World Conference on Artificial Intelligence: Brain-computer interfaces are expected to treat als, aphasia and more

liu, tempo Date: 2021-08-13 09:37:42
Views:46 Reply:0

The “craniotomy free minimally invasive implantable high-throughput flexible brain-computer interface” independently developed by Shanghai Institute of Microsystems, Chinese Academy of Sciences can be applied to the diagnosis and treatment of major clinical brain diseases and brain function exploration. It is an important means to solve als, high paraplegia, epilepsy and other major brain diseases, and is also the core technology of the “human-machine” ternary integration of everything perception.


The World Artificial Intelligence Conference 2021 will be held at the Shanghai World Expo Center from July 8 to 10. July 5, Jiemian News reporter visited the Shanghai Institute of Microsystems, Chinese Academy of Sciences, to enjoy the charm of AI technology in advance. Here, the reporter saw the independent development by the microsystem “non-craniotomy minimally invasive implantation of high-throughput flexible brain machine interface”.


Brain-computer interface (BMI) is a channel that connects the brain to a computer or other external device to establish a real-time bidirectional connection between the brain and the device for information exchange. The research of brain-computer interface (BCI) started in the late 1990s, involving interdisciplinary techniques such as neurophysiology, computer science and engineering.


AI technology


Traditionally, brain-computer interfaces (BcI) have been studied to uncover and exploit the workings of the brain and to create therapies to restore movement and sensation to patients who have lost the ability to move. Recent studies have shown that the use of brain-computer interfaces has expanded from the original goal of being an extension of humans, controlling robots and other external devices in place of human limbs.


According to introducing, micro system which is developed from the Chinese Academy of Sciences Shanghai keyhole craniotomy implantable high-throughput flexible brain-computer interface technology, in the animal experiment stage at present, is applying for clinical trial ethics, after taking clinical ethics approval can be used for clinical significant of diagnosis and treatment of brain disease and brain function, is gradually cold syndrome, paraplegic, an important means of major brain diseases such as epilepsy, It is also the core technology of the “human-machine” ternary integration of all things perception.


Craniotomy free minimally invasive implantable high-throughput flexible brain-computer interface system is divided into four parts: front-end flexible deep electrode device, mid-end conversion unit, back-end EEG acquisition and transmission module, and biocompatible packaging material. At present, the system has been applied to a variety of animal models, such as mice, rabbits and monkeys, and can achieve postoperative acute signal acquisition (within 30 minutes after surgery) and stable neural signal tracking for up to 8 months.


The research of brain computer interface is developing rapidly and has a wide clinical application prospect. Some time ago, two brain-computer interface experiments were widely discussed.


On April 9, 2021, Neuralink, a technology company founded by Elon Musk that specializes in brain-computer interfaces, released a video showing Pager, a 9-year-old rhesus monkey, playing a computer game called MindPong by using “thoughts.” According to the video, the principle of the experiment is to implant a brain-computer interface into the brain of Pager, train the Pager to learn how to control the stick to play games, record the brain neural signals during the process of controlling the stick and the corresponding arm movements, so as to decode the brain signals of Pager. When Pager attempts to move the cursor after disconnecting from the stick, the resulting brain signals are decoded by the brain-computer interface system and converted into cursor movements, enabling “mind” play.


In a May 13, 2021 cover article in nature, researchers at Stanford University have created a brain-computer interface system that combines brain-computer interfaces with artificial intelligence to help a high paraplegic “write.” This “writing” process is not carried out by hand, but through the brain-computer interface system, the brain control “handwriting” neural signal decoding, and then the decoded information into letters in real time, so as to form text, realizing the “brain writing” free from the restrictions of the body.


In the Stanford study, researchers inserted two hard electrodes into subjects’ brains. The volunteer, a 60-year-old man who had lost almost all movement from the neck down to a spinal cord injury, was asked to imagine himself writing letters. He recorded the neural signals in his brain and used them as material to train an AI algorithm to decode the signals and predict the subjects’ hand movements.


The above two experiments both use implantable brain-computer interface, which can obtain brain neural signals with less noise and higher accuracy compared with non-implantable brain-computer interface. This kind of implantable brain-computer interface is also used in the “Craniotomy free Minimally invasive Implantable High-throughput Flexible Brain-computer Interface” attended this World Artificial Intelligence Conference.


Deputy director of the Shanghai institute of microsystem and information technology of Chinese Academy of Sciences, the researchers TaoHu on interface news pointed out that the two cases is one of the most successful worldwide implantable brain-computer interface experiment, which also can be seen that the implantable brain-computer interface should be put into use the difficulties faced by, tao Iliad their summed up as “bad” and “can’t use” two points.


This is reflected in the insufficient number of electrode channels. As the number of electrode channels increases and the number of neurons that can be simultaneously recorded and regulated increases, the ability to interpret, control and respond to brain signals increases exponentially, he said. A normal human brain has more than 80 billion neurons. Musk’s rhesus monkey experiment has only about 1,000 channels, so the number of channels that can be collected from the brain is very limited. Therefore, the urgent problem for bCI is how to increase the number of channels.


Tao explained that “unusable” means that the implantable brain-computer interface has very limited use in the human body. The implantation of the electrodes for the BCI requires extensive craniotomy, which means cutting the patient’s skull open to expose the brain, inserting the electrodes, and then covering the skull, all of which can be very traumatic. So it’s usually only people who are severely handicapped in their lives, like the high paraplegics in the Stanford study, who get the surgery after weighing the risks and benefits.


In addition, he noted, the Stanford experiment used hard electrodes, which are not suitable for brain implants. Because the brain is wrapped around the skull, the head is normally free to move, but when the head moves, the brain swings inside the skull, causing the electrodes attached to the skull to cut the brain, causing inflammation in the brain, and forming nerve scars in the brain that wrap the electrodes around, causing the electrodes to fail. The Stanford experiment was made possible by the inability of high paraplegics to move their heads


In general, the main difficulties faced by implantable brain-computer interface are the small number of electrode channels, the large incision of implantation operation, and the damage to the brain caused by hard electrode implantation and the failure of electrode. The “craniotomy free minimally invasive implantable high-throughput flexible brain-computer interface” project of the Chinese Academy of Sciences is a targeted solution to these difficulties.


Tao Hu told reporters that in order to solve the problem of insufficient electrode channels, the non-craniotomy minimally invasive implantable high-throughput flexible brain-computer interface changes the traditional manual or semi-automatic assembly mode, and uses integrated circuit to make electrodes. Integrated circuit can be tens of thousands or even billions of devices integrated in a chip, the use of integrated circuit can be large-scale increase in the number of channels in a single device, improve neural signal acquisition and regulation of data volume.


According to the report, in order to solve the problem of hard electrode, the non-craniotomy minimally invasive implantable high-throughput flexible brain-computer interface uses flexible electrode. The flexible electrode is the equivalent of a wire in the brain, and it can follow the brain around without cutting. In the current animal experiment, eight months after the bCI was implanted, the animals were free to move while awake, and the signals were stable and still being recorded.


Neuralink’s experiments also use flexible electrodes, but as Tao explains, Neuralink’s flexible electrode requires a steel needle to be inserted into the brain, so it still requires a craniotomy, which can cause large wounds. The project uses a special method of temporarily solidifying flexible electrodes that are inserted into the brain and then softened, so there is no need for a steel needle. The implant requires a small incision of less than a millimeter in the skull, which is 100 to 1,000 times smaller than the incision Mr Musk uses. The significance of such a small wound is that the body can heal itself, and the recovery period is much shorter than that of craniotomy.


At present, craniotomy free minimally invasive implantable high-throughput flexible brain-computer interface technology has been completed in mice, rats, rabbits, monkeys, monkey has been the closest animal model to human.


In the process of implantation, the brain computer interface accurately locates the brain area to be implanted through the brain stereoscopic locator, visualizes the position of the electrode in the brain spectrum, and uses the automatic surgical robot to complete the minimally invasive implantation. The process achieved 10 micron level precision control, saving 80% of the time cost of implantation, and the final implantation size is smaller than the diameter of the infusion pinhole.


Tao told reporters that the brain-computer interface will be used in clinical research in the near future, mainly for ALS, high paraplegia, and aphasia. All three are motor perception impairments that restore speech and motor skills through brain-computer interfaces. Although the patient cannot speak, he can use his mind to synthesize language and can choose which language to synthesize. Although patients cannot move their own limbs, they can use their thoughts to control external mechanical devices.


At present, the project team is trying its best to promote clinical research and carry out the examination and approval of clinical ethics of scientific research with neurodegenerative diseases such as ALS and aphasia as the key points, hoping to really apply the technology to medical treatment.

Leave a comment

You must Register or Login to post a comment.