https://www.cas.cn/syky/202512/t20251217_5092713.shtml
According to this article, progress has been made in Brain-Computer Interface Applications in everyday life.
Thus, the CAS Center for Excellence in Brain Science and Intelligence Technology, and other institutions completed their second clinical trial of an invasive brain-computer interface (WRS01) to enable a paraplegic patient to stably control a smart wheelchair and a robotic dog via EEG signals, achieving autonomous movement and object retrieval in real-life scenarios.
This patient suffered high-level paraplegia due to spinal cord injury in 2022 and received the brain-computer interface system developed by the Center for Excellence in Brain Science and Intelligence Technology in June 2025. After several weeks of training, the patient was able to stably control a computer cursor and a tablet computer.
The team further expanded the system to control three-dimensional physical devices, achieving continuous, stable, and low-latency control of the smart wheelchair and robotic dog, helping the patient complete multiple functional activities in complex daily scenarios.
There were a series of technological breakthroughs. In the neural information extraction stage, the team developed a high-compression-ratio, high-fidelity neural data compression technology and innovatively integrated data compression methods such as peak frequency power, adjacent pulse interval, and peak pulse counting. This hybrid decoding model can still efficiently extract effective signals even in noisy environments, improving overall brain control performance by 15% to 20%.
To address signal instability caused by acoustic, optical, and electromagnetic interference in real-world environments, as well as fluctuations in patients’ physiological and psychological states, the team introduced neural manifold alignment technology. This technology extracts stable low-dimensional features from high-dimensional dynamic neural signals, enhancing the decoder’s environmental adaptability and cross-day stability.
The team simultaneously developed online recalibration technology, allowing patients to fine-tune decoding parameters in real-time during daily use without interrupting operations for specialized calibration. This ensures consistently high system performance and a user experience that becomes smoother with use.
The research, through a custom communication protocol, compressed the end-to-end latency from signal acquisition to command execution to less than 100 milliseconds, below physiological latency levels, resulting in a smoother and more natural control experience for patients.
The research team discovered that as patients become more proficient with brain-controlled peripherals, task-related neural activity gradually shifts from widespread neuronal involvement to being dominated by a few highly efficient neurons. This reduces cognitive burden and achieves “internalized” control of the peripherals, explaining the formation process of “voluntary control” from a neural mechanism perspective.
Regarding application expansion and social integration, the team collaborated with local disabled persons’ federations to guide patients in participating in online data annotation and other tasks.