The brain’s connection to the Internet and real-time data interaction are no longer just a fantasy. In fact, the brain’s function depends on neuronal circuits, and the parts of contact between neurons are synapses that play a key role in combining information transfer with memory storage and processing. Electronics has made important progress in simulating neurons and synapses, as well as brain-computer interfaces.
Recently, researchers in the UK, Italy and Switzerland have created a hybrid neural network that uses the Internet to transfer signals from biological neurons to artificial neurons. The scientific journal Nature Scientific Reports published the team called Memristive synapses brain connect and silicon spiking neurons (memory synapsations connect ingest neurons” brain and silicon spike neurons papers.
Biological neurons and artificial neurons via Internet connection
This paper focuses on the memory connection between the brain and silicon spike neurons that simulate the transmission of real synapses and the plasticity of the brain and the silicon spike.
It is worth noting that the study was conducted by researchers from three countries:
The University of Southampton in the United Kingdom has developed nano-scale mematas to mimic synapses;
The University of Padua, Italy, has trained biological (mouse) neurons;
The University of Zurich in Switzerland, in collaboration with the Federal Institute of Technology in Zurich, created artificial neurons on silicon microchips that enable artificial and biological neurons to communicate in real-time in both directions by using a hybrid neural network created by amnesors.
Specifically, the researchers used a memphoresistor paired with a titanium oxide microelectrode of metal film to connect silicon neurons to neurons in the hippocampus of mice (shown below), demonstrating a three-neuron brain-silicon network in which memry resists synapses undergo long-term enhancement or suppression driven by the rate of neuron discharge.
Based on the above description, it is not difficult to see that a key element of this study is the nano-scale memritis.
Typically, spike processing is managed by a digital von Neumann-based hardware-run statistical algorithm. However, depending on the peak signal and processing strategy of near-biological organisms, neuromorphic electronicdevices and structures can be said to be relatively better computing choices.
On this basis, the researchers found that nanoscale memrist mimics the plasticity of synapses, and that it is the titanium connection that mimics the piezoelectric transfer and plasticity processes of brain synapses.
As the figure below shows, the system is actually a one-way circuit – the titanium memryresistor MR1 stores synaptic weights as a blocking state, and the ultra-thin capacitor microelectrode CME transmits stimulation to biological neurons and adjusts them by memory weights. THE BN SPIKE IS RECORDED BY THE DIAPHRAGM CLAMP MICROELECTRODE, THEN PROCESSED BY THE MEMRESISTOR MR2 AND INJECTED BY CURRENT, WHICH IS EVENTUALLY TRANSMITTED TO THE SECOND SILICON NEURON ANPOST.
Another important element is artificial neurons. Artificial neurons are known to be arranged on silicon microchips. Similar to biological neurons, artificial neurons can send, receive, and receive readable and measured signals.
This is the first time in the world that the cross-border connection between artificial and biological neurons has been achieved in this way. The researchers also say they hope the study will be the beginning of the neuroelectronic Internet.
In response to this breakthrough, Themis Prodromakis, Professor of Nanotechnology and Director of the Centre for Electronic Frontiers at the University of Southampton, said:
We are very excited about the new developments. On the one hand, it lays the foundation for a new scenario that has never been seen before in natural evolution, in which biological and artificial neurons connect and communicate in global networks, and on the other, it offers new prospects for neurorepair technology, paving the way for research into replacing parts of the brain with AI chips for brain dysfunction.
Musk: Building a Third Brain
In fact, the “neuroelectronic Internet” is also a brain hole for Elon Musk, Tesla’s CEO, SpaceX CEO and CTO, and Sun City Chairman.
It is well known that the human brain is divided into two layers:
Peripheral systems: used to control emotions, long-term memory and behavior, etc.
Cerebral cortex: Dealing with complex thoughts, reasoning, and long-term planning.
In response, Musk argues, our brains actually have a third layer — digital devices such as computers and mobile phones that humans cannous. So he put forward the concept of “brain-computer interface”, which connects the brain to a computer, perfects the third layer of the human brain, and complements the other two layers.
As a doer, Musk founded Neuralink, a brain-computer interface company, in 2017.
Neuralink was founded to help people with severe brain injuries (e.g. stroke, cancer, congenital diseases, etc.). After implanting the device in the brain, the patient can control the phone or computer on their own.
At the July 16, 2019 launch of Neuralink, Musk revealed that a monkey involved in the experiment was able to control the computer with its brain, and introduced the main technologies that have been introduced (:
Thread, which is only 4-6 m wide and can transmit large amounts of data;
Neurosurgical robots that can insert six Threads containing 192 electrodes into the brain per minute;
Custom chip, better read, clean, amplify signals from the brain;
The N1 sensor, which can be embedded in the human body and transmits data wirelessly, reads fewer neurons than the current wired system – four sensors will be implanted in the human brain (three in the motion area and one in the sensor area), which are wirelessly connected to an external device mounted behind the human ear and controlled by the iPhone application.
In fact, Neuralink has begun experiments on mice, and if the platform is stable, Thread will be implanted through robotic surgery. Of course, this process requires review by the U.S. Food and Drug Administration.
But it’s not just Neuralink that’s working on brain-computer interfaces, but Facebook’s Building 8 division, universities such as mIT, and research institutions are doing just that.
Given the engineering barriers, including wireless transmission, the level of human awareness of the brain, public concerns about the safety of such technologies, and so on, it is not easy to achieve brain-machine connectivity.
As Musk himself said:
Science and technology will not make their own progress, scientific and technological progress needs to pay, need ambitious goals and a lot of hard work, need to have great dreams to inspire people to create.