Today, the Center for Excellence in Brain Science and Intelligent Technology of the Chinese Academy of Sciences announced new progress in the second invasive brain-computer interface clinical trial conducted by the center in cooperation with domestic scientific research institutions and medical units. This clinical trial has technically achieved a major transition from two-dimensional screen cursor control to three-dimensional physical world interaction.
The patient who underwent this brain-computer interface clinical trial was a middle-aged male patient. Due to an unfortunate fall, the patient suffered a spinal cord injury that resulted in quadriplegia in 2022. After more than a year of rehabilitation, his condition has not improved, and only his head and neck can move. In June 2025, the patient was implanted with the brain-computer interface system developed by the scientific research team. Initially, after 2 to 3 weeks of training, the patient was able to control computer cursors, tablets and other electronic devices with his thoughts. This was also the behavioral level achieved by the research team's first clinical trial of an invasive brain-computer interface. In order to further improve the implanter's ability to interact with the surrounding environment, the research team has successfully expanded the brain-computer interface application scenario from a two-dimensional screen to a three-dimensional physical world through the introduction of more new technologies. At present, the system has enabled users to achieve operation speeds close to ordinary people using mobile phones and computers through the "thoughts" of their brains, as well as the ability to initially control embodied intelligent robots.
The invasive brain-computer interface consists of two parts, the front-end sensor and the back-end processor. The sensor at the front is only about one hundredth as thick as a hair. The sensor is embedded into the brain of the implanter about 5 to 8 mm, and the skull of the implanter is thinned by 3 to 5 mm, and then the back-end processor is embedded. The whole process is a minimally invasive process.
Experts say that the front-end sensors are equivalent to the network cables connecting the brain, responsible for connecting to the outside world and uploading and downloading various information. The back-end processor is responsible for converting these weak neural activities of the brain into digital signals, which is a language that the machine can understand. In this way, the implanted person can control external devices through their thoughts and assist their life.
It is understood that continuous, stable, and low-latency precise control are the main features of the invasive brain-computer interface system released this time. In order to achieve these goals, the scientific research team developed high compression ratio and high fidelity neural data compression technology, and innovatively integrated several data compression methods: "peak frequency band power adjacent pulse interval" and "spike counting". This hybrid decoding model can efficiently extract effective information even in an environment with relatively noisy neural signals, improving the overall brain control performance by 15% to 20%.
In addition, the scientific research team has also conquered key core technologies such as "stable neuropopulation alignment across the sky" and "online recalibration", so that the system can fine-tune decoding parameters in real time and silently during daily use of the patient, making it more comfortable for the implanter to use it. At the same time, the end-to-end delay of this system from signal collection to command issuance to peripherals is also compressed to less than 100 milliseconds, which is lower than the human body's own physiological delay. This makes the patient's control experience smoother, and thoughts and actions are almost synchronized. On this basis, the scientific research team is currently still studying more application scenarios to adapt to the different needs of implant recipients.
Pu Muming, academician of the Chinese Academy of Sciences and academic director of the Center for Excellence in Brain Science and Intelligent Technology of the Chinese Academy of Sciences, said that it has been confirmed that the electrodes are safe and long-term stable in the brain, and that the recording and decoding of signals are also stable. This is a necessary step for invasive brain-computer interfaces to move toward practical medical applications. In the future, related technologies will expand to more applications, such as decoding language information in the brain.