Scientists build a brain

[ughaibu]

Has anyone heard about this? There's not a great deal of information here: ABC News: Scientist Build a 'Brain' From Rat Cells

[CraigD]

Has anyone heard about this? There's not a great deal of information here: ABC News: Scientist Build a 'Brain' From Rat Cells

What an amazing exercise in nerve-computer interfacing! I was aware that this field was very advanced, but was frankly astonished at DeMarse and colleagues accomplishment of keeping a airplane simulator flying straight and level using a dish of 25,000 rat brain cells with 60 electrodes stuck in it – and, I imagine, a pretty nice bit of interfacing hardware.

 

It helps keep the accomplishment in perspective, however, to note that German engineers solved the same problem in 1945 using completely mechanically linked gyroscopes and pendulums in the guidance system for the V-1 “flying bomb”. Though it seems very complicated, keeping an aircraft - simulated or otherwise – flying straight and level is not all that hard.

 

PS: I’d give no more credence to the claim of Russian scientists to have created a “human level” artificial brain than the article does. Russian science enterprises have a history of making wildly exaggerated to outright fraudulent claims like this.

[ughaibu]

An interesting point, for me, is the disparity in times that it took the "brain" to learn the task, "usually, within 10 to 15 minutes". It's difficult to assign a lexical value to the time that it should take an artificial brain to repeatedly learn the same task, but the variation, even if only within usual results, strikes me as, at least, significant.

That the article states "25,000 neurons from the brain of a rat" suggests that they used a portion of a previously extant brain, compared with (for example) the scenario of taking a single cell from each of 25,000 individual rats or having neurons that could be randomly re-wired before each test. All in all, I'd like access to a more detailed report.

[Qfwfq]

In substance they interfaced a small bit of a rat brain and observed it learning to fly.

[InfiniteNow]

All in all, I'd like access to a more detailed report.

You may need to sign up for an account for full access, but you can at least see the abstract and maybe take the reference to a university library to find the articles themselves:

 

Welcome to IEEE Xplore 2.0: Computation within cultured neural networks

In this paper we present three related areas of research we are pursuing to study neural computation in vitro. Rat cortical neurons cultured on 60 channel multielectrode array (MEA) allow the researcher to measure from and stimulate sixty different sites across a small population of neurons grown in vitro. Using this system we can send stimulation patterns into the network and study how these living neural networks compute by measuring its outputs. Our first series of studies uses chaotic control techniques to study the dynamics and potentially control the behavior of cortical network. At the same time, we are beginning to apply a model of computation called the liquid state machine or LSM model developed by Wolfgang Maass to provide a firm mathematical framework from which to proceed with our investigations. Each of these components is integrated into a third area investigating the role of computation and feedback using a real-time sensory-motor feedback robotic flight system.

 

 

Wiley InterScience :: Session Cookies

We have developed a novel method for on-chip cultivation of neural cells in a flexible agarose-microchamber array on a glass slide. The agarose microchamber is a micrometer-order cavity constructed on the surface of an agarose layer by molding a 50-µm-high square/circular micro-cast of thick SU-8 photoresist. In addition, the shape of the agarose microchamber was rearranged by using the photothermal etching method, in which we used an infrared (1064-nm) focused laser beam as the heat source to melt and remove a portion of agarose gel at the heating spot. With the photothermal etching method, we can also manufacture narrow tunnel-shaped channels between microchambers. When nerve cells were cultured on the agar-microchamber array chip, the nerve cells in two adjacent microchambers connected through the photothermal-etched channel after 48 hours of cultivation. Those results suggest the potential of an agarose-microchamber array integrated with the photothermal etching method for the next stage of single cell cultivation and measurement of nerve cells, such as real-time control of cell interactions during cultivation. © 2003 Wiley Periodicals, Inc. Electr Eng Jpn, 146(2): 37-42, 2004; Published online in Wiley InterScience (http://www.interscience.wiley.com). DOI 10.1002/eej.10215

 

 

 

Here are some other search results of related work on Google Scholar:

 

DeMarse nerve cell - Google Scholar

 

 

Cheers. :cup:

[ughaibu]

Thanks for the replies, "a small population of neurons grown in vitro" appears to contradict Qfwfq. I'm not wildly bothered except that this would seem to be ammo against the determinists. My apologies for not searching further myself, I'm a little upset and I plan to get drunk (nothing to do with this thread).

[InfiniteNow]

No worries, and good point. If this thread made you upset enough to get drunk, I'd suggest you needed counseling. I am reassured to know this is not the case, and you get drunk for other more salient reasons like the rest of us. ;)

 

 

Neurons can be cultured much like bacteria strains. :cup:

 

 

 

Neuron networks

Neuron cultures in vitro are advantageous, compared to in vivo systems or brain slices, because of the easiest access they provide for numerous techniques and characterization methods; on the other hand, these cultures suffer a major lack when compared to those two systems, namely the fact that connexions between neurons are randomly established.

 

The figure here shows a neuron culture in vitro: the cellular bodies (white) are randomly distributed on the surface as are, also, the dendrites and axones. Different research groups had proposed experimental methods aiming at controlling the geometry of neuron networks but they all failed at preserving the normal activity of the neurons.

 

 

The new method based on a specific surface treatment elaborated at L.D.F.C. allows precisely:

• the control of the geometry and connectivity of the networks over a 1-month period;
  
• the preservation of intact neuronal functionalities.

[ughaibu]

Thanks again. The main question, re the time to learn and questioning deteriminism, concerns the state of the "brain". I guess there're a bunch of questions about apoptosis and reproduction, that would influence the brain's state. If one could view results over time, it would help. It's an extraordinary story, the words in which it's couched wouldn't satisfy me for over the fence gossip.

[ughaibu]

Another small point: there wasn't necessarily a contradiction between my quote and Qfwfq's post, the actual brain could've been grown from a few mother cells. But, in such a case I still assume the wiring was consistent during all flight tests.

My mistake.