Science

Synthetic brain comes a step closer with creation of artificial synapse

Synthetic brain comes a step closer with creation of artificial synapse
A microscopic view of the carbon nanotube field-effect transistor used in the fabricated synapse (Images courtesy USC)
A microscopic view of the carbon nanotube field-effect transistor used in the fabricated synapse (Images courtesy USC)
View 2 Images
A microscopic view of the carbon nanotube field-effect transistor used in the fabricated synapse (Images courtesy USC)
1/2
A microscopic view of the carbon nanotube field-effect transistor used in the fabricated synapse (Images courtesy USC)
A schematic of USC's synapse circuit
2/2
A schematic of USC's synapse circuit

It's probably still going to be a while before autonomous, self-aware androids are wandering amongst us. That scenario has come a little closer to reality, however, with researchers from the University of Southern California having created a functioning synapse circuit using carbon nanotubes. An artificial version of the connections that allow electrical impulses to pass between neurons in our brains, the circuit could someday be one component of a synthetic brain.

The USC Viterbi School of Engineering team was led by Professors Alice Parker and Chongwu Zhou. Parker has been looking into the feasibility of creating a synthetic brain for the past five years, as part of the BioRC Biomimetic Real-Time Cortex project.

The circuit itself consists of highly-aligned carbon nanotubes that are grown on a quartz wafer, then transferred to a silicon substrate. It mimics an actual synapse insofar as the waveforms that are sent to it, and then successfully output from it, resemble biological waveforms in shape, relative amplitudes and durations. In other words, it can take in the type of impulses generated by real neurons, and send them on in a form that could be further processed by other neurons - it can even vary the strength of those impulses, much as real synapses do in a biological process that is thought to facilitate learning.

A schematic of USC's synapse circuit
A schematic of USC's synapse circuit

"This is a necessary first step in the process," said Parker. "We wanted to answer the question: Can you build a circuit that would act like a neuron? The next step is even more complex. How can we build structures out of these circuits that mimic the neuron, and eventually the function of the brain?"

While Parker stated that synthetic brains are probably still decades away, she believes that the technology could ultimately be used in prosthetic nanotechnology for treating traumatic brain injuries, or for designing intelligent systems that could be used to make cars safer, among other applications.

10 comments
10 comments
ralph.dratman
I\'m very glad to know that \"It\'s probably still going to be a while before autonomous, self-aware androids are wandering amongst us.\" In fact I have no desire for any of those at all.
I used to imagine that having autonomous toys would be interesting, but then I tried out a simple sphere device that would wander around on the floor. It really was autonomous, and it abruptly dawned on me that even a beloved little dog can be quite annoying at times, not to mention a child, no matter how wonderful. Instantly I realized that I did not want any autonomous anything in my house unless there were going to be a LOT of fun, help or some other benefits as a result.
I like my personal space. Dogs and children and spouses have wonderful positive effects on our lives, at least some of the time. An android? I doubt it.
No love, no comfort, therefore no android!
ichidos3
Well I\'m pretty sure nobody intends to make androids so they can simply take up your personal space Ralph... It would be great if we had androids to take the menial jobs we are already outsourcing. I would love to have them around, but really we\'re not ready for them as a species. Once we do create intelligent machines and turn one on, is it unethical to turn it off? I can\'t wait to hear what all the various religious leaders say when we build an intelligent android, or better yet what the android says about religion. Will people protest if an engineer gets almost done building one and then stops because it\'s just not the right time? I mean, isn\'t robobortion wrong? I wonder if the robots that happen to be sold to a very wealthy family will feel a false sense of entitlement over those who were bought by a family of slightly lower socioeconomic status. We\'re not ready for androids, they would have far too much to teach us.
Drew Ryan
Watch the Transcendent Man. This is only the beginning. Self aware robotics and human robot integration is coming soon.
Paul Mason
While interesting, this isn\'t critically unusual. The value of this is likely more to do with size and a reduction in the number of transistors needed to perform the function. A few questions I have that are rather critical are: What\'s the size of the synapse, what\'s it\'s heat output, and how many transistors does it use (and does it use the new memristor tech IBM developed)? If it\'s very small and very low heat due to few transistor components then synapses can be packed tightly together and connection length can be kept short, making 3d constructs more viable (a necessity for handling the extremely high neuron/synapse count in a brain).
Beyond that, this isn\'t as exciting as it may seem. We\'ve been able to mimic neuron behavior for decades. The first artificial neuron was proposed in 1943, predating mainframes (!!). Essentially a neuron is a junction that takes an ionic impulse (the so-called electrical impulse that is in actuality a \'wave\' of opening ion gates along the axon) and chemically sums the impulses it receives both temporally (several impulses over time) and spatially (several impulses from different neurons at the same time) until a threshold value is reached. This process, as we know it now, is actually quite simple and can be expressed as a rather simple sigma equation.
What we don\'t know, and therefore this device doesn\'t do, is fully understand the role of the many disparate neuro-transmitters that our brains use. There are two primary types: GABA and glutamate which comprise over 90% of what our brains use, but there are also a myriad of other neurotransmitters that serve numerous functions (some of which can be mimicked just be adding modulation circuitry to the synapse). We don\'t yet know all the neurotransmitters in the brain, and we don\'t yet know why the brain uses many of them.
We also don\'t know how the speed of the ionic \'wave\' down the axon effects brain function (it varies on the presence of myelin sheaths and axon width), the role neural cross-talk plays in intercommunication, or how neurons form new synapses. We don\'t know if the threshold and spike value variations in neurons is a biologic byproduct or critical to brain function, and we don\'t know if the effects of any of these variables need to be simulated within an artificial neural network to generate cognizance.
Daniel Gary
I just wonder how much closer this gets the man-machine interface.
webmaster
If those nanotubes can mimic post synaptic potentials and cortical evoked responses, then it can probably grow itself into a neural matrix. Apply thesame harmonics that we have in our heads to a quartz matrix using Chladni\'s law.
Simon Lee
\"Resistance is futile, you will be assimilated!\"
Alice Parker
Love getting comments! Paul, these nanotubes are a billionth of a meter in diameter, so the transistors are quite small. The chance for lower power operation than conventional electronics is certainly possible. This simple synapse is only a single transistor but it would take more transistors to have variability in neurotransmitter release and reuptake, and receptor concentration, like our conventional synapses. And while we implement mechanisms as fast as the neuroscientists report their understanding, there are so many open questions, which makes all of this very exciting and very much a small step towards an enormous goal. We\'re working on memristors, variability and other issues you mention, as well. However, implementing each mechanism or new technology in the laboratory takes a significant investment in time and energy, so achieving synaptic plasticity, for example, the next obvious step, is months away.
Samantha Renault
Immortality?
A few decades or more from now. Imagine technology like this, used to transfer information between the human brain and a computer. Now slice up a small part of your brain and scan it into a computer (like the rat hypothalmus scanned into IBMs blue gene) then connect the computerized brain slice with the rest of your organic brain, and let it reintegrate itself whole again. Repeat the process till your entire organic brain has been scanned into a computer.
Scanning a brain into a computer this way means you don\'t die, you just lose tiny parts of your brain one at a time briefly. If you just destroyed and scanned your brain in one go you\'d be making a digital copy of yourself and killing the original. And brain scanning likely involves brain destruction in the foreseeable future. So what I described instead is a way of scanning your brain into a computer without dieing.
The science fiction of today sometimes becomes the science fact of tomorrow.
Shradhapati Singh
I read some of the comments that we dont want anoroids around us but it came to my mind that study of brain function on artifical or synthetic brains can help in creating cures in case of injury to human brain and possibly improve life of those disabled due to non functioning of nerves or brain.I have seen people getting bedridden and having vegetative life due to paralysis.I hope some cure might become possibilty for the paralytics