Carnegie Mellon University professor on Musk’s brain-computer interface: no neurodecoding progress.

“It is not clear whether Neuralink devices can be used to decode devices in human intent and control environments. As far as I can see, this ability has not been proved. Elon Musk’s brain-computer interface company, Neuralink, has returned to the global spotlight.

Professor He Bin, a brain-machine interface scientist and head of the Department of Biomedical Engineering at Carnegie Mellon University, told reporters that Musk demonstrated developments in engineering areas such as miniaturization of brain-machine interface devices and wireless communications, but did not appear to show “any significant progress” in neurocoding.

On August 29th, with more than two million followers, Musk broadcast a new generation of brain-machine interface implants, surgical robots and a piglet implanted with a brain-machine interface device. It was implanted into the brain-machine interface device two months ago and is in good condition.

Carnegie Mellon University professor on Musk's brain-computer interface: no neurodecoding progress.

A new generation of Neuralink robots.

During the broadcast, when a piglet’s mouth and nose touch an object, the brain-machine interface device captures the signal emitted by the neurons, presents a dot image on the display and emits sound, which shows that its brain signal can be captured in real time.

According to Musk, the “neural connection” of the new generation of brain machine interface implants is only the size of a coin, diameter 23 mm, thickness of 8 mm, fully wireless transmission, battery life of a whole day;

Musk’s presentation represents an important development in neurotechnology in engineering and surgical robotics, he said.

But he also noted that Musk did not show any significant progress in neurodecoding this time. “It’s important to narrow down the implantable brain-computer interface device from a recording and signal transmission perspective, but ultimately people need to prove what the system can do to decode brain signals and correctly interpret the state and intent of the brain.”

Based on this, many experts believe musk’s display of a brain-computer interface system is “a breakthrough in brain surgery without explaining brain signals.” In this regard, He Bin agreed.

Referring to the future of “neural connection” technology in humans, He said he believes “neural connection” technology could be applied to humans, such as patients who need heavy intervention, who could benefit from this invasive brain implant.

Brain-computer interfaces (BCI) refer to the transformation of brain-computer signals with related instruction signals by connecting the brain in some form to external devices. The concept was put forward in the 1970s. Brain-computer interfaces are mainly divided into invasive and non-invasive forms. Invasive brain-machine interface systems require surgery to implant chips and electrodes in the brain.

Because of the noise of brain electrical signals and the complex unknown mechanism behind the neural encoding, the decoding and translation of neural information is the core problem of brain-computer interface. About a decade ago, scientists decoded speech from signals in the human brain for the first time, but the accuracy and speed of decoding was much lower than the natural speed of speech.

Carnegie Mellon University professor on Musk's brain-computer interface: no neurodecoding progress.

He Bin.

He Bin is an excellent scientist in the field of non-invasive brain-machine interface, and his research team uses non-invasive brain-machine interface technology to realize the idea of controlling drone flight and crossing obstacles. In 2019, his team developed a brain-machine interface that can be connected to the brain without invasive connections, allowing people to control the continuous, rapid movement of the robotic arm.

In early 2020, He Bin was interviewed by the National Science Review, an academic journal, about the progress of brain-computer interfaces. In it, he talks about two aspects of the brain-computer interface. The first is to detect human brain signals, decipher the meaning of which, so as to control the machine, the other aspect is to send electrical, magnetic, acoustic signals into the brain, to stimulate or regulate the activity of neural circuits, mainly used in disease treatment.

Commenting on the brain-computer interface technology released by Neural Connections in 2019, he noted that if the news reports were accurate, he thought the “neural connection” was a major breakthrough in technology, a big step forward from the technology currently in the lab. In the future, this technology may be used in patients with epilepsy who require surgery. But he also believes that in the general public, the creation of technology is difficult to promote, no matter how fine the electrodes. Non-invasing brain-machine interface technology will have greater application prospects.

In recent years, in addition to the engineering advances represented by “neural connections”, researchers in the field of brain-computer interfaces have also made progress in understanding neuroscience mechanisms and invasive auditory brain-computer interfaces, He told reporters. On the other hand, non-invasive brain-machine interfaces have also made great progress, such as the application of artificial intelligence technologies such as deep learning to better decode brain signals, the development of brain electrical signals based on the continuous movement of the robotic arm.

Journalist Zhang Wei.