Brains on a Chip

jhouserizer writes "The New Scientist magazine has an article reporting on new advances in keeping brain tissue alive (and working) on a "chip", with electrodes that can monitor the brain activity. Could this be a step toward computers that can learn as humans learn?"

Comments

The future?

[FosterSJC]

This can't really be the future of computing can it? I mean, we all are aware of the biggest difference between computers and thu human brain. Humans have great pattern recognition, while computers have great calculating/processing powers. Slicing pieces of brain and attaching them to chips hardly seems likely to enhance either the brain's computational ability or the chip's recognizing abilities. If anything, this is a step forward in facilitating communications between man and machine. I could see uses in reversing paralysis, but thought-upgrades or what have you are a long way off.

Re:The future?

[Myco]

You're being too literal here -- assuming that the end result will look just like a particular intermediate step. Nobody's saying we're going to build computers out of brain tissue. Well, maybe somebody is but that's not the point.

In order to potentially imitate the human brain, we still need to learn a lot more about it. Since there's no manpage for the brain, it's a black-box problem and we have to reverse engineer it by trying various combinations of inputs and outputs (as well as analyzing the physical structure, of course). This new technique allows us to do so more effectively, hence improves our ability to understand. That's all.

The key to immortality

[Anonymous Coward]

Replace the neurons one by one with an equivalent electrical doodads until the whole kaboodle is machine. Then you'll be effectively immortal, unless you skimped and purchased from the Shack or something.

A step towards learning computers? Not quite...

[stefanlasiewski]

"The New Scientist magazine has an article reporting on new advances in keeping brain tissue alive (and working) on a "chip", with electrodes that can monitor the brain activity. Could this be a step toward computers that can learn as humans learn?"

Not quite, The scientists are really just developing a a way for keeping your brain alive while watching TELEVISION!

Hocum or over protective of company IP?

[Deanasc]

I've been trying to find something, anything on this breakthrough besides a press release. It sounds like cutting edge neuroscience but I haven't found anything peer-reviewed in any of the journals. I leave it up to you to draw your own conclusions.

Re:Hocum or over protective of company IP?

[Alik]

That's because it's not a breakthrough, it's commercialization of stuff that's been done in the academic labs for a couple years. Amazingly, it *does* take time to go from neat idea to marketable product. Electrode arrays can be a real bitch to fabricate.

See:

Biosensors & Bioelectronics 16:527-33
Brain Research Protocols 2(4):229-42
Journal of Neuroscience Methods 101(1):31-42
Journal of Neuroscience Methods 114(2):135-48

Plus, as the other poster mentioned, Jerry Pine's work. However, AFAIK, Pine's no longer working on that project, having found other interests. Pity, because the neurowells (as noted by Peter Fromherz in the New Scientist article) give the kind of single-unit interface that might be very desirable. I know that my group is planning to use similar approaches.

Future of computers learning ...

[rotwhylr]

The LAST thing we should want is a computer that learns like a human. With my luck I'll probably buy one that complains about homework and gets F's before dropping out of school and mooching off me for life.