Let’s imagine some pictures: people with broken arms play the piano in the concert hall, people with lost legs dance on the stage, autistic children return to normal social and cognitive levels, the elderly gradually get rid of Alzheimer’s disease, insomnia can go to sleep quickly without drug control, and ordinary people and even patients with high paraplegia can type in space Video manipulation and brush shaking through the tiktok manipulation of electronic devices.
Space typing
Some of the above scenarios have become a reality, and some are becoming a reality step by step, thanks to the empowerment of brain computer interface technology.
what is brain computer interface?
Brain computer interface is the interface between brain and machine. It establishes a path that can directly transmit signals between brain and external equipment, so as to realize the information exchange between brain and equipment. Brain computer interface technology is the next generation technology of artificial intelligence. It is a fusion intelligence of artificial intelligence and human intelligence.
In recent years, neuralink, a brain computer interface company invested by Elon Musk, has attracted much attention, which is the focus of Musk’s work after autopilot and space exploration.
In 2019, neuralink released microchip and surgery Siasun Robot&Automation Co.Ltd(300024) ; In 2020, neuralink experimented on animals for the first time. Chips and electrodes were implanted into the heads of three piglets to successfully read EEG signals; In 2021, neuralink announced that it hoped to make paralyzed patients stand up through brain computer interface technology.
According to media reports, neuralink completed round C financing in July this year, raising a total of US $205 million, which is considered to be the largest financing in the field of brain computer interface so far.
It is reported that at present, the companies that have raised more than 200 million US dollars worldwide and used for the research and development of new brain computer technologies, in addition to Musk’s neuralink, there is also brainco brain technology, a Chinese team founded in 2015.
Founded in 2015, brainco is the first Chinese team selected into Harvard innovation lab. In 2019, the Chinese Academy of Sciences released the top 20 list of global AI enterprises, including brainco.
neuralink and brainco represent two technical paths of brain computer interface – invasive and non-invasive.
As the name suggests, invasive surgery should be performed on the brain. Electrodes are placed on the human cerebral cortex and even inserted into the brain to directly collect the electrical signal changes in the human brain.
The non-invasive method is to collect very subtle EEG changes and signal changes in people’s brain without damaging the human brain, such as nuclear magnetic resonance, infrared, EEG and other technologies, and then identify the information in people’s brain.
According to Han Bicheng, founder and CEO of brainco brain technology, in all brain computer interface laboratories, more non-invasive experiments are done than invasive ones, accounting for about 70%; Others are invasive, and only a small part of them are human, and the rest are animal experiments.
the invasive solution is the problem of severe brain diseases , such as Parkinson’s disease, severe paralysis, etc. Its advantage is that it can accurately monitor brain wave signals, which is mostly used to help patients with basic disability to realize prosthetic instrument operation. Its disadvantage is that craniotomy has high risk, small application area and very little information.
the non-invasive solution is a broader brain related disease problem , such as Alzheimer’s disease, sleep disorder, depression, autism, human functional rehabilitation and human next-generation interaction. Its advantage is that it only needs to analyze the surface signals of the large cerebral cortex through relevant equipment, and can directly collect and process signals, Without surgical intervention, it is easier to be applied to wearable devices that can monitor the state of brain waves in real time to realize auxiliary functions, but the difficulty is that the signal is extremely weak.
Yang xiongli, academician of the Chinese Academy of Sciences, believes that there are still severe challenges in brain computer interface technology. Both invasive and non-invasive have many problems. “The invasive type is how to integrate the electrode recording the signal with the organism, how to analyze the useful signal from the organism, how to extract the signal we need from the signal pile mixed with noise, and how to prevent signal interference between different brain regions. There is still a lot of work to be done.”
In this regard, Han Bicheng also said that brain computer interface technology has high barriers, mainly including three directions:
First, the number of neurons in the brain is very large. The human brain has 86 billion to hundreds of billions of neurons. This is a complex group. It is very difficult to study its whole mechanism and its correlation and role.
Second, the human brain is very special, and its plasticity is very strong. A person is moody or moody, because the connectivity of the brain and the enhancement and weakening of the whole function are always changing. How to use algorithms to deal with this real-time change is a very difficult place in brain computer interface technology.
Third, the brain is wrapped in the skull. It is also technically difficult to take the signals in the brain.
These three points make it very difficult to collect high-precision EEG signals and analyze them in the whole field of brain computer interface.
The technical barrier and problem of non-invasive brain computer interface is how to make a breakthrough in materials science, analyze and remove the noise, so as to collect extremely weak human body signs and signals, and then analyze the ever-changing thoughts in the brain, he added.
In order to achieve precise acquisition and portability of EEG, the BrainCo team made an important breakthrough in signal acquisition and analysis. invented the innovative electrode material hydrogel storage electrode, and applied for a patent. The technology has broken through the difficulty of large-scale bioelectrical signal collection. At the same time, it overcomes the noise processing difficulties of portable bioelectricity acquisition equipment and greatly improves the signal acquisition accuracy of micro equipment. In 2018, brainco brain technology began to commercialize the technology.
According to Han Bicheng, as early as more than ten years ago, Europe and the United States invested heavily in brain computer interface and brain science, with early research and development and large investment. At present, many companies can mature and manufacture brain computer interface products. However, China brain computer interface company has just started, and the technology is still in an early stage requiring large-scale investment.
“China has also made a lot of achievements, including the laboratories of Zhejiang University and Tsinghua University, which have recently published a lot of very good scientific research articles. China has a very large demand population, a very large market space, a very good experimental environment and conditions, and very good engineers at the product and Application level. Although we started late, our development speed is actually now It’s faster than the United States. We do a material test in the United States to coordinate factories, materials and environmental protection applications. These processes may take 9 months. In China, almost all supply chains can be found in an instant, which may be completed in two months. But from the national level, the biggest difficulty now is how to find talents with mixed knowledge system, which takes time to accumulate, so we should increase talent training and investment in this field. ”
from the world’s first brain controlled intelligent hand to intelligent bionic leg
On June 12, 2014, at the opening ceremony of the world cup in Brazil, the team of Professor Nicolelis, academic founder in the field of brain computer interface and academician of the French Academy of Sciences, demonstrated the first brain controlled Siasun Robot&Automation Co.Ltd(300024) exoskeleton, which enabled Brazilian paraplegic athlete Juliano Pinto to kick off for the world cup. This basically means the arrival of brain computer interface technology.
In the same year, Han Bicheng became a doctoral student in the Harvard brain science center and began to form a team of brainco. At first, the team did the bottom technology pre research in an underground garage of Harvard and thought about how to apply these technologies.
At that time, an intern from MIT in the team lost his right hand due to a laboratory accident. It was the same student who proposed, “it would be better if we could make a hand that can be controlled by our brain and let it be controlled like my own hand.”
Han Bicheng introduced that although the previous brain computer interface research has done the research on the control of the manipulator, it adopts an invasive scheme. Surgery must be performed on the patient’s head to realize the control of the manipulator. The non intrusive solution does not have any industry benchmarking products at all.
This is not easy. The hand is the most precise organ of the human body. There are 34 different muscles in the arm. It is very difficult to accurately capture the consciousness signal transmitted by each muscle in a non-invasive way and restore it to the original work intention.
but the social significance of this matter is obvious.
According to the data of China Disabled Persons’ Federation, there are 24.72 million disabled persons with physical disabilities in China. However, it is understood that the utilization rate of prosthetics for people with physical disabilities in China is very low. The utilization rate of prosthetics, including various forms, may be less than 1%, mainly because it is difficult to afford the price of hundreds of thousands of prosthetics.
In 2019, brainco made the world’s first brain controlled mechanical prosthetic product, brainrobotics intelligent bionic hand, brainrobotics collected the EMG and neuroelectric signals at the end of the disabled’s residual limbs, restored the movement intention of the disabled through the algorithm of deep learning, and let the prosthetic perform corresponding actions to achieve “hand with heart”. This product won the top 100 best inventions of time magazine in 2019 and was on the cover of the magazine.
Brainrobotics intelligent bionic hand plays the piano
“It took us five years for the team to make a hand. This hand can move every finger very flexibly, and can rely entirely on the user’s own imagination. At present, this hand is also the only intelligent prosthetic in the world that can be directly controlled through consciousness, and the user does not need to do any surgery, and directly puts a group of high-throughput information on the residual arm The collector collects his electromyography and neuroelectricity. ”
at present, brainrobotics intelligent bionic hands can be purchased and installed through the assembly center, some of them are uniformly purchased by the government or charity. Nearly 1000 units are shipped a year, covering 80% of the provinces, and the price is about 1 / 5 of the same type of products.
On December 21, brainrobotics intelligent bionic leg was officially released. According to the relevant person in charge of brainco, the traditional prosthesis can not receive instructions from the brain like the normal leg, and can not make real-time gait adjustment according to the environment and muscle conditions. Such a prosthesis is only a support tool.
In the past, the difficulty of intelligent bionic leg was how to judge the action intention of different disabled people and how to adjust the algorithm in different environments. “The flexibility of the intelligent lower limb is much less than that of the intelligent upper limb, but the difficulty is that different roads bear the weight of the human body, the pressure changes during the contact between the sensor and the skin, and the treatment of noise is very challenging.”
Brainrobotics intelligent bionic leg dance
In order to achieve such a fast and accurate judgment, the intelligent bionic leg needs to collect up to 20000 EMG and neuroelectric data points in one second and intelligently identify the intention in order to understand the user. After 2 years, the intelligent bionic leg was officially put into mass production.
In addition to the intelligent prosthesis in the rehabilitation scene, brainco also works in the education and health scene. At present, there are educational products such as focus focus focus improvement system to improve students’ focus. On November 27, 2019, the scientific journal sensors published papers by nataliya kosmyna and Pattie MAEs, two scholars from the MIT Media Lab. The results show that the test scores of testers using this system are nearly 40% higher than those of non users (random group).
it is understood that the official price of focus focus focus improvement system is 4980 yuan. At present, it has entered 400 educational institutions.
It is reported that according to the calculation of quantum analysts, at present, the market scale of brain computer interface optical equipment in China has reached the level of one billion yuan. It is estimated that by 2040, the market scale of brain computer interface pure equipment in China is expected to reach 56 billion yuan, with an average annual compound growth rate of 21%.
According to Zhentai intelligent analysis, the market of pure brain computer interface equipment in China will have an obvious amplification effect after combined with specific use scenarios, which may reach 100 billion in the future.
Han Bicheng told the all media reporter of Nanfang finance and economics that in the future, the company will also invest in R & D in two key directions: delaying the process of Alzheimer’s disease and solving sleep disorders. “Next year, relevant products will be launched and all will be commercialized.”
meta universe entrance: neural interface platform
2021 is the first year of metauniverse. In October, Facebook was officially renamed meta.
In fact, Facebook’s layout of the metauniverse began many years ago and has acquired at least two companies. Among them, oculus, a VR headwear equipment company acquired with us $2 billion in 2014, and Ctrl labs, a brain computer interface company acquired with us $10 in 2019, is another mysterious company.
CTRL labs produces armband equipment that can collect EMG and neuroelectric signals on the arm. Through it, the real interactive actions in real life can be projected into the virtual world. In other words, this is the key equipment to realize the human-computer interaction of Zuckerberg’s version of the meta universe.
What is the meta universe?
The meta universe can be understood as a digital parallel universe. Therefore, no matter how to build the meta universe, it is impossible to continue to feel the meta universe through the mouse and keyboard. The entrance problem is always an unavoidable topic.
For the deep cultivation under the concept of meta universe, Han Bicheng said that he would not be a VR terminal product, but chose to be a supplier of underlying technology and an import product of meta universe, so as to provide underlying technology output for all kinds of technology enterprises in the direction of meta universe. “We are willing to be an android in this field, share products and technologies with other manufacturers, and adjust our SDK according to the results that each manufacturer wants to achieve. For example, if it wants to have very fine operations, such as typing in the VR world, then we can provide a high-channel information collector.”
Brainco team believes that the establishment of meta universe is indispensable to the reform of human-computer interaction mode. in the PC era, human-computer interaction is carried out through keyboard and mouse, and then there is the rise of AI technology of multi touch and gesture control; Every change of human-computer interaction is from machine centered to human centered natural interaction, and the ultimate form of human-computer interaction should be human instinctive interaction.
Human muscles will produce very weak electrical signals whether in a relaxed state or in a moving state. Brainco has built a system combining sensor hardware and signal processing algorithm, which can accurately collect this signal and accurately identify people’s action intention through this signal. This year, the sensing system has been comprehensively upgraded to create a new neural interface platform.
At the scene, brainco demonstrated the functions of gesture recognition, space puzzle and space typing. Through this platform, real interactive actions in real life can be projected into the virtual world, which can do more than mouse and keyboard, breaking the constraints of field and space in the two-dimensional world and entering the interaction in the three-dimensional world, just like natural limb movements, Without mouse, keyboard handle or any controller, you can directly touch, command and create with your own hands.
(21st Century Business Herald)
