The brain-computer interface that Elon Musk has made a viral hit is actually not nearly as popular as it appears.

Elon Musk brought brain-computer interface (BCI) into the public eye. The visions of enabling paralyzed patients to regain control of their limbs, giving aphasic people the ability to express themselves, restoring sight to the blind, and even directly reading human thoughts have continuously raised people's expectations for this technology. However, in the eyes of Zhao Yun, founder and CEO of NewCloud Medical (EMBA 2024, China Europe International Business School), brain-computer interface is still a long way from the future people imagine.

He has long worked in hospitals and the pain management field, so he pays more attention to specific practical issues. What diseases can a certain technology actually solve? Do doctors really need it? Can patients get improvement after using it? And ultimately, how many people are willing to pay for it? In his view, the future of BCI is worth looking forward to, but for the technology to evolve into a real industry, many things have only just begun.

Brain-computer interface is still in its early stage, far from the "mind-reading" scenario

In recent years, BCI has been a hot topic, with Elon Musk and Neuralink frequently appearing in news coverage. Many people ask: Has the technology developed to the point where it can tell what people are thinking? In fact, we are far from reaching that stage.

The so-called current brain reading is mostly about identifying a specific type of clear brain signal. For example, when a patient wants to raise their hand, the system collects the corresponding neural signal, decodes it, and then makes a robotic arm or other equipment complete the movement. Being able to read this movement intention relatively accurately is already a remarkable achievement in itself.

But this is completely different from knowing exactly what a person is thinking in their mind.

Some abnormal neural signals generated by diseases are relatively easier to identify. For instance, in cases of pain, movement disorders or other neurological diseases, the brain will show abnormal activities, which we can try to capture. However, there are still many unsolved questions about how human thoughts and memories are formed and stored.

Therefore, at least based on the current technological level, I don't worry too much about the so-called mind-reading technology.

From the perspective of technical routes, BCI can currently be roughly divided into invasive, semi-invasive and non-invasive categories.

Invasive BCI penetrates into brain tissue, collecting relatively clear signals, but it faces challenges including trauma, immune reaction and long-term stability.

Non-invasive BCI is more user-friendly, requiring no implant surgery and easier to be accepted by people. But the signals have to pass through the scalp and skull, losing a lot of information when reaching the equipment, so there is still much room for improvement in accuracy.

Semi-invasive BCI falls in the middle of the two. It does not go deep into brain tissue, aiming to strike a balance between signal quality and safety.

At the current stage, it is almost impossible for one single route to have all advantages. Invasive, semi-invasive and non-invasive BCI all face the same "impossible triangle": it is very difficult to achieve the best performance at the same time in terms of signal quality, trauma level and long-term safety. At this stage, BCI is still seeking balance among the three factors.

That's why I always think this industry is still in its very early days.

At present, the most widely seen applications of BCI are still movement intention recognition and motor rehabilitation. Researchers are working on language decoding, vision restoration, Parkinson's disease treatment, depression intervention and even more neural function reconstruction, but most of these applications are still in the research or early clinical exploration stage, far from mature and universal products.

BCI is indeed very popular, but technical popularity does not equal industrial maturity.

Why might pain management applications enter clinical use earlier?

When we develop pain-focused BCI, we do not first have a BCI technology and then look for its applicable scenarios. On the contrary, after years of work in the pain management field, we found that existing methods still cannot solve many problems, which led us step by step to the BCI track.

I used to work in hospitals for a long time, and later participated in the construction of pain management departments. After the departments were gradually established, we found a very realistic problem: doctors are still lack of proper tools.

Chronic pain is a typical example.

Acute pain is essentially a human protective mechanism. People will immediately withdraw their hand when touching something hot, and after a fall, they will naturally be more careful the next time they go to the same place. Without pain sensation, it would be very difficult for people to protect themselves.

But when pain lasts for a very long time, the situation changes. The tissue damage may have already healed, but the pain remains. At this point, the pain itself becomes a problem that needs to be treated.

What's more troublesome is that pain is very difficult to measure.

Body temperature is relatively easy to measure, with clear numbers showing whether it is 37℃ or 38℃. Pain is not like that. For the same stimulus, some people think it is tolerable, while others feel extremely uncomfortable. It involves not only neural activities, but is also affected by factors such as emotion and memory.

Therefore, we have been thinking about whether we can turn pain, which used to be mainly described subjectively by patients, into a type of neural signal that can be collected and analyzed.

This is exactly where BCI can play its role.

Our goal is not to make people completely lose pain sensation. Pain itself has protective effects. A more realistic approach is to control pain within a range that patients can accept.

When the system detects that the abnormal pain level exceeds this range, it will analyze based on the collected signals, adjust the stimulation parameters, and intervene in the nervous system. If the process of collection, decoding, stimulation and feedback can be continuously completed, a closed loop is formed.

I believe pain management is likely to become one of the earliest landing scenarios for BCI, which is related to its own characteristics.

A very important reason is fast feedback. Patients can quickly feel whether their pain is relieved. It is unlike some rehabilitation therapies that require months of training before the effect can be evaluated.

In addition, many motion-focused BCI systems require external equipment such as robotic arms and pneumatic gloves for cooperation, making the whole system relatively complex. Pain-focused BCI mainly revolves around neural signal collection and stimulation, with a relatively more direct closed loop.

There is also a very realistic factor, which is the large patient base and huge demand. There are a large number of chronic pain patients, and when the pain is severe enough to affect their sleep, work and daily life, they will have a very strong willingness to improve their pain condition.

Therefore, when I evaluate a medical technology, I do not first ask how advanced it sounds. I prefer to look at several questions: are there a sufficient number of clear target patients, can the effect be perceived by patients, and is there indeed a lack of such treatment method in clinical practice?

If all these conditions are met, the technology will have more opportunities to move from the laboratory to clinical practice.

Of course, pain management is only one direction of BCI.

Looking further ahead, without the help of external peripherals, BCI can enable people with mobility impairments to regain control of their limbs, help people who have lost language ability to express themselves again, provide visual information to people with vision loss, or regulate some abnormal neural activities, all of which can become BCI applications.

Essentially, what BCI does is not that mysterious. On one hand, it understands the signals sent by the nervous system more accurately, and on the other hand, it finds ways to re-communicate with the nervous system.

This is exactly where its real value worth expecting lies.

Getting regulatory approval is not enough, the final judgment depends on whether patients are willing to use the product

It is not surprising that BCI has attracted a lot of attention from capital in recent years in my opinion.

What investors value is not just a single medical device, but the broader possibilities behind it. If BCI really becomes a new way for humans to interact with the external world in the future, its impact will go far beyond the treatment of several diseases. That's why people are willing to bet on this future in advance.

But when it comes to real industrial development, we finally have to return to reality.

For medical products, from the laboratory to clinical trials, then to regulatory approval, entry into hospitals, and finally real use by patients, the challenges at each step are different.

Many current BCI projects are still in the clinical trial stage. At this stage, patients usually do not need to bear the full cost, so the large number of people willing to participate in trials does not mean that after the product is launched on the market, people will also be willing to pay for it out of their own pockets.

After entering the commercialization stage, many realistic problems will emerge.

The cost of the equipment, the cost of the surgery, how much can be reimbursed by medical insurance, how much the patient has to pay themselves, and how much improvement they can actually get after the treatment, all these factors will affect the final choice.

Therefore, in my view, getting access to payment channels including medical insurance reimbursement is certainly a good thing, but it is just the beginning.

With payment channels in place, the industry can truly test the actual demand. In the past, trials were free of charge, and next we need to see whether patients are willing to choose this product in real clinical scenarios.

Enterprises also need to gradually solve two problems. One is to reduce costs through engineering optimization and large-scale production, and the other is to identify more suitable indications. What kind of patients have the strongest demand, what kind of curative effect is the most clear, what kind of improvement is worth patients paying for, all these need to be gradually verified in the market.

The companies that can survive in the end are not necessarily those with the most fancy-looking technologies, but those that truly solve clinical problems.

Policies are of course very important. BCI is not an industry that can mature in just a few years, it requires a long period of scientific research investment, clinical verification and industrial accumulation. With continuous policy support, China has great opportunities in this field.

At present, the United States is moving faster in some cutting-edge technologies, while China is catching up at a very rapid pace. Especially in the stage of engineering transformation and clinical application, China has its own unique advantages.

While favorable policies can help the industry develop more smoothly, the final performance still depends on the product itself.

Whether a product can run stably, whether doctors are willing to use it, whether patients can get effective improvement after using it, and whether the price is acceptable, all these are unavoidable core issues.

Therefore, I think it is too early to judge which company will become the final industry leader. The companies that are at the forefront today may not still be the leading players ten years later. There is still huge room for change in this industry.

I am optimistic about the long-term development of BCI.

In the past, the Internet realized the connection between people and information, and the mobile Internet made this connection available anytime and anywhere. If one day we can understand and transmit neural signals more accurately, the connection between humans and machines, as well as between humans and the environment, will continue to evolve. I call this concept "Brain Internet".

It is hard to say when this future will arrive.

But for the medical field, we do not need to wait for that far future. As long as one day we can read the neural signals sent by diseases earlier and more accurately than now, and make timely interventions, this technology will already be of great significance.

Rather than focusing on when BCI can realize mind-reading, I care more about when it can first "read" diseases accurately.

This article is from the WeChat official account "China Europe International Business School" (ID: CEIBS6688), written by Zhao Yun, authorized for release by 36Kr.