Stem cells mimic human brain

'Mini-brains' help researchers to study neurological diseases in living human tissue.

With the right mix of nutrients and a little bit of coaxing, human stem cells derived from skin can assemble spontaneously into brain-like chunks of tissue. Researchers provide the first description and application of these ‘mini-brains’ today in Nature1.

“It’s a seminal study to making a brain in a dish,” says Clive Svendsen, a neurobiologist at the University of California, Los Angeles, who was not involved in the study. “That’s phenomenal.” A fully formed artificial brain might still be years away, he notes, but the pea-sized neural clumps developed in this work could prove useful for researching human neurological diseases.

Researchers have previously used human stem cells to grow structures resembling the eye2 and even tissue layers similar to the brain's cortex3. But in the latest advance, scientists developed bigger and more complex neural-tissue clumps by first growing the stem cells on a synthetic gel that resembled natural connective tissues found in the brain and elsewhere in the body. Then, they plopped the nascent clumps into a spinning bath to infuse the tissue with nutrients and oxygen.

“The big surprise was that it worked,” says study co-author Juergen Knoblich, a developmental biologist at the Institute of Molecular Biotechnology in Vienna. The blobs grew to resemble the brains of fetuses in the ninth week of development.

Under a microscope, researchers saw discrete brain regions that seemed to interact with one another. But the overall arrangement of the different proto-brain areas varied randomly across tissue samples — amounting to no recognizable physiological structure...

But we still have the problem that we don’t know what is actually happening in the meat space.

Neural nets are very interesting, but I have read that the major problem is understandability of the coding in neural nets. From an engineering standpoint, tweaking a neural net is not so easy, or even possible in a reliable way, since the data is “learned”, not preset. I am not talking about the code emulating a neural net, but the weights and values that result from training a neural net.

Ie., the only way to tweak a neural net, is really to force more learning into it.

It’s not like compiled C-code or interpreted languages like Java running on a telephone switch, where you can halt everything, change a variable, and then let things proceed without taking down the entire system.

Modifying the weight or value of a node in a neural net would have unpredictable results in most cases except the simplest example.

On the other hand, something like a human brain, without the nasty moral restrictions, could be a quite capable control platform for weapons, etc. Just don’t complain when it decides it hates its masters and attempts to wipe them out and anyone associated with them!

Of course, it begs a moral question: if the neural net is based on human dna, whose to say we haven’t instantiated a human soul to go along with it, but we force it to our purposes? I’d rather keep the neural nets in hardware myself.

We still don’t know how the brain works well enough to be able to reliably control such a construct, either.

So, in the end, I don’t like the idea. If we had successfully done full successful simulations of the human brain and gotten a reliable artificial intelligence out on the other end, then perhaps I’d be more comfortable dropping back to meat for neural computing. For now, meat brains are still the best neural computational constructs known to man, and if we understood it and its relationship to humankind, and there weren’t any moral issues, I’d be OK with going that direction till the silicon (or diamond or grapheme) becomes available that can emulate it faster and more reliably.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.