Scientists at Stanford University have developed a new microchip modeled on the human brain called "Neurogrid", which works 9,000 times faster and more efficiently than computers.
Neurogrid gives scientists the chance to better understand the human brain and create new forms of computing patterns after brain circuits, according to Yahoo! News.
"From a pure energy perspective, the brain is hard to match," said Kwabena Boahen, bioengineer at Stanford and leader of the development of Neurogrid.
Boahen added that PCs are not only slower than the human brain, but they take 40,000 times more power to run, Mumbai Mirror reported.
The Neurogrid circuit board includes 16 custom-designed "Neuroscore" chips, which together simulated 1 million neurons and billions of synaptic connections. The research team hoped to give certain synapses the ability to share hardware circuits during their experiments.
Boahen is reportedly working with other scientists from Stanford to give the chip the ability to control prosthetic limbs for paralyzed people, Mumbai Mirror reported.
"Right now, you have to know how the brain works to program one of these," Boahen said. "We want to create a neuro compiler so that you would not need to know anything about synapses and neurons to be able to use one of these."
The chip would translate brain signals into movements of the limb, and the brain would not be overheated, Yahoo! News reported. Neurogrid has another potential use in controlling humanoid robots.
Other organizations have made efforts to use computers to model the human brain, including The European Union, which has started the Human Brain Project to simulate a mind on a computer, and the U.S. Brain Research through Advanced Innovative Neurotechnologies (BRAIN) Initiative, which will develop new technology to measure brain circuit and neuron activity. Boehan said that these efforts, along with others, were able to make different balances in performance and capability. But ultimately, Neurogrid was the most effective.
The prototype chip cost almost $40,000 to make, Yahoo! News reported.
Researchers are looking to use techniques in modern manufacturing to cut the cost 100-fold.
