China has made new progress in invasive brain-computer interface (BCI) tehcnology, achieving complex real-world operations such as mind-controlled wheelchairs.
BCI is categorized into three types: invasive, semi-invasive, and non-invasive. The new progress in the invasive BCI was made by the Chinese Academy of Sciences (CAS) Center for Excellence in Brain Science and Intelligence Technology, in collaboration with Huashan Hospital affiliated with Fudan University and other institutions. A middle-aged man with quadriplegia caused by a cervical spinal cord injury can now steer a wheelchair for a stroll and command a robotic dog to retrieve takeout food using only his thoughts, a major breakthrough by Chinese scientists in clinical trials of invasive BCI technology.
In March, the CAS Center for Excellence in Brain Science and Intelligence Technology, together with Huashan Hospital, implanted electrodes smaller than 1 percent of the diameter of a human hair into the brain of a patient with motor dysfunction, enabling mind-controlled chess playing and car racing. This was China's first clinical trial of an invasive BCI.
Currently, the second participant in the BCI clinical trial has also successfully undergone implantation, enabling new applications.
Previously, BCI technology was largely limited to "two-dimensional screen" tasks such as cursor control. This time, the research team has advanced beyond that, moving patients from virtual screen interactions to performing "three-dimensional actions" in the real world.
In real world settings, patients controlling wheelchairs with their thoughts must navigate complex, dynamic environments, such as avoiding obstacles and executing emergency stops, with zero tolerance for delay between intention and action.
This demands that external devices like wheelchairs become an integrated extension of the patient's own body. To achieve this, Chinese scientists have overcome multiple core technical challenges in BCI development. First, the team developed neural data compression technology and a hybrid decoding model. Their efficient coordination enhances the overall performance of the brain control system by 15 percent to 20 percent. Second, for stability in interpreting fluctuating signals, the team created a "neural manifold alignment" technique. This ensures the decoder consistently reads the core intent of the neural signals, which are subject to interference from factors like emotional state and environmental changes. Third, to achieve self-evolving capability, the team developed an "online recalibration" technology. Unlike traditional BCI systems that require users to stop for tedious calibration, this new system allows for real-time parameter adjustments during daily use.
Fourth, regarding the synchronization rate from "thought to action," the team reduced the system delay to under 100 milliseconds for controlling external devices.
This is lower than the approximately 200 milliseconds it takes for a normal person's brain command to translate into a physical action.
It is these breakthroughs that have enabled the patient implanted with the latest electrodes to interact with the real world using his thoughts.
The patient now works as an online auditor for vending machine, participating in online data annotation tasks, becoming China's first patient to achieve online employment with the aid of an invasive brain-computer interface system.
Chinese scientists achieve breakthrough in invasive brain-computer interface trial
