Connect with us


Developing Artificial Neurons That Can Mimic the Human Brain



the human brain has incredible abilities. With just two bananas a day, you can continue to perform complex tasks with energy. This ultra-high efficiency of the brain is due to neurons, the basic units that make up neural tissue.
Neurons have membranes with nano-sized holes called ion channels, which open and close depending on the stimulus received. the flow of ions generated in this process generates an electric current, which emits an action potential, an important signal that allows neurons to communicate with each other.

All of this can also be done with artificial intelligence (AI). However, artificial intelligence must consume tens of thousands of times more energy than the human brain. So scientists think that if computers can be made like the human brain, they will need much less energy.
One of the ways scientists try to mimic the brain’s biological system is to harness the power of ions. The power of ions are the charged particles that the brain depends on to produce electricity.
However, a study was recently published that has built prototypes of artificial neurons that can generate the same types of electrical signals that neurons in the human brain use to transmit information. Created by researchers such as the French National Center for Scientific Research (CNRS) and the Paris School of Sciences and Humanities, these artificial neurons use ions like our brains, not electrons.

ions in water form a worm-like structure,

Artificial neurons

Neurons change the ionic parametricity of the cell membrane through electrical stimulation or neurotransmitters, in which the reverse phenomenon of the + pole and the reverse voltage occurs. This is called activity and it is a medium that transmits all the reactions that occur through the brain, such as smell, sound and movement.
Neurons begin to accept more cations that are attracted to anions within cells to generate activity. At this time, the voltage across the cell membrane opens the cell gate called “voltage-gated ion channels” and increases the charge much more until the cell returns to normal immediately after reaching its peak. These signals are transmitted to other cells, allowing the information to move the brain.
The researchers modeled a very thin layer of water between the graphene sheets to mimic voltage-gated ion channels. In computer simulations, the researchers added an electric field to the model and found that the ions in the water formed a worm-like structure.

To mimic the behavior of neurons, the researchers ran a simulation that connected two channels with another component. The results showed that the model generated peaks in electrical activities, such as activity, and recalled the consistent properties of ions when they conducted more electricity and in the two different states they performed.
The memories of these ions’ previous state actually lasted for several milliseconds (one thousandth of a second), roughly the same time it took for neurons to generate activity and return to a resting state. This is usually quite a long time for ions that operate in nanoseconds (one billionth of a second) or less.
Our brain uses the opening and closing of ion channels to produce these kinds of memories. The findings were published in the international journal Science on August 6.

it helps us understand how the brain processes information.

The new model is a version of an electronic component called a memlister, with unique properties that maintain historical information. A memristor is a memory and register synthesizer that both remembers the previous state.
However, scientists are interested in this new study of fluid memlisters because conventional memlisters are solid, since they do not use liquids like the brain.
“Practical computers that function like brains are still a long way off, but this study may help us better understand how the brain processes information and develop new theories of computation, like the brain,” said Gina Adam, professor at George Washington University. in the USA, who did not participate in the study.
With the help of digital and theoretical tools, French researchers have discovered a way to reproduce physical mechanisms that assemble special groups to release activity. In other words, he has discovered the ability to transmit information from one artificial neuron to another.
The pioneering work is being carried out in France in collaboration with scientists at the University of Manchester in the UK, where researchers are working on the next phase of research that experimentally proves that the new system can run basic learning algorithms.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *