December 29, 2002
Hybrot Robot Operated By Rat Brain Neurons

Steve Potter of Georgia Institute of Technology has built a hybrid rat neuron robot called a hybrot.

In his experiment, Potter places a droplet of solution containing thousands of rat neuron cells onto a silicon chip thatís embedded with 60 electrodes connected to an amplifier. The electrical signals that the cells fire at one another are picked up by the electrodes which then send the amplified signal into a computer. The computer, in turn, wirelessly relays the data to the robot.

Endless science fiction parallels come to mind. How about the Star Trek original series episode where Spock's brain was stolen in order to use it run a planet? Rolf Pfeifer of the University of Zurich, Switzerland foresees the use of neurons to make self-healing computer systems. Neuronal stem cells could be induced to form new connections to repair damage. Picture hybrot battlebots that would be silicon-biological hybrids that would be extremely difficult to kill.

If human neurons were used to make hybrots then how many neurons would it have to have before we'd hear demands for the recognition of hybrot rights? But if the hybrots were designed to desire to kill a large portion of humanity how could they be granted rights?

Details from a Georgia Tech web site:

Existing computational paradigms do not approach the capabilities of even simple organisms in terms of adaptability and real-time control. There are computational mechanisms and network architectures in living neural systems that are missing from even the most sophisticated artificial computing systems. This project consists of the development of computational systems that incorporate both living neuronal networks and artificial elements, including robotic testbeds and signal-processing circuitry. These hybrid neuronal-robotic systems (ĎHybrotsí) will provide a platform for discovering, exploring, and using the computational dynamics of living neuronal networks to perform real-time tasks in the physical world. Cultured networks of molluscan and mammalian neurons will be interfaced to robotic systems via multi-electrode array substrates capable of distributed, spatio-temporal stimulation and recording of neural activity. Unlike brains in animals, in vitro networks are amenable to detailed observation and manipulation of every cell in the network. Both high-speed optical recording, and time-lapse microscopy will be employed. By embodying the networks with actuators and sensors, the dynamical attractor landscape of neuronal networks will be studied under the conditions for which they evolved: continuous real-time feedback for adaptive behavioral contr

Update: The neurons can live for up to 2 years.

"We call it the 'Hybrot' because it is a hybrid of living and robotic components," he said. "We hope to learn how living neural networks may be applied to the artificial computing systems of tomorrow. We also hope that our findings may help cases in which learning, memory, and information processing go awry in humans."

The team uses networks of cultured rodent brain cells as the Hybrot's brain, and has essentially given the cultured neural networks a body in the form of a mobile robot. Potter's group hopes the research will lead to advanced computer systems that could some day assist in situations where humans have lost motor control, memory or information processing abilities. The neural interfacing techniques they are developing could be used with prosthetic limbs directly controlled by the brain. Advances in neural control and information processing theory could have application, for example, in cars that drive themselves or new types of computing architectures.

Inside Potter's lab, a droplet containing a few thousand living neurons from rat cortex is placed on a special glass petri dish instrumented with an array of 60 micro-electrodes. The neurons are kept alive in an incubator for up to two years using a new sealed-dish culture system that Potter developed and patented. The neural activity recorded by the electrodes is transmitted to the robot, the Khepera, made by K-Team S.A, which serves as the body of the cultured networks. It moves under the command of neural activity that is being transmitted to it, and information from the robot's sensors is sent back to the cultured net in the form of electrical stimuli.

Central to the experiments is Potter's belief that over time, the team will be able to establish a living network system that learns like the human brain.

Share |      Randall Parker, 2002 December 29 10:06 PM  Cyborg Tech

daniele said at May 18, 2003 1:46 AM:

how is the neuron-web fed? Can the cell reproduce and get ill? Is the whole system temperature-sensitive?

animal said at May 18, 2003 2:33 AM:

Those are neurons!!! they don't reproduce, they just die in 15 days in best conditions!!! (37į, 5% CO2, good medium) That means the robot can work no more than a couple of hours!!!

Dario said at May 18, 2003 2:53 AM:

I prefere neural networks simulation by hw or sw. I find this horrible and thinking that in the future a human brain could be used in such a situation makes me scared. Anyway a mouse brain has evolved to be useful for a rat... Most of the connections and the paths are made to solve biological, physiologic and survival needs... How do a robot may need these?

Hurricane said at May 19, 2003 1:45 AM:

Why yo think their only reason to exist is to do something bad? Seems like you fear everything that's "other".

Are you one of the mad guys who would shoot an alien before it could say a word, because you just shit in your pants?
(Well... typical for americans...;)

Dario said at May 19, 2003 3:07 AM:

I think I have been misunderstood... I don't think this will do something bad... I didn't even say that! The fact that *touches* me is taking a mouse brain and put it into a robot... Think if your brain would be enclosed into an airplane to take place of the control system... what would be of your human istincts? So if you think to a rat, even if you may think that it is an inferior form of life, it may be not so happy of this *nice* experiment. Just this.
Just to be fair, I'm a computer engineer and I study robotics too, and I'm not American, I'm Italian. :)

Perispipo said at May 22, 2003 12:36 AM:

Do they kill a rat to extract the neuron cells?!

Giorgio said at May 22, 2003 8:05 AM:

I think that it's not possible to implement the kind of technology described by the article. Maybe i'm wrong, anyway a more detailed article could be useful to understand better. Bye.

Giorgio said at May 22, 2003 8:05 AM:

I think that it's not possible to implement the kind of technology described by the article. Maybe i'm wrong, anyway a more detailed article could be useful to understand better. Bye.

Jim said at August 13, 2003 10:00 AM:

Seems doable to me. But I imagine all the robot does is humm with random inputs. So electrical stimuli is fed back to the network. So? In order to train the network you need to give it some form of positive re-inforcement to desired fitness levels. How is that accomplished? Feed it? Temperature stimulation? Curious...

Mando said at September 24, 2003 3:21 PM:

Iím kind of skeptical about getting any real from this experiment anytime soon. In my opinion I feel that we donít have enough of an understanding about how the brain works in a biological setting to branch off into what would happen in a machine of this sort. I agree with one of the other comments about how the robot would basically be receiving semi-random input from the ratís brain. I imagine that when the experiment first started it was like sticking a probe to various wires and seeing which ones will spark. One other thing, I am an American.

naveen gupta said at September 15, 2004 2:02 AM:

i am naveen gupta doing btech in electronic and communication in india . i am intrested in making aproject on machine which directly operated from human brain . i wana to know that it is possible to tranfer brain signal ( i.e. nuerons signal) to machine using external hardware(i.e electrode series outside the human brain).please respond me
Thanking you

justyourdave said at May 31, 2005 3:00 PM:

Why does the brain die after only two years?

name said at September 19, 2005 10:25 PM:

fascinating... wondering about the usefullness of the feeedback system employed... and didn't we see a person in some european country doing this sort of thing years ago? of course not with the attached "robot" but at least culturing brain cells in an electrode array... imho the poor rat brain cell culture got gypped... it would be better off as a rat... it looks like another example of going public with "innovative"tm work prematurely in order to secure a patent...

anon said at December 21, 2006 6:53 AM:

hey, I'm actually involved in this (in europre) just to let you know, only the cells are used the connections are lost in the process so there is no knowledge or behaviour copied at the beginning, - neurones reform connections over about a week - simmilar neural culture experiments have been going for years but without the cells being able to influence their input there is no way to see what changes happen to produce behaviour and connections are basically just forming randomly (or it is extremely aftificial) this way you can directly monitor how neurons work together to create behaviour (not possible with current brain scanning technology). - basically if you took a kitten and put it in a completely white room the visual part of its brain would fail to develop - it needs feedback and that is the idea with giving the cell culture the ability to affects its own input - and hopefully learn.

naveen - yes using EEG you can control a cursor on a screen but it takes training and a lot of concentration google 'EEG brain machine interface'

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright ©