What the brain says, the brain-computer interface knows

In August 2020, Musk, the founder of Tesla, announced to the media a major achievement of his company-the latest brain-computer interface technology.

Researchers implanted a brain-computer interface device the size of a coin into a pig’s brain to establish a connection between the brain and the outside world. When the pig’s nose touches an external object, the neural activity in its brain responds accordingly, which indicates that the implanted device has collected correct neural information. Two months after the implantation device was put in, they put the piglet on a treadmill, and the rhythm of its walking could be seen from the nerve signals in its brain. Musk said that this device has already applied for innovative medical devices, and the next step can be implanted in the human brain.

Some experts borrowed the words of American astronaut Neil Armstrong to comment on this technology: “This is a small step taken by individuals, but it is a big step taken by mankind.”

Science fiction blockbusters move towards reality step by step
So, here comes the question of the “eating melon” masses: Can people use their brains to control objects? In the movie “Avatar”, the scene in which the consciousness of all creatures can be connected with the machine, will it become a reality? With electrodes or chips installed in the brain, can you manipulate everything in your imagination and make everything connected, intelligent, and human-computer interaction?

If the scene described in “Avatar” is still a bit far away from us, then another science fiction film made by this film crew, “Alita: Battle Angel”, tells the story of the doctor who picked up a little girl from the garbage dump. The brain, made her a body, realized the story of the communion of human brain tissue and machinery, and it got closer and closer to us.

Clinically, paralysis is a common symptom of the nervous system, manifested as a decrease or loss of voluntary motor function, which is caused by nerves, neuromuscular junctions, or muscle diseases.

In recent years, the principle of brain-computer interface technology based on motor imagination involves the information related to the movement of the limbs contained in the brain electrical signal. When the brain starts to perform the task of motor imagination, by digging deep into the corresponding motion characteristics, it can identify the human body’s motion intention, and then drive the external equipment such as the robotic arm to truly control it by mind.

Paralyzed patients are therefore expected to use their minds to lift paper cups, flip through newspapers, and express their intentions. How to collect as few EEG signal channels as possible while ensuring high accuracy of motion intention recognition is an important guarantee for the productization and marketization of portable wearable devices.

Realize “speaking” with the brain
Dr. Cao Tianao of Harbin Institute of Technology pointed out that the electrical signals of the human brain contain a lot of rich physiological information, such as fatigue, concentration, emotion, pain location and intensity, sleep depth, anesthesia depth, and limb movement intentions. By digging deeply into the characteristics of EEG signals, the current physiological state of the human body can be reflected objectively and accurately. Therefore, the brain-computer interface technology does not rely on peripheral nerves and muscle tissues, and can directly construct a communication channel between the human brain and external devices. Even for patients who are paralyzed due to strokes and other diseases, the movement intention can be detected from their EEG signals , The real realization of “speaking” with the brain.

Then, what specific information about the movement intention is included in the EEG signal? How to capture EEG signals to determine exercise intention? How to ensure the accuracy of exercise intention judgment? In response to these problems, Professor Sun Jinwei from the School of Instrumentation Science and Engineering of Harbin Institute of Technology and his team members Cao Tianao, Wang Qisong and others have been published in many international journals such as the Journal of Physics and Measurement Science and Technology since 2018. Published a series of academic articles, solved the above problems from a professional perspective, and cast a “golden key” in order to open the “door” of unknown human EEG signals.

Find shortcuts to collect motion intention signals
Many patients with limb paralysis have minimal brain damage and still have good functions. Only the nerves, neuromuscular junctions or muscles that transmit motion signals are injured. Therefore, brain-computer interface technology can be used to explore EEG signals and explore the intention of motion. Cao Tianao explained that the patient needs to perform a motor imagination task, that is, to imagine the movement of one of his own limbs, but in fact the limb remains motionless, and the patient just enters the scene of pure limb movement imagination. At this time, the brain waves will show corresponding characteristics. After the researchers remove the noise of the collected brain signals, and extract the motion-related features, the patient’s motion intention can be predicted and identified.

The traditional idea is to use as many channels of EEG signals as possible to collect comprehensive information, use algorithms to extract features of motion intentions, and finally use professional classification methods to improve the accuracy of intention prediction. However, this method makes the experiment operation more complicated, the calculation is more time-consuming, and the equipment is larger, which is not conducive to the marketization of the actual product.

On the basis of controlling the accuracy of motion intention recognition within a certain range, Sun Jinwei’s team carried out weight analysis on the acquisition channels of EEG signals through the improvement of the analysis method, and judged the contribution of each channel one by one. The channel weights are expanded and sorted, and then the relationship between the number of acquisitions of high-weight channels and the accuracy of the final motion intention recognition is observed. It was found that a satisfactory recognition rate can be achieved with approximately 25% of the channels. When the number of acquisition channels is increased, the accuracy rate does not change significantly, and even shows a downward trend.

Professor Sun Jinwei emphasized that this “subtraction” strategy of selecting the best brain area channel can ensure the accuracy of brain movement intention recognition and at the same time, with as few acquisition channels as possible, it can give accurate predictions of the movement imagination in the mind. . Paralyzed patients can control external equipment through motor imagination, for example, by adjusting the mechanical arm to complete different actions, grasping and transmitting objects, drinking water and eating, etc., truly achieving the purpose of “mind control”.

Paralyzed patients can also “walk like flying”
The brain-computer interface decodes and reveals the brain nerve signals in the process of human thinking activities, and builds a direct “information highway” between the brain and the outside world. It is effective against amyotrophic lateral sclerosis, Parkinson’s disease, essential tremor, and high position. The treatment of severe movement disorders such as paraplegia is of great significance, and a brand-new breakthrough is expected.

Professor Zhang Jun, director of the Rehabilitation Center of Heilongjiang Agricultural Reclamation General Hospital, believes that paralyzed patients can use their EEG signals to cooperate with external intelligent equipment such as manipulators, prostheses, exoskeletons, and rehabilitation robots to express their intentions and perform actions. People who are bedridden can use EEG signals to manipulate external intelligent manipulators to help them move objects and finish drinking and eating. Amputation patients often live inconvenience and are shrouded in the double shadow of the body and the mind. The use of smart prostheses can complete actions in accordance with their own brain instructions, such as using the brain-controlled smart prosthetic leg to stand up, or even drive the prosthetic leg to walk, up and down. Stairs, and high-speed running, etc., return to the state of “steep steps like flying”.

Patients with stroke and spinal cord injury can also rely on brain-controlled intelligent robots to carry out rehabilitation training and complete cumbersome, difficult and detailed designated actions. Zhang Jun believes that in the rehabilitation process, the patient’s brain’s electrical activity reflects his strong desire and willingness to actively participate. This active participation induces the restoration of the brain’s nerve pathways, and the nervous system is remodeled and repaired, which is expected to cause injury. The long-lost limb motor functions “lost and regain” the disabled.

Looking to the future, Dr. Cao Tianao pointed out that workers in many types of work often have to move heavy goods, and problems such as lumbar muscle strain and hunchback are unavoidable for a long time. Once the smart exoskeleton is assisted, movers can use their own brain electrical activities to direct the exoskeleton “Hercules” to assist in moving large objects. There is no doubt that the competent weight of an exoskeleton is often several times the maximum strength of the human being. Workers equipped with exoskeleton can replace forklifts, making hard work simple and easy.

Such a vision, thanks to the efforts of Chinese and foreign scientists, is no longer out of reach.