Electronic memory may bring bionic brain one step closer
Using a matrix of nano-sized memristors, researchersworking at the Royal Melbourne Institute of Technology (RMIT) and theUniversity of California, Santa Barbara claim to have constructed the world’sfirst electronic memory cell that effectively mimics the analog process of thehuman brain. By storing memories as multiple threads of varying information,rather than a collection of ones and zeroes, scientists believe that thisdevice may prove to be the first step towards creating a completely artificial,bionic brain.
Working at the MicroNano Research Facility of RMIT, the researchersbelieve that the breakthrough not only carries them closer to reproducing keyaspects of the human brain electronically, but could also one day assist inproviding effective treatments for neurological conditions – such asAlzheimer’s and Parkinson’s diseases – by studying such diseases outside thebody using artificial brains. Eventually, even cybernetic implants couldconceivably be developed from this technology.
"This is the closest we have come to creating a brain-like system withmemory that learns and stores analog information and is quick at retrievingthis stored information," said Dr Sharath Sriram, co-leader of the RMITFunctional Materials and Microsystems Research Group."The human brain is anextremely complex analog computer … its evolution is based on its previousexperiences, and up until now this functionality has not been able to beadequately reproduced with digital technology."
The memristors (a contraction of the term "memory resistors") usedin the new device are electronic components whose electrical resistance is notconstant like standard resistors, but is determined by the history of theelectric current that has previously flowed through it. In other words, amemristor "remembers" the direction and magnitude of the electricityapplied to it. It is also non-volatile in that it retains that "memory" whilstturned off and does so until it has another electric charge applied. In thisway, memristors act like brain neurons because the information contained onthem is proportional – not simply "on" or "off" like binary computer data.
"We have now introduced controlled faults or defects in the oxide materialalong with the addition of metallic atoms, which unleashes the full potentialof the 'memristive' effect – where the memory element's behavior is dependenton its past experiences," says Dr. Hussein Nili, lead author of the study."The ability to create highly dense and ultra-fast analog memory cellspaves the way for imitating highly sophisticated biological neuralnetworks."
This current work exploits earlier research at RMIT, where ultra-thin filmoxide material 10,000 times thinner than a human hair was used to createNano-scale, ultra-fast memory components. These memory components now form thebasis of the analog storage device recently developed, and may be the forerunnerto the components required to emulate the complex artificial intelligencenetwork needed to eventually produce a bionic brain.
"This new discovery is significant as it allows the multi-state cell tostore and process information in the very same way that the brain does,"says Dr. Nili. "Think of an old camera which could only take pictures inblack and white. The same analogy applies here, rather than just black andwhite memories we now have memories in full color with shade, light andtexture, it is a major step."
Work by other researchers in this area, suchas that from scientists at Northwestern University, constantly increasesthe promising development of memristors as a possible electronic analog for thememory function of the human brain. Coupled with the new breakthrough from RMITand others, future artificial memory devices may also realize benefits in brainexperimentation for scientists to imitate the human brain so that there is norisk from invasive procedures working on a live animal or human.
"If you could replicate a brain outside the body, it would minimizeethical issues involved in treating and experimenting on the brain which canlead to better understanding of neurological conditions," says Dr. Nili.
The results of this research were recently published in the journal AdvancedFunctional Materials.