The fact that humans know little about brain activity makes the market of Nuggets still seemingly far away.
Editor’s note: This article is from “Tencent Technology” “Silicon Valley Cover” section, author: Jiao Han, authorized reprint.
[Editor’s note]
March announced last month that its brain-computer interface company, Neuralink, has made breakthrough new developments and experiments on animals have been successful. Facebook then announced that it could directly read words and phrases in the human brain through implantable electrodes. In addition to these two companies, there are many Silicon Valley startups involved in this field. The difficulties and challenges in the field of brain-computer interface have not blocked everyone’s enthusiasm. However, the fact that humans know little about brain activity makes the market of Nuggets still seemingly far away.
[Key Points]
-
Last month, Musk announced that its brain-computer interface company, Neuralink, has made breakthrough new developments, and then Facebook announced that it can directly read words and phrases in the human brain through implantable electrodes.
-
The scalable high-bandwidth brain-computer interface system developed by Muscat Neuralink has successfully carried out animal experiments, which will begin clinical trials as early as next year.
-
In addition to Neuralink, American startups Paradromics, Cyberdynamics and Kernel are conducting research on implantable brain-computer interface technology.
-
Compared to the risky invasive brain-computer interface technology, companies such as Ctrl-labs and Facebook have conducted brain-computer interface technology research through non-invasive methods such as skin contact.
-
Size and technology companies are excited about brain-computer interface technology, but the biggest obstacles in front of them are still insurmountable: humans still know very little about brain activity.
(This article is about 4,000 words, it takes about 5 minutes to read the full text)
The following is the body of the article:
In the middle of last month, Elon Musk announced a series of new developments in its brain-computer interface company, Neuralink. The Silicon Valley star founder excitedly described a “skin-like” robot that can implant ultra-slim lines into the depths of the brain. According to Neuralink, this system has less damage to the brain and will eventually be able to read and write large amounts of information.
Illustration: Neuralink Brain Machine Interface Rendering
Muske said that Neuralink has used the sewing machine robot to perform at least 19 operations on the animal. The “line” circuit that was successfully placed in the animal’s brain allows the animal to directly control the machine with the brain, with a system success rate of 87%. Musk said that the Neuralink research team has succeeded in allowing “a monkey to control the computer through the brain.” It also revealed that Neuralink is applying to the US Food and Drug Administration (FDA) for clinical trials of humans as early as next year.
Illustration: Sewing Machine Robot in Neuralink Brain Computer Interface System
Muske pointed out that these results can solve neurological diseases and eventually be used to “achieve a symbiotic relationship with artificial intelligence.” He also said, “We can do a lot of things to deal with brain disorders and injuries, but this process will be very slow.” “This will be a slow process, and we will gradually increase the problem to be solved until we finally create a complete Brain machine interface.”
And at the end of July, Facebook also announced the latest research results of its brain-computer interface project. According to published in NatureA paper in the Communications magazine shows that the University of California, San Francisco, and the Pittsburgh-based Facebook Reality Labs researchers built a “brain-computer system” based on non-implantable wearable devices that accurately decodes testers in the brain. Conceived words and phrases.
Illustration: Facebook’s “brain-computer system” flow chart
In the experiment, the researchers placed an electrode on the surface of several volunteers who were receiving epilepsy and used a technique called cortical electroencephalography (ECOG) to directly record the cerebral cortex. In the process of volunteers answering test questions, the researchers track the brain’s language and pronunciation area by signal, and use the machine learning algorithm built by the speech analysis model to decode the specific speech from the acquired signal.
But this system also requires electrodes to be implanted in the brain and is highly invasive. The researchers said that the system only identified 24 standard answers to nine questions, with a system accuracy of 61% to 76%. The effect is still far below the real-time decoding speed that Facebook expects to achieve.
But Facebook says the study helps people with language disabilities to regain communication skills.
Computer Interface Technology and Neuralink
Humans always have an obsession with developing brain function. Archaeologists have discovered that as early as the prehistoric human period, many ancient people would apply a “ring saw”, which is to dig a hole in the top of the skull. Archaeologists speculate that they are doing this “to gain some mysterious power.”
