• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

UF scientist grows "brain" in a dish that flies simulated plane.

Status
Not open for further replies.
R

Ripclawe

Banned
http://www.napa.ufl.edu/2004news/braindish.htm


GAINESVILLE, Fla. --- A University of Florida scientist has grown a living “brain” that can fly a simulated plane, giving scientists a novel way to observe how brain cells function as a network.

The “brain” -- a collection of 25,000 living neurons, or nerve cells, taken from a rat’s brain and cultured inside a glass dish -- gives scientists a unique real-time window into the brain at the cellular level. By watching the brain cells interact, scientists hope to understand what causes neural disorders such as epilepsy and to determine noninvasive ways to intervene.

As living computers, they may someday be used to fly small unmanned airplanes or handle tasks that are dangerous for humans, such as search-and-rescue missions or bomb damage assessments.

“We’re interested in studying how brains compute,” said Thomas DeMarse, the UF professor of biomedical engineering who designed the study. “If you think about your brain, and learning and the memory process, I can ask you questions about when you were 5 years old and you can retrieve information. That’s a tremendous capacity for memory. In fact, you perform fairly simple tasks that you would think a computer would easily be able to accomplish, but in fact it can’t.”
Click to expand...

When DeMarse first puts the neurons in the dish, they look like little more than grains of sand sprinkled in water. However, individual neurons soon begin to extend microscopic lines toward each other, making connections that represent neural processes. “You see one extend a process, pull it back, extend it out – and it may do that a couple of times, just sampling who’s next to it, until over time the connectivity starts to establish itself,” he said. “(The brain is) getting its network to the point where it’s a live computation device.”

To control the simulated aircraft, the neurons first receive information from the computer about flight conditions: whether the plane is flying straight and level or is tilted to the left or to the right. The neurons then analyze the data and respond by sending signals to the plane’s controls. Those signals alter the flight path and new information is sent to the neurons, creating a feedback system.

“Initially when we hook up this brain to a flight simulator, it doesn’t know how to control the aircraft,” DeMarse said. “So you hook it up and the aircraft simply drifts randomly. And as the data comes in, it slowly modifies the (neural) network so over time, the network gradually learns to fly the aircraft.”

Although the brain currently is able to control the pitch and roll of the simulated aircraft in weather conditions ranging from blue skies to stormy, hurricane-force winds, the underlying goal is a more fundamental understanding of how neurons interact as a network, DeMarse said.
Click to expand...
 
C

Chrono

Banned
heh first thing I thought when reading this was Ghost in the Shell: Stand Alone Complex. :lol

I don't have any science degree so I don't know what I'm talking about.... but I wonder if one day (20-30 years from now) it becomes possible for a brain like that to develop a consciousness. If it has the same power as a human brain and they let it interact with as much information.... Tachikoma? :D
 
ManaByte

ManaByte

Gold Member
6305308802.01.LZZZZZZZ.jpg
 
H

happyfunball

Member
What the??? Im a biologist and I still have no clue what the fuck he just did? This can't be as monumental as they are making it out to be...
 
T

teiresias

Member
Sounds just like a neural network that one would build in hardware or software, where you have a mesh of neuron models and feed it data and feedback in order to train it. This has been done in software models for a while, but it's weird to see someone doing the same process with actual biological neurons.
 
3

3phemeral

Member
I still don't understand how the brain recognizes what the "correct" way is to fly. They mentioned flight conditions as one of the variables used to test the brains ability to respond, but in what ways does the brain know how to fly? Are they consequences like, the plane will crash or fly too high, and if it does, how does the program relay the information to the brain as 'incorrect.' Otherwise, I'm just assuming that the computer is sending data on how to fly the plane, while the brain slowly incorporates that information through the establishment of neural networks until it no longer needs the computer to respond since, by that time, it will already have a flight database for reference.
 
N

nitewulf

Member
3pheMeraLmiX said:
I still don't understand how the brain recognizes what the "correct" way is to fly. They mentioned flight conditions as one of the variables used to test the brains ability to respond, but in what ways does the brain know how to fly? Are they consequences like, the plane will crash or fly too high, and if it does, how does the program relay the information to the brain as 'incorrect.' Otherwise, I'm just assuming that the computer is sending data on how to fly the plane, while the brain slowly incorporates that information through the establishment of neural networks until it no longer needs the computer to respond since, by that time, it will already have a flight database for reference.
Click to expand...

neural nets are built using small blocks. each small block is governed by a strict set of rules that they must maintain in order to survive. in this example, each particular block's existance could depend on keeping the flight level. if the place crashes, the information will be fed back (hence its a feedback system!) to the "brain", and the "brain" will be trained to recognize that output condition as unacceptable. it will be trained to accept values within a certain range, ie, the range where flight conditions will be normal. and each instant the simulation will keep sending the brain signals, if one signal goes too high or too low, the brain will manipulate the controls (ie, pitch, roll, yaw) to make the necesarry adjustments to bring the flight level, and output conditions to normal, acceptable levels.
 
3

3phemeral

Member
nitewulf said:
neural nets are built using small blocks. each small block is governed by a strict set of rules that they must maintain in order to survive. in this example, each particular block's existance could depend on keeping the flight level. if the place crashes, the information will be fed back (hence its a feedback system!) to the "brain", and the "brain" will be trained to recognize that output condition as unacceptable. it will be trained to accept values within a certain range, ie, the range where flight conditions will be normal. and each instant the simulation will keep sending the brain signals, if one signal goes too high or too low, the brain will manipulate the controls (ie, pitch, roll, yaw) to make the necesarry adjustments to bring the flight level, and output conditions to normal, acceptable levels.
Click to expand...


I understand that it's a feedback system. I well aware of the concept behind the communication involved. What I'm wondering is how the brain differentiates between "unsuccesful" and "successful." You don't just send out a command that says, "Your plane has crashed, Game Over. Please Try Again." It's easy to say that it's "trained" to do this and that, but how do explain the process indicating to the brain that certain values are not within an acceptable range?

-edit- Re-reading your post, I just realized that I didn't see the whole "dependent on survival". That would make sense, but under what conditions does the brain experience this? *zap zap* Maybe they could use electro-shock. :)
 
S

Seth C

Member
3pheMeraLmiX said:
I understand that it's a feedback system. I well aware of the concept behind the communication involved. What I'm wondering is how the brain differentiates between "unsuccesful" and "successful." You don't just send out a command that says, "Your plane has crashed, Game Over. Please Try Again." It's easy to say that it's "trained" to do this and that, but how do explain the process indicating to the brain that certain values are not within an acceptable range?
Click to expand...

Exactly. That's what I don't understand. Unless they're really using another computer (software app) to control the brain cells, in which case all they are is an organic pass-thru.
 
R

radioheadrule83

Banned
My understanding (or lack of) was that the neurones reach out to one another and form a network. Then they maybe threaten the survival of that network, thereby teaching it how to survive: aka fly a plane. Pretty decent.

"Hello, this is a brain in a dish speaking. We are now at an altitude of 35000 feet, and if you look out to your right you will see the expanse that is the pacific, and your life flashing before your eyes. Enjoy the flight!"
 
F

Funky Papa

FUNK-Y-PPA-4
Nice, a bunch of rat neurones can play Flight Simulator while I still need a calculator to resolve a simple equation.
 
L

Lord Error

Insane For Sony
I saw this news the other day. I too don't understand how exactly they condition the 'brain' to understand that he did the wrong thing by crashing the plane. I doubt it's something like electroshocks, because how would a bunch of neurons understand that something zapped them, and that they should try different output/configuration next time?

Admittedly, I don't know much about the biology involved. A good friend of mine who's a chemistry Phd, works on the university in Gainesville. I'll ask her to inquire about this project and tell me how it works in more detail.
 
V

Vgamer

Member
I dont understand how a rat brain can fly a plane when it takes tons of training for humans to fly one. Makes no sense to me.
 
3

3phemeral

Member
Marconelly said:
I saw this news the other day. I too don't understand how exactly they condition the 'brain' to understand that he did the wrong thing by crashing the plane. I doubt it's something like electroshocks, because how would a bunch of neurons understand that something zapped them, and that they should try different output/configuration next time?

Admittedly, I don't know much about the biology involved. A good friend of mine who's a chemistry Phd, works on the university in Gainesville. I'll ask her to inquire about this project and tell me how it works in more detail.
Click to expand...

I was just joking with the electro-shock treatment :)

Ah.. it'd be interesting to know what she says. From my brief stint in "Brain and Behavior," conditions that would threaten the life of the cell would be the lack of a proper nutritional environment, which the neruon utilizes to sustain optimal functioning conditions -- but there are a few other mechanisms in the brain that detect these deficiencies and motivate the person to eat, drink, or whatever they need to alleviate the problem. How they accomplish this without those fail safes is a little weird to me.

Vgamer said:
I dont understand how a rat brain can fly a plane when it takes tons of training for humans to fly one. Makes no sense to me.
Click to expand...

Hrm.. I suppose that in this instance, there's no real thinking involved. Ignoring how it's trained, it's just a mass of neurons devoted to a specific task, conditioned to respond to set of variables with no other interaction. Think of all the things in your life that you juggle consistently -- your daily activities, emotions, etc. The mass of rat brain cells don't have any of these things to worry about, because they're devoted to one task, whereas you have multiple things to work through on a daily basis.

I doubt these brain cells would be able to problem solve, say if an another plane headed on a collision course entered the equation. If it's never been exposed to this variable, it won't be able to react and would probably crash because of it.
 
Phoenix

Phoenix

Member
teiresias said:
Sounds just like a neural network that one would build in hardware or software, where you have a mesh of neuron models and feed it data and feedback in order to train it. This has been done in software models for a while, but it's weird to see someone doing the same process with actual biological neurons.
Click to expand...

Yep. Isn't really too surprising since both models are just trying to mimic what we know about the brain.
 
B

BugCatcher

Member
The article makes some assumptions on what this technology could be used for, and I think it's causing people to read a bit too much into this experiment. The purpose was primarily to study how networks of neurons interact, but it's been hyped up to sound like they grew brains and conditioned them to perform some arbitrary behavior. It can't be trained, because there is no punish/reward mechanism in the experiment. Here is how it works, straight from the horse's mouth:

We did not report LTP because it is NOT LTP. In fact, we are using and effect reported by Eytan, D., Brenner, N., and Marom, S., Selective Adaptation in Networks of Cortical Neurons. Journal of Neuroscience, 2003. 23(28): p. 9349-9356 in which "high" frequency stimulations (once every second) was reported to depress the response of the network while "low" frequency stimulations resulted in an enhanced response. For our system we tied the network's response to the control surfaces, dedicating stimulations on one channel for pitch, and a second for roll control. Each channel is stimulated separately, and the response (PSTH) is recorded. Control movements are proportional to the current error from straight and level by mapping the error (0 to 180 degrees) to the interval 0 to 100 ms of the PSTH and integrating the difference in response before training, to the current or enhanced or depressed levels. The more error, the more the control surface is moved. The networks only gradually control the aircraft since the Marom effect requires over 15 minutes to develop. The two frequencies are then used to adjust these weights (i.e. number of spikes in the PSTH) to produce optimal flight. The neurons/network don't seek optimal flight in the classic sense. Instead, we adjust the weights (using high and low Freq. stims) in the network to produce that result. It is a very simple system and our only interest in it is in terms of those changes within the network and the possibility to extend it to more of the network than just two or three different channels. Hope that helps.. Tom DeMarse
Click to expand...

Here is a simpler explanation:
the network is not "learning". Rather, they are setting up the system so that the inherent properties of the neurons cause the correct response to the feedback it receives from the environment.

The real knowledge about the task is built into the systems that interface with the neurons.
Click to expand...

It may not be as cool as the article would lead you to believe, but it's still impressive work.
 
3

3phemeral

Member
That makes more sense that they 'encourage' specific neuronal connections to take place. Would have been cool if it was learning, though. That'd be something to think about.
 
Status
Not open for further replies.

Similar threads

Ken Kutaragi is the Greatest 'emperor' Of All Time and remains SIE's only true visionary
Opinion 2 3 4
156
16K
GAF machine replied
Cyber Chip - Brain/Microchip Hybrid introduced to accelerate AI
4
959
Nhranaghacon replied
Xbox Black Friday starts on November 17
News Hype Platform 2 3
121
10K
Leonidas replied
I have a theory about the existence of souls and ghosts.
Opinion 2 3 4 5 6
255
23K
nush replied
Satya Nadella: Annual letter to stockholders. Content on new platforms will continue
4 5 6 7 8
352
20K
noob smokes replied
Top Bottom