On May 9th, Pacific Time, it marked exactly 100 days since the first human trial participant under Elon Musk’s Neuralink underwent surgery to implant a brain-machine interface device.

However, a blog post released by Neuralink the day before revealed that within weeks of the January surgery on patient Noland Arbaugh, some of the wires embedded in the brain tissue had come loose, affecting the rate of information transmission and resulting in the loss of some data.

Neuralink did not disclose why the electrode wires unexpectedly contracted. However, according to media sources citing insiders, Neuralink believes one possible reason is that air remaining in Arbaugh’s skull after the surgery could have caused his motor cortex to retract. Some professionals also believe that the issue may be related to the way the implant was connected.

For Neuralink, which is seeking to conduct more extensive clinical trials, any malfunction could potentially delay the approval process by the U.S. Food and Drug Administration (FDA). However, as of the time of writing, the FDA has not made any comments on the incident.

Less than 100 days after the implantation, Neuralink’s first trial subject experienced a malfunction

In a tweet on May 9th, Elon Musk wrote: “Neuralink’s first human subject has successfully passed 100 days.”

Photo/X

The “first trial subject” of Neuralink is 29-year-old Noland Arbaugh, who suffered a spinal cord injury in a diving accident eight years ago. On January 28th this year, he underwent surgery to implant Neuralink’s brain-machine interface device, which uses 1024 electrodes to record neural signals, distributed across 64 threads finer than human hair.

Photo/Neuralink

However, a Neuralink blog post attached to Musk’s tweet revealed that after Arbaugh’s surgery in January, the device implanted in the trial patient experienced numerous mechanical failures. Within weeks of being implanted into the human brain, some of the electrode threads embedded in the brain tissue came loose. By late February, the data captured by Noland Arbaugh’s implant began to decrease.

Photo/Neuralink Blog

Neuralink did not disclose the cause of the unexpected contraction of the electrode threads. As a solution, Neuralink stated they modified the recording algorithm to be more sensitive to neural group signals, improved the technology for converting signals into cursor movements, and enhanced the user interface. These fixes “resulted in a rapid and sustained improvement in BPS (bits per second), surpassing the initial performance post-implantation.”

Despite the decline in the implant’s functionality, Arbaugh was still able to complete a live chess demonstration. Last Saturday evening, local time, Arbaugh also live-streamed on platform X, demonstrating how he uses the implant to navigate the computer screen and play games.

Neuralink indicated that although some neural threads in Arbaugh’s brain tissue have contracted, he uses the brain-machine interface system for about 8 hours on weekdays and up to 10 hours on weekends.

Photo/Neuralink Blog

Why did the malfunction occur?

According to The Wall Street Journal, insiders revealed that Neuralink suspects one possible reason for the malfunction could be air trapped in Arbaugh’s skull post-surgery, known as pneumocephalus, which might have caused his motor cortex to retract. This issue does not seem to pose a threat to Arbaugh’s safety. Despite this, researchers are considering the possibility of removing the device implanted in Arbaugh’s skull.

Bloomberg reported that some professionals believe the problem may have arisen because the electrode wires are connected to a device inside the skull, rather than directly to the surface of the brain tissue. They note that brain tissue can move significantly within the cranial space, and traditionally, brain implants are placed directly on the surface of the brain tissue, allowing them to move like a ship on water. “For brain implants, the retraction of electrode wires is abnormal.”

Dr. Eric Leuthardt, a neurosurgeon at Washington University School of Medicine in St. Louis, said, “Engineers and scientists did not realize the extent to which the brain moves within the cranial space. Merely nodding or suddenly moving the head can cause several millimeters of disturbance.”

Currently, Neuralink is attempting to implant their devices in more human subjects, aiming to implant their N1 brain implant in 10 patients within this year.

However, for Neuralink, which is seeking to conduct more extensive clinical trials, any malfunction could potentially delay the FDA approval process. As of the time of writing, the FDA has not responded to media requests for comment on the matter.

The concept of “human-machine symbiosis” still has a long way to go

According to the latest documents submitted by Neuralink to the SEC last November, the company has raised at least $323 million in funding, pushing its valuation over $5 billion. Despite being one of the highest-valued brain-machine interface enterprises globally, Neuralink does not hold an absolute leading position.

Controversies surrounding Neuralink have been ongoing for the past few years, with animal testing risks being one of them. Before being granted permission for human trials in May 2023, Neuralink was twice denied by the FDA due to a higher-than-average mortality rate in experimental animals.

In November of the same year, four U.S. congress members requested the SEC to investigate whether Elon Musk committed securities fraud by allegedly misleading investors about the safety of the brain implants being developed by Neuralink. This was after Musk claimed on platform X that “no monkeys have died as a result of Neuralink implants.”

Moreover, many have criticized Neuralink’s technological path and Musk’s grand vision. The Physicians Committee believes that Musk’s goal of achieving “symbiosis with artificial intelligence” through Neuralink’s technology does not necessarily align with the best methods for treating patients. Non-invasive brain-machine interfaces have proven to make progress in improving patient health, and Neuralink’s invasive implantation method may pose unnecessary risks.

Gao Xiaorong, a long-term professor at Tsinghua University who has been involved in brain-machine interface research for 20 years, mentioned in a media interview that Neuralink has not completely resolved the safety issues of its implantation devices. Although safety has improved compared to earlier pig trials, Neuralink has not published papers or disclosed technical details.

Miguel Nicolelis, a neuroscience professor at Duke University Medical School and known as the “father of brain-machine interfaces,” also stated that invasive brain-machine interfaces are meant for scientific research and are not the best option for patients. Implantation methods should be limited to very severe cases.

According to foreign media reports, brain-machine interface technology is an emerging field of research still in its early stages, involving multiple disciplines such as computer science, neuroscience, cognitive psychology, biomedical engineering, mathematics, signal processing, clinical medicine, and automatic control. There are still numerous problems to be solved. For instance, how to accurately output information from the brain and how to input the correct information into the brain?

Neurons in the brain are responsible for inputting and outputting information, and the brain-machine interface intervenes in this process. The entire cerebral cortex has a volume of about 500,000 cubic millimeters, containing approximately 20 billion neuron cell bodies, with an average of about 40,000 neurons per cubic millimeter of cortex. In addition, there are glial cells in the brain, roughly equal in number to neurons, as well as blood vessels. The total length of capillaries in one cubic millimeter of cortex can reach up to one meter.

To precisely capture or feedback brain signals, it is necessary to detect signals from specific neuron cell bodies within this cubic millimeter area or stimulate certain cell bodies to emit signals required by engineers. The complexity of this task is evident.

All these factors indicate that Musk’s vision of “human-machine symbiosis” still has a considerable distance to traverse.

Disclaimer: The content and data in this article are for reference only and do not constitute investment advice. Verify before using.