NeuroAI NEUROAINEUROAI.SITE
ESC

Beinao-1 (北脑一号): 128-channel wireless implant returns movement and speech to paralyzed patients

Beijing Brain Institute's semi-invasive wireless BCI has been implanted in six patients, decoding speech for the aphasic and restoring limb control after spinal injury.

2026-10-23 · 963 words · NeuroAI
Beinao-1 (北脑一号): 128-channel wireless implant returns movement and speech to paralyzed patients

A 67-year-old woman who had lost the ability to speak because of ALS (肌萎缩侧索硬化症) sat in front of a screen. She thought the words "I want to eat" (我想吃饭), and the characters appeared. No keyboard, no voice — only a thin implant resting on the surface of her brain.

That scene is not a demo video from a lab far away. It is one of the early human cases of Beinao-1 (北脑一号), a home-grown brain-computer interface (脑机接口) built by the Beijing Brain Science and Brain-Inspired Research Institute (北京脑科学与类脑研究所) together with its spin-off Beijing Xinzhida NeuroTech (北京芯智达神经技术有限公司).

What Beinao-1 actually is

Beinao-1 is a semi-invasive (半侵入式), wireless, fully-implanted BCI system. "Semi-invasive" means the hardware sits beneath the skull but above the dura mater (硬脑膜) — the protective membrane around the brain — rather than piercing brain tissue. The core sensing piece is a flexible, high-density electrocorticography (ECoG) electrode thin enough to be described as cicada-wing-like and smaller than half a palm, which records neural activity across 128 channels.

The internal unit has no battery of its own. It is powered and communicated with through an external wearable device using short-range wireless transmission, which removes the need for a follow-up surgery to replace a depleted battery. According to the institute, the effective channel rate after implantation exceeds 98%, meaning almost all of the 128 pathways keep working.

Luo Minmin (罗敏敏), director of the Beijing Brain Institute and the trial's lead scientist, has framed the design choice plainly: placing electrodes on the dura captures enough signal to decode intent while avoiding the higher surgical risk of driving electrodes into the cortex.

The clinical results so far

The first batch of human implants took place in February and March 2025, initially in five patients, and by early 2026 six people had received the system. The reported cases span three hard problems:

  • Speech. The ALS patient who typed "I want to eat" is described as the first person worldwide to recover language decoding through a wireless, fully-implanted BCI. The system turns imagined speech into on-screen Chinese characters.
  • Upper-limb control. A 30-year-old quadriplegic patient, identified as Xiaorui (小锐), was implanted at Peking University First Hospital (北京大学第一医院) in February 2025. Within days he was using intention to drive a robotic arm to bring a cup of water to his mouth. After about six months of training he could grasp wooden blocks and pinch small glass balls; eleven months in, his upper-limb function kept improving.
  • Walking. A patient paralyzed for five years by spinal-cord injury regained the ability to stand and walk with unpowered braces and a walker after brain-guided rehabilitation. Another high-cervical injury patient learned to brain-control a cursor and play simple games while recovering partial upper-limb movement.

The project was selected among the 2025 China and World Top Ten Science and Technology Advances, voted by members of the Chinese Academy of Sciences and Chinese Academy of Engineering.

Why "wireless" and "no battery" matter

Most invasive BCI reports focus on the electrode. Beinao-1's quieter innovation is the system around it: a miniaturized host, a new-generation short-range wireless link, and real-time decoding algorithms that run across multiple scenarios. The absence of an internal battery is a real clinical advantage — implanted pulse generators that rely on batteries eventually need replacement surgery, and every return to the operating room carries infection and complication risk. A wirelessly powered design pushes that risk out of the picture for the life of the implant.

Beinao-2 and the road to market

The same institute is developing Beinao-2 (北脑二号), a fully invasive system with a 1,024-channel wired interface. In 2024 it was reported to have let a monkey intercept a moving target using intention alone — a step toward higher-resolution invasive recording.

Clinically, Beinao-1 is moving from research toward product. A procedure staged in March 2026 was described as the final full-process rehearsal before the system enters a National Medical Products Administration (NMPA, 国家药监局) registration trial. In March 2026, China's government work report mentioned brain-computer interfaces (脑机接口) as a future industry for the first time, a signal that the regulatory and industrial track is being laid in parallel with the science.

Trend reading

According to Li An, chief scientist at BrainNet (脑机网), the real bottleneck for Chinese BCI is no longer the electrode but the closed-loop algorithm and the long-term biocompatibility data, which take years to accumulate before any product can claim durability.

Honest limitations

This article relies on institute statements, Chinese state-media reporting, and conference disclosures rather than peer-reviewed publications of the full datasets. The six implanted patients are a small sample, and the longest follow-up described is on the order of a year or so — far short of the multi-year evidence regulators and clinicians will want. The speech-decoding claim, while striking, has not been independently replicated in a published trial that we could locate. Beinao-1 is also a research system, not yet a commercially approved device; timelines to market depend on the NMPA registration trial, which had not reported results at the time of writing. We have not covered the full list of competing approaches — invasive, non-invasive (非侵入式), and other semi-invasive programs — in depth, and cost, reimbursement, and equitable access remain open questions.

What readers can do now

  • If you or a family member has a spinal-cord injury, stroke, or ALS-related communication loss, ask treating neurologists whether a hospital near you runs BCI or neuromodulation evaluations; Beijing Tiantan Hospital (北京天坛医院) and major teaching hospitals have opened assessment clinics.
  • Treat any "cures paralysis" claim on social media with skepticism; today's BCI restores specific, trained functions, not whole-body recovery, and candidacy depends on strict age and medical criteria.
  • Follow NMPA (国家药监局) device-approval announcements directly if you track medical-device investment or clinical access, since the registration trial result will be the watershed event.

Related coverage

Cite this article

NeuroAI (脑机网), "Beinao-1 (北脑一号): 128-channel wireless implant returns movement and speech to paralyzed patients," 23 October 2026, https://neuroai.site/a/na-brain-beinao1-wireless-bci

Machine-readable: BibTeX · RIS · JSON · Markdown

More in “Brain–Computer Interface” → · Back to home · Markdown version