NeuroXess Files IPO Prospectus: Pioneering China's First Fully Implantable, Wireless, and Functional Brain-Computer Interface

Huashan Hospital's Mao Ying and Chen Liang Professor Team Deeply Discuss the First "Three-All" Brain-Computer Interface Clinical Trial with NeuroXess Team

NeuroXess Founder and Chief Scientist Tao Hu (right) Playing a Game with a Patient (via Brain-Computer Interface Mind Control)

NeuroXess Founder and Chief Scientist Tao Hu (right) Playing a Game with a Patient (via Brain-Computer Interface Mind Control)

By Staff Reporter Gao Yang

"Did you know? He's already better at playing games than I am!" The day before yesterday, at the 2025 TCCI Brain-Computer Interface and Artificial Intelligence Forum, Tao Hu, founder and chief scientist of NeuroXess, said with a hint of "resentment." The "he" referred to by Tao Hu is a young patient who has been paralyzed from the neck down for eight years. After being implanted with the latest achievement from NeuroXess and undergoing two months of training, he can now skillfully play games, shop online, and control smart home devices adapted to the brain-computer system.

This clinical trial was led and completed by the team of Professors Mao Ying and Chen Liang from Huashan Hospital. The device implanted in the patient is China’s first and the world’s second fully-implantable brain-computer interface product with an internal battery — a "three-complete" system — which makes NeuroXess one of the very few companies globally to have mastered and validated a fully-implantable solution with an internal battery.

When stories of patients regaining their lives through brain-computer interfaces appear in the media, can those who "can't move, can't see, and can't speak" rely on this new technology to "save their lives"?

Innovative "China Solution" Stays Away from the Brain's Built-in Battery

The patient's surgery was performed on October 30 this year. The so-called "three-all" refers to full implantation, fully wireless, and full functionality.

"The product fully implants all core modules within the body, with no cables or interfaces exposed on the surface, physically eliminating the risk of infection commonly associated with traditional external systems," introduced Tao Hu. "At the same time, it integrates wireless power supply and wireless data transmission functions, allowing patients to completely摆脱体外设备的束缚, enabling 'full-scene' application."

Unexpectedly, NeuroXess innovatively placed the battery module under the skin near the chest. "Everyone has experienced their phone heating up while charging. When that happens, the phone's temperature rises by a few or even more than ten degrees Celsius," said Tao Hu, drawing an analogy. "However, if the brain’s temperature increases by more than 2 degrees Celsius, it can lead to fainting."

On the other side of the ocean, Musk's Neuralink has encountered the same problem. Their solution is: the battery remains inside the skull, using a data-lossy transmission method to keep the battery temperature within a safe range at the cost of sacrificing data.

"We have moved the heating unit away from the brain and implanted it subcutaneously in the chest, which is more tolerant to temperature, significantly enhancing the system's safety," Tao Hu told reporters. NeuroXess has also considered that if there is a need for upgrades or maintenance in the future, "thoracic surgery" is clearly more acceptable to patients than "cranial surgery."

It is worth mentioning that "full wireless" was a recurring expectation raised by patients and their families during the R&D process. If brain-computer interfaces could only rely on external power sources, then a "power outage" would leave the person "without power" as well; with an internal power source, during postoperative rehabilitation and daily life, patients wouldn’t have to "drag around a tail," truly achieving "wireless freedom" in terms of mobility and dignity.

Tao Hu introduced that the cortical flexible electrode used in this product does not directly penetrate the brain, which significantly reduces short-term damage to patients and improves long-term stability. It can also cover higher-level language functional areas.

From Startup to Mastery: Mind-Controlled "Hands Follow Brain"

Transferred to a general ward one day after the surgery, and successfully achieved mind control upon the first activation five days post-operation; after systematic training, the brain-controlled decoding rate in standard tests reached 5.2 bits per second, comparable to international top levels… Numbers don’t lie. The new generation of brain-computer interface products from NeuroXess demonstrates excellent safety and superior performance.

"Patients can control 'Mario Kart' computer battles and 'Snake' multiplayer online battles via thought, and even win. He can manipulate each finger with a pneumatic hand and flexibly brain-control humanoid robots."Robot." Introduced Professor Mao Ying, president of Huashan Hospital Affiliated to Fudan University.

"I can't beat him in the coin-grabbing game," Tao Hu said with a smile. In terms of gaming operations, the basic principle is that the brain generates signals, which are transmitted through the central nervous system to the peripheral nerves to control external muscles; however, by using the "external aid" of a brain-computer interface, one can achieve "hands moving as the brain commands" without any "middleman" delaying the process. From this perspective, Tao Hu's defeat is quite understandable.

"Currently, a two-hour charge can sustain a patient's daily training volume," he told the reporter. The industry consensus on the baseline for evaluating a product’s charging capability is that charging time should not exceed usage time. At present, the patients are undergoing relatively mild training sessions, one hour each in the morning and afternoon; however, this 28-year-old young man, who has been paralyzed for eight years, after experiencing the benefits, voluntarily suggested that the training duration could be longer and the game difficulty could be higher.

When we bring the timeline back to the operating table, a brain-computer interface surgery sounds like a complex, highly challenging, and time-consuming "tug-of-war." However, Mao Ying said that this is a "minor surgery."

"Neurosurgery has evolved from being highly invasive to minimally invasive, and now it is approaching non-invasiveness. With neuro-navigation technology, we can locate lesions with sub-millimeter accuracy — within 1 millimeter — making surgeries much safer," Mao Ying explained. Brain-computer interfaces may seem to require the implantation of "foreign objects" into the brain, but in reality, many patients already have biocompatible foreign objects in their bodies. These "implants" have been clinically tested for decades and have proven to be able to coexist peacefully with the brain.

"Moreover, every implant surgery has undergone extremely rigorous ethical论证, including the patient's适应证 and the safety of the materials, etc." Mao Ying补充道.

Solving Real Pain Points: Brain-Computer Interface Is More Than Just "Showing Off Skills"

Patients who undergo brain-computer interface surgery often demonstrate recovery progress by playing games. Is this "showing off skills," or has the definition of "rehabilitation" for the new generation of brain-computer interfaces transcended the traditional scope of "restoring basic motor functions," moving towards the broader goal of "rebuilding meaningful life and social participation"?

Shan Chunlei, Dean of the Yuanshen Rehabilitation Research Institute of Shanghai Jiao Tong University School of Medicine and Director of the Rehabilitation Center at Tongren Hospital, said that the goals of modern rehabilitation medicine are not limited to the recovery of motor functions but also include patients' cognitive functions, language abilities, emotional well-being, and even daily social interactions.

"Through brain-computer interface technology, it can indeed accelerate the remodeling of damaged neural circuits. They can achieve playing games, reaching for objects, and interacting with people." said Chuanlei Shan.

Two years ago, Shan Chunlei led the establishment of the "Brain-Computer Interface and Rehabilitation Committee" under the Chinese Association of Rehabilitation Medicine, which brought together rehabilitation physicians, brain-computer interface scientists, as well as experts in neuroscience, linguistics, psychology, and computer science. The initiative had two objectives: to keep rehabilitation physicians updated on the current state of brain-computer interface technology, and to help brain-computer interface developers understand patient needs, facilitating the practical application of their products.

During the interview, Tao Hu told the reporter a story: a few days ago, he was communicating in a patient group for ALS patients, and everyone was very interested in the new achievement of NeuroXess, asking continuously whether this product could be used on ALS patients and how many hospitals were conducting clinical trials.

It is not difficult to understand that, from the perspective of modern medicine, if you want a person to "stay alive," you just need to put them on a ventilator. However, for patients with ALS (Amyotrophic Lateral Sclerosis), losing motor function and relying on a ventilator to live is considered "living without dignity." "For enterprises, a deeply moving statement from patients is 'Even if there is only a 1% chance of hope, we are still willing to try,'" said Tao Hu.

In the current era of rapid artificial intelligence development, two "branches" have gained significant attention: embodied intelligence and large language models. Interestingly, brain-computer interfaces are closely tied to both. "Embodied intelligence, whether in robots, robotic arms, or unmanned vehicles, involves motion control, and patients can use brain-computer interfaces to control them."

Notably, NeuroXess's wide-area multi-point sampling architecture provides richer raw neural signals with more pronounced spatial characteristics for large AI models, facilitating the training of more powerful and versatile decoding algorithms.

At the forum that day, Tao Hu's speech won applause from the audience: All choices of technical paths and the development of functional scenarios are guided by addressing the real pain points of patients. From achieving delicate finger movements through mind control to maneuvering a wheelchair; from regaining social enjoyment via online gaming to completing daily basic tasks with humanoid robots — every function realized represents an incremental restoration of dignity to patients' lives. "We firmly believe that brain-computer interfaces should not be tools for showing off technology but bridges connecting life and hope," said Tao Hu.

Passing Through the "Trumpet Mouth" to Return to Clinical Practice for the Benefit of Patients

Mao Ying revealed that, at this stage, it is very important to establish the definition of brain-computer interfaces and the various potential issues that may arise in the future, in a standardized manner. Of course, whether at the national or municipal level, or within the medical community, there is active promotion for the implementation of these standards. However, "a standard cannot be perfect at all times; we need to continuously update it according to developments."

"We have only taken a very small step. In the current wave of brain-computer interface (BCI) hype, it is essential to remain calm," pointed out Mao Ying. In the future, the brain-computer interface will be a bottleneck-like "funnel," with a large number of BCI scientists and companies currently converging towards this "funnel." At present, this "funnel" represents clinical application scenarios. From a medical perspective, this "bottleneck" is extremely strict. "Out of 1,000 brain-computer interface companies rushing towards the 'bottleneck,' only a few will make it through."

Brain-computer interfaces bring together multiple cutting-edge technologies, and with urgent demand, their development is an inevitable trend. "The current dynamic landscape of rapid advancements is conducive to technological progress, but ultimately, it must return to the fundamental goal of serving clinical needs and benefiting patients," said Mao Ying.

Brain-Computer Interface: Integration of Medicine, Industry, and Research is Particularly Important. This year, the first brain-computer interface cluster area in China was launched in Shanghai. The cluster area is less than 500 meters away from the Hongqiao Branch of Huashan Hospital. "Doctors can directly communicate clinical issues to industry R&D personnel; industry R&D personnel can also enter operating rooms or come to patients' bedsides to gain a more direct understanding of clinical applications."

"Will the development of brain-computer interfaces ultimately move towards 'non-invasive'?" This is a question everyone is concerned about.

"From the perspective of benefiting the majority, non-invasive is definitely the optimal state. Currently, when it's absolutely necessary, we can only resort to invasive methods. As technology advances, we will strive to minimize harm to patients as much as possible," said Mao Ying.

"The common people all understand a basic logic: 'If you can take medicine, don't have surgery.' 'If you can have minor surgery, don't go for major surgery.' However, now some 'superficial' technologies cannot solve the problem, so invasive methods are the only option," said Tao Hu. "But NeuroXess is working to clarify some issues: What is the signal correlation between invasive and non-invasive methods? Where are the capability boundaries of each? These are very important for the future development of brain-computer interfaces."

Chunlei Shan believes that invasive means craniotomy. Surgery always comes with risks, and the cost of invasive procedures is also higher. Therefore, invasive methods have a certain range of selection — those with severe paralysis, especially quadriplegia, severe dysarthria, or other serious functional disorders, who have a strong desire for rapid rehabilitation, can afford it economically, and for whom non-invasive brain-computer interfaces cannot help achieve their goals. "The two approaches are not contradictory; they are chosen based on the patient's condition and needs."