Imagine moving a wheelchair or commanding a robot dog using only your thoughts. This year, researchers in the Chinese mainland have turned this vision into reality with an invasive brain-computer interface (BCI) trial that goes beyond virtual screens.
BCI technology spans three categories: invasive, semi-invasive and non-invasive. The latest breakthrough comes from the Chinese Academy of Sciences Center for Excellence in Brain Science and Intelligence Technology, in partnership with Huashan Hospital affiliated with Fudan University.
In March 2025, ultra-thin electrodes – each thinner than 1 percent of a human hair – were implanted into the motor cortex of a middle-aged participant with quadriplegia. The participant soon steered a wheelchair outdoors and directed a robotic dog to fetch takeout using only neural signals.
Recently, a second participant received the implants, unlocking new applications and paving the way for broader clinical use.
Key technical advances fuelled this leap into real-world control:
- Neural data compression and hybrid decoding: Boost system performance by 15-20 percent.
- Neural manifold alignment: Ensures stable interpretation of fluctuating signals amid emotional and environmental changes.
- Online recalibration: Allows real-time adjustments without interrupting daily tasks.
- Ultra-low latency: Thought-to-action delay reduced to under 100 milliseconds, outperforming the average human reaction time of 200 milliseconds.
By integrating external devices as seamless extensions of the body, this trial marks a shift from "two-dimensional screen" interactions to "three-dimensional" actions in dynamic environments. Participants can now navigate obstacles, execute emergency stops and interact intuitively with assistive technologies.
As clinical trials continue, these developments promise to redefine assistive care and open new horizons in neurotechnology for people with motor impairments worldwide.
Reference(s):
Chinese scientists make breakthrough in brain-computer interface trial
cgtn.com




