How China’s changing the brain-computer interface race

(Edited and refined by James Loo, with the assistance of AI translation.)

A patient paralysed from the neck down was able to pick a pen again after a coin-sized chip was implanted in his brain.

This is not a scene from a science-fiction (sci-fi) movie, but a real case widely reported by Chinese and US media. In November 2024, 36-year-old Dong Hui, who had high-level paraplegia after a car accident, underwent brain surgery in which doctors placed a coin-sized wireless implant inside his skull.

Despite previously being unable to move below the neck, with training he was able once again to perform movements such as grasping and writing. His case has become one of the definitive examples of China’s brain-computer interface (BCI) technology moving from clinical trials towards real-world application.

The concept of BCI can be traced back to 1973. At the time, American scientists proposed a forward-looking idea: could electrical signals generated by the brain be used as computer input, bypassing muscles and peripheral nerves to enable direct communication between the brain and a computer?

This idea continued to evolve over the decades, gradually moving from the laboratory to the operating theatre. Today, BCIs can help paralysed patients control cursors, robotic arms and robotic gloves, and are even being used in attempts to restore language communication and assist the recovery of motor function.

Brain-machine scientist Feng Chao, associate editor of the international peer-reviewed journal Cognitive Neurodynamics, told Lianhe Zaobao (LHZB) that BCIs are “the future bridge connecting carbon-based life and silicon-based life”.

However, the human brain is extremely complex, soft and fragile. How to capture neural signals while minimising damage to brain tissue is the core challenge in the development of BCI technology.

The commercialisation compromise

Based on the depth of contact between electrodes and brain tissue, the development of BCIs can be divided into non-invasive, semi-invasive and invasive approaches.

Feng said non-invasive methods collect brain signals using devices placed outside the scalp, making them relatively safe — but signal precision is limited. Invasive methods by contrast involve surgically placing electrodes inside the skull, or even implanting them within brain tissue, to obtain higher-quality neural signals. However, the surgical risks are greater and there is a higher benchmark when it comes to long-term stability and ethics.

Semi-invasive methods fall in the middle. They usually require surgery to place electrodes inside the skull but outside the dura mater, without penetrating it or entering the brain tissue itself. This was the type of surgery Dong Hui underwent. Brain signals generated when the patient intends or imagines motion are decoded into commands recognisable by external devices, which then drive a pneumatic glove to perform the corresponding actions.

As of the end of last year, 32 patients in China with cervical spinal cord injuries, including Dong Hui, had undergone this semi-invasive BCI surgery. The system, called NEO, was jointly developed by BCI startup NeuroXess and Tsinghua University. In March this year, China’s National Medical Products Administration approved the related product for market launch.

Its approval marks an acceleration in China’s BCI industry. In 2025, China’s Ministry of Industry and Information Technology set a timetable for the sector: breakthroughs in key technologies by 2027 and a significant improvement in industrial innovation capacity by 2030. At this year’s Two Sessions — the annual meetings of the National People’s Congress and the Chinese People’s Political Consultative Conference — China included the development of BCIs in the government work report for the first time, clearly identifying it as a future industry to be cultivated.

Driven by strong policy signals, financing for related companies accelerated. NeuroXess filed for listing guidance in February and completed the process on 9 June. BrainCo, which uses non-invasive technology to control prosthetic limbs, also submitted a confidential listing application to the Hong Kong Stock Exchange in January. The two companies are competing to become the “first BCI stock”.

According to the Shanghai Securities News, in the first five months of 2026, there were more than 33 financing deals in the BCI sector, with the total exceeding 4.6 billion RMB (about US$679 million), surpassing the total for the whole of 2025.

Differing approaches push boundaries

To external observers, though China started later in the BCI field, it is rapidly catching up with the US. Technology magazines Wired and MIT Technology Review have both cited the approval of NeuroXess’s NEO system for market entry as an example, noting that China, through a “middle ground semi-invasive approach”, has moved faster than the US in commercial applications. The US currently has no similar implanted BCI product approved for commercial use in patient treatment.

However, the reports also noted that the US still leads in obtaining more precise signals and in pursuing bolder invasive methods.

“It would be more accurate to say that the US is currently more inclined to pursue the ‘upper limit of performance’, which is to use implanted BCIs to achieve higher-precision and higher-bandwidth human-machine control as quickly as possible, focusing on complex problems such as helping paralysed patients communicate and control robotic arms. China meanwhile, is advancing along multiple tracks. It is promoting invasive systems while also emphasising non-invasive and semi-invasive ones, because it is more focused on how to bring the technology into medical care, rehabilitation, industry, education and consumer scenarios as quickly as possible, forming an industry that can actually be implemented.”

Feng described the US as pushing against the “ceiling” of BCIs, while China is more akin to laying the “floor” and “networks”. In a truly mature future industry, it is unlikely to adopt a singular path — invasive systems may be responsible for breaking through performance limits while non-invasive systems would focus on expanding scale, and semi-invasive systems would strike a balance.

Max Riesenhuber, a professor of neuroscience at Georgetown University Medical Centre and the co-director of its Center for Neuroengineering, told LHZB that, in theory, the deeper electrodes go into the brain, the better the signals obtained — and the higher the risks. “But the question is: what do you want to use the signals for? NEO obtains suitable signals and, because of its low risk, was the first to be approved for paralysis treatment.”

Riesenhuber said: “Compared with the US, China’s advantage lies in its ability to put all these pieces in place in a top-down manner until the goal is achieved.”

He said this kind of centralised decision-making is especially important for the development of BCIs, because “a BCI is not a single technological breakthrough, but a systems project requiring the joint coordination of neuroscience, engineering, clinical translation, regulation and medical insurance payments”.

He also noted that China has a larger patient base and higher trust in government-backed fields, which helps companies recruit trial participants more quickly, accumulate data, improve products and accelerate real-world applications.

Beyond medicine: strategic BCI applications

Beyond medical care, the prospects for BCI technology are also supported by more scenarios, including anti-depression treatment, sleep regulation and emotion management. Riesenhuber also specifically mentioned potential applications in the military and high-risk working environments.

He said the military significance of BCI does not lie in sci-fi-esque “mind reading”, but in enhancing human-machine collaboration when combined with artificial intelligence (AI). This could include monitoring personnel attention, assisting with target identification, and even enabling people to remotely control robots or exoskeletons. “These strategic applications concern how far the potential of the human brain can ultimately be applied.”

However, academics and industry players alike caution that strategic potential and capital enthusiasm also require calm reflection.

On the one hand, BCIs deal with the human body’s most complex and sensitive organ, and long-term safety, data privacy and ethical boundaries still require observation and discussion. The influx of capital could also bring investment waste and overhyped marketing.

On the other hand, beyond medical care, BCIs have yet to produce large-scale consumer applications. Kong Xiang, former sleep medicine specialist at a leading Beijing hospital and current BCI entrepreneur, said in an interview that for BCIs to move from state recognition to broader applications, the key lies in deep integration with different industries and specific scenarios.

Kong said the true future value of BCIs lies not only in “reading the brain”, but even more in “writing from the brain”. To achieve more comprehensive applications, BCIs still need to be deeply integrated with different industries and specific scenarios.

He cited non-invasive brain-computer intervention for anxiety- and stress-related insomnia as an example, commenting that the difficulty does not lie in collecting and decoding EEG signals but in making those signals the “baton” that drives the product. Through real-time brainwave monitoring, the system would generate personalised closed-loop neuromodulation plans, adjust them both intelligently and dynamically in real time and optimise personalised plans over the long term. This would truly realise the goal of different solutions for different people or even multiple solutions for the same person.

Riesenhuber also said that, just as in the early development of AI, the next question BCIs must answer is not just what the technology can do, but which applications are truly valuable. The ultimate contest between China and the US may be over who can first identify application scenarios that are clear and real enough to bring this technology genuinely out of the laboratory and into the real world.

This article was first published in Lianhe Zaobao as “让高位截瘫患者握笔写字 中国脑机接口抢先商业化”.

Related: Is China a threat, or just out-engineering everyone? | Can China win the AI race with cheap power?