All you need is an algorithm which senses various inputs and uses a self-optimizing function to find the best response. None of that bullshit.

Firespawn
Firespawn

@Nojokur
not wanting rape hand jobs
Gay

lostmypassword
lostmypassword

Bump.

Ignoramus
Ignoramus

@Sharpcharm

Adding to that, what about power? I hardly think a mechanical arm could run on the body's heat or the little bit of electricity that runs through our bodies. It would need its own power supply, and that would need constant recharging.

Gigastrength
Gigastrength

@Ignoramus
well yeah, just how you need to eat to charge yourself, you'd need to charge the limb. Though I don't think it would be that hard, leave it to charge over night or some shit and you're good I guess?

StrangeWizard
StrangeWizard

Isn't there already research into this? iirc working prototypes of this exist already. The main problem is development for such a small subset of people and of course the invativeness of testing something that connects to actual nerves.

lostmypassword
lostmypassword

@Supergrass
What it can't do is make you faster or let you safely jump higher because remember, you're still wrapped in meat. So if we move a servo faster than your muscles can follow along, they'll rip.

Is this true? Like if someone is in a car crash and their arm opens too fast from the impact it'll tear their bicep, even if the muscle doesn't reach the end of its range?

Methnerd
Methnerd

@TreeEater
It's possible to detect and measure a twitch because its amplitude is so much greater than the noise for that twitch. It remains difficult to differentiate continuous contraction from no contraction because there is a fair share of integration done at the effector ends, not to mention filtering the "useful" information from the errant electrical activity in the nerves.

PackManBrainlure
PackManBrainlure

@TreeEater
It should be rather easy
Then why haven't you done it?

eGremlin
eGremlin

@Burnblaze
Because those that need it cannot pay for it. Those that can pay for it are smart enough not to lose limbs.

Going to war and losing a limb is not exactly a profitable endeavor, which would be your primary limb recipient. I bet it's cheaper to train a new soldier than buy an old one a new limb and rehabilitate him with it.

JunkTop
JunkTop

@Burnblaze
1. it's hard to connect mechanical parts to nerves so they will works as intended
2. no good way to power them

Soft_member
Soft_member

@lostmypassword
Lol.

Inmate
Inmate

@WebTool
there are programs that work with signals directly from the brain instead from the arm, that is using a kind of helmet to read brain activity. I've seen the chimps do it

Emberburn
Emberburn

@Burnblaze
What is stopping people from creating prosthetics that are equal
everything

First and foremost, interfacing with nerves. We can interface with nerves, just not very well. Interfacing with individual nerves is difficult and current interfaces tend to degrade over time. Although we have made some progress with the degradation, at least we're getting to the point where interfaces don't kill nerves.

Another big problem is physically attaching the prosthetic. One can use straps and suction cups to attach to the stump, but leads to an irritated stump and not much load capacity. We can bolt the prosthetic to bone, but then you have a metal bolt going through the skin. This can lead to infections in the metal bolt which are difficult for the body to fight because the metal doesn't have blood flowing through it. In addition, because metal is so much stiffer than bone, load isn't transferred very well to the bone so we get less load capacity and damage to the bone.

In addition, current actuators don't have as much torque density as muscle, but we're making progress there. Artificial muscle is not needed, and rotary electric actuators can do everything that regular muscle can

Methshot
Methshot

Fundinggggg

farquit
farquit

@Gigastrength

That sounds shitty though. I already hate having to charge my smartphone every 3-4 days.
Unless the charge in the augmented body part holds for at least a week, it will be DOA

Stark_Naked
Stark_Naked

@farquit
You're a dummy.

Soft_member
Soft_member

@CouchChiller
hurr this major field of scientific inquiry is only facing problems because *muh favorite field* is too euphoric to try but if they did they would totally teleport behind the secrets of nature
lmao

Sharpcharm
Sharpcharm

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3497935/

ZeroReborn
ZeroReborn

@Stark_Naked
?

StrangeWizard
StrangeWizard

@ZeroReborn
Current leg prosthesis hold a charge for about a day and they sell just fine.

Garbage Can Lid
Garbage Can Lid

@StrangeWizard

For people who have the option between that or no leg.
I thought we were talking about replacing working limbs with better ones.

Nojokur
Nojokur

@Garbage Can Lid
@Garbage Can Lid
@StrangeWizard
(You)

I thought we were talking about replacing working limbs with better ones.
You might want to reread your opening post, dummy.

Disable AdBlock to view this page

Disable AdBlock to view this page