A fully implanted brain–computer interface has allowed a man with complete paralysis to control digital devices, a smart wheelchair, and robotic dogs using only brain signals, according to researchers from the Chinese Academy of Sciences. The work represents one of the most comprehensive demonstrations to date of invasive BCI technology operating outside laboratory conditions.

The system was developed by the Centre for Excellence in Brain Science and Intelligence Technology (CEBSIT) and tested in a clinical trial involving a patient with a high-level spinal cord injury. The researchers report that the interface provided stable, low-latency control across multiple devices and supported sustained daily use, including paid remote work.

The patient, identified as Mr Zhang, was left with paralysis below the neck after a fall in 2022. After more than a year of conventional rehabilitation produced little improvement, he joined the BCI clinical trial. In June, surgeons at Huashan Hospital in Shanghai implanted a wireless invasive brain–computer interface known as WRS01.

**Also Read: **EngineAI T800: Humanoid Robot Performs Incredible Martial Arts Moves

The system consists of tiny front-end sensors inserted via flexible electrodes, along with a processor chip embedded into a shallow groove in the skull. Power delivery and data transmission are handled wirelessly through a specially designed external wearable hat worn by the patient.

Following two to three weeks of training, Zhang was able to control a computer cursor and digital devices using neural signals alone. As reported by the South China Morning Post, he said in a video released by the research team, “It has been more than three years since the incident, and now I can finally work again.”

The ability to operate a computer enabled Zhang to perform paid remote work, verifying product dispensing from vending machines. This makes him the first documented case of a participant in a BCI trial carrying out sustained paid employment through an implanted neural interface. Zhang described the work as challenging but meaningful, calling it “a valuable opportunity.”

Beyond digital interaction, the team extended the interface to control physical devices. Zhang learned to operate a smart wheelchair and issue commands to robotic dogs, which can fetch items such as takeaway food. The wheelchair allows him to navigate outdoor environments and move down stairways without assistance.

The transition from screen-based control to three-dimensional physical movement presents significant technical challenges. Errors and delays that are manageable in digital interfaces can become critical when translated into real-world motion. To address this, the researchers focused on minimizing system latency.

Natural neural conduction in the human body takes about 200 milliseconds. By developing a custom communication protocol, the team reduced the end-to-end delay from neural signal acquisition to command execution to less than 100 milliseconds. This response time falls below the body’s own physiological delay, contributing to smoother and more natural control.

The study also documented changes in neural activity over time. As Zhang became more proficient, task-related brain activity shifted from involving broad neural populations to being dominated by a smaller, more efficient subset of neurons. This reorganization reduced mental effort and made actions feel more intuitive.

**Also Read: **World’s Smallest Autonomous Robot That Can Sense, Think, and Act

Pu Muming, an academician of the Chinese Academy of Sciences and academic director of CEBSIT, said the trial confirmed the long-term safety and stability of implanted electrodes and neural signal decoding. He noted that these factors are essential for invasive brain–machine interfaces to advance toward clinical application.

The work has also drawn comparisons with Elon Musk’s “Telepathy” project at Neuralink, which has reported human trials focused on basic interactions like playing a video game and simple tasks. In contrast, the Chinese system has already been demonstrated in sustained daily use, including control of physical robotic devices.

Based on the results, the team has announced an upgraded system, WRS02, with 256 recording channels. The first clinical trial of the new system is expected to begin soon. Future applications may include decoding speech directly from brain signals.

While the study does not resolve the broader ethical, surgical, and scalability challenges associated with invasive BCIs, it demonstrates that stable, real-world use is possible. The findings suggest that brain–computer interfaces are beginning to move beyond proof-of-concept demonstrations toward practical assistive technologies for people with severe paralysis.

Sources: Chinese Academy of Sciences, South China Morning Post.