I, for one, welcome our new monkey overlords with their thought-controlled robotic henchmen.

Of course, the real point of this research is that if monkey's can do it, so can humans. So the serious question about the future that we all have to face is whether we will go with straight-up thought-controlled robotic henchmen, or whether we will develop a human-monkeybrain interface whereby what we think will be carried out by the monkeys, and what the monkeys think (as instructed by us) will by carried out by the robots. I personally prefer this model, in that I get not only henchman, but a tiered reporting structure as well.

But that's just me

Also, if it matters, this research might have some kind of obscure side benefits for victims of paralysis:

In a major step toward helping victims of paralysis walk again, researchers at Duke University Medical Center today announced that they had proved monkeys can use their brainpower to control the walking patterns of robots.

The Duke researchers, working with the Computational Brain Project of the Japan Science and Technology Agency, implanted Idoya, a rhesus monkey, with electrodes that gathered signals from her brain's motor and sensory cortex cells as she ambled along on a specially built child-size treadmill. The electrodes recorded the cells' responses as the monkey walked on the treadmill at different speeds; simultaneously, sensors on Idoya's legs tracked their patterns of movement. The information was transmitted in real time from their lab in Durham, N.C., to control the commands of a five-foot-tall humanoid robot (see video here) in Kyoto, Japan.

That part all seems a little far-out to me. But who knows? Maybe the human interest angle will help them keep the work funded.

Via GeekPress.

Posted by Phil Bowermaster at 06:28 AM | Permalink | Comments (2)

How cool is this?

Can cyborg moths bring down terrorists?

A moth which has a computer chip implanted in it while in the cocoon will enable soldiers to spy on insurgents, the US military hopes.

...

At some point in the not too distant future, a moth will take flight in the hills of northern Pakistan, and flap towards a suspected terrorist training camp.

But this will be no ordinary moth.

Inside it will be a computer chip that was implanted when the creature was still a pupa, in the cocoon, meaning that the moth’s entire nervous system can be controlled remotely.

...

“This is going to happen," said [Rod Brooks, director of the computer science and artificial intelligence lab at Massachusetts Institute of Technology (MIT)], "It’s not science like developing the nuclear bomb, which costs billions of dollars. It can be done relatively cheaply.”

This is the power, literally, to be a fly on the wall. If we can send one of these, we can send a swarm. The next step is to farm the data processing to AI's. The AI could wade through thousands of hours of useless junk to find important conversations.

Posted by Stephen Gordon at 03:47 PM | Permalink | Comments (7)

Thomas DeMarse of the biomedical engineering department at the University of Florida has developed a "living computational device" from 25,000 neurons extracted from a rat embryo.

Then he taught it to fly a jet fighter. The F-22 to be precise.

The 25,000 neurons were suspended in a specialized liquid to keep them alive and then laid across a grid of 60 electrodes in a small glass dish.

Under the microscope they looked at first like grains of sand, but soon the cells begin to connect to form what scientists are calling a "live computation device" (a brain). The electrodes measure and stimulate neural activity in the network, allowing researchers to study how the brain processes, transforms and stores information.

In the most striking experiment, the brain was linked to the jet simulator. Manipulated by the electrodes and a desktop computer, it was taught to control the flight path, even in mock hurricane-strength winds.

"When we first hooked them up, the plane 'crashed' all the time," Dr DeMarse said. "But over time, the neural network slowly adapts as the brain learns to control the pitch and roll of the aircraft. After a while, it produces a nice straight and level trajectory."

The article doesn't say, but DeMarse must have found a way to reward the brain for flying straight (or punish it for crashing) using hormones like serotonin. Otherwise, why would this brain-in-a-dish prefer level flight to crashing?

The implications are profound. DeMarse' first goal is to study brain function. Until this development, scientists were only able to study a few neurons in a petri dish. Now DeMarse can observe how these neurons work together to compute. Obviously this is important brain research, but it could also be very important computer research. It could also be important to researchers interested in learning how to get a brain to directly communicate with a computer.

Individual neurons are slow by comparison to transistors, but a brain is superior to a contemporary computer in many ways - pattern recognition, redundant fail proofing (the loss of a few neurons doesn't lead to a crash), self-organizing, and after a crash (a stroke) it can rewire itself. This could lead us to develop computers that are more like a brain.

In the meantime it might lead to hybrids - computers with electronic and biological components.

It could also be another route to greater-than-human intelligence. If this brain-in-a-dish is possible, why couldn't this, ultimately, be ramped up to a 20 pound brain? Such a brain would not be limited by a size that is practical to be carried around in a human skull. Nor would it have to be concerned with the "mundane" tasks of managing a body.

Posted by Stephen Gordon at 04:33 AM | Permalink | Comments (3)

Last August I wrote briefly about a small company by the name of Correlogic. The Wheeling Intelligencer brought me up to date with an article published last week. Correlogic has developed a method to glean information from the data generated from a mass spectrometer.

A mass spectrometer is a machine that can ionize the molecules in a sample of blood serum and then propel those molecules down a tube about a meter in length. The machine will then measure the number of ions that hit the end of the tube during any given moment.

This process will effectively separate the molecules by mass. It is then possible, said [Ben Hitt of Correlogic] to determine whether there is, say, a relatively large amount of a molecule with a molecular weight of 500 versus one with a molecular weight of 510. And in proteomics, the "presumption is that the mass spectra represents proteins in the serum"…

A single blood sample can return as many as 180,000 data points. The trick is to mine that data for information.

"It's almost revolutionary," said Hitt, "Given time to develop the test, we ought to be able to have some one go see their physician, take a very small amount of blood, send it to the lab, get a mass spectrum of that blood" and then determine if the patient has a number of possible cancers.

He said it could be possible to use a single sample to test for lung cancer, colon cancer and liver cancer, as well as ovarian and breast cancer in women and prostate and testicular cancer in men..

If a simple blood test was developed that could detect multiple cancers, then the test could become routine. Early detection would allow early treatment and greatly improved survival rates.

Ultimately this technology could be placed into an implanted medical device. This would be a simpler system than what Phil and I have thought about recently (here and here). Instead of a system that would need to send nanobots out into the body to correct damage, this system could simply take blood samples and test them for cancer and other problems (high LDL cholesterol, heart damage, intoxication, blood-sugar, hormone imbalances, etc.). It could then inform you (or your doctor) via cell phone or PDA if there is a problem.

Perhaps this could even be an added function to medical devices like the Cardioverter Defibrillator.

Posted by Stephen Gordon at 02:03 AM | Permalink | Comments (2)

A team led by scientists at Sangamo Biosciences in Richmond, California has reported in the journal Nature that they have developed a sophisticated new process for editing DNA without bombarding the genome with foreign genetic material.

They treated the cells in test tubes with the company’s proprietary type of “zinc finger nucleases” (ZFNs)... ZFNs are proteins made up of “fingers” of around 30 amino acids and stabilised by a zinc atom. Each finger binds to a specific combination of DNA bases and is attached to a DNA-cutting enzyme called a nuclease.

By using different combinations of amino acids, they can be designed to latch on to DNA at exactly the place where the mutated gene lies and cut it. This triggers the body’s natural repair process, called homologous recombination, which corrects the gene where the DNA was cut, The researchers provided the cells with a copy of the correct gene as a template.

This could be the beginning of a huge step forward. Previous forms of genetic therapy often caused cancer. Scientists would bombard the genome with the desired information and hope it stuck in the right place. It's like blindly lobbing paint balloons at a road sign hoping to cover graffiti without obscuring the speed limit. Sometimes it worked, often it didn't.

This new method is more like word processing. These scientists are hopeful that this advance will yield useful therapies for single gene mutations like that which causes sickle cell anemia or the "bubble boy" disease.

This new method might ultimately have applications beyond treating genetic diseases. Recently Phil and I have been speculating about the possibility of biological cyborgs. These people would remain largely biological, but would have an implanted processor for directing the work of biological nanobots. These "nanobots" might be our own cells redirected as the processor sees fit - cleaning up arteries, compensating for poor eating habits, etc. Obviously such a system would need a sophisticated method for communicating with cells. It would have to speak the language of cells.

A DNA "word processor" might fit the bill.

UPDATE: USAToday has much more. Via KurzweilAI.

Posted by Stephen Gordon at 01:45 PM | Permalink | Comments (3)

My friend Paul has been one for years. As has Stephen's father-in-law. And over the weekend, the Blogfather announced that his wife would be joining the ranks of the early-stage human/machine hybrids.

Glenn comments on the cyborgization of America:

Soon, probably within a decade or two, we'll see such devices becoming common, and multipurpose, and -- most importantly -- aimed at people who don't have anything in particular wrong with them. Perhaps a 'body computer?' It could measure heart rate, blood chemistry, diet and exercise levels, etc., and export its data to outside devices so that the owner, or a physician, could monitor the owner's health. Perhaps it could take preemptive action, releasing clotbusting drugs at the onset of a heart attack or stroke, or steroids in the event of an allergy attack, providing on-the-spot first aid for many serious problems. Still more advanced versions could fine-tune things in a variety of ways, until we gradually reach the stage in which our bodies are pervaded with nanodevices that maintain health and repair damage without our even thinking about them.

I'd like one of those now, especially if it could also treat migraine headaches. They're working on it! They're already working on vagus-nerve stimulation for epilepsy and depression, and even neural stimulation implants that promote female orgasms. (What, nothing for us guys?*) Since these devices are based on two things -- electronics and biological knowledge -- that are improving by leaps and bounds, we're likely to see a lot more of them, and we're likely to see them become cheap enough, and capable enough, and reliable enough that they'll attain widespread use. Which I favor, though not everyone will agree.

As we noted yesterday, developments on the biotech/nanotech front promise to make our integration of machine components as unobtrusive as possible. But yesterday's development raises an interesting question: what do you call a cybernetic organism whose cybernetic components are organic?

Orgcyborgs comes to mind, but maybe its a little clumsy. Perhaps we should reserve the term "cyborg" just for the organic cyborgs, and use the word borg to refer to the old-school mechanical cyborgs.

I'm just free-forming, here. Just kind of opening it up for discussion.

*Well, right, since guys wouldn't benefit from that at all. What guys need is something really useful -- maybe something to make us, I don't know, more regular. No, not punctual, I mean -- criminy, just forget it.

Posted by Phil Bowermaster at 08:47 AM | Permalink | Comments (4)