In a rehab room in central Sichuan, a woman who could not lift her arm imagines the motion. A band of sensors on her scalp catches the attempt.
No surgery, no incision. Just a closed loop of intention, machine, and movement, repeated until the brain begins to rewire itself.
This is no longer an experiment behind glass. At Mianyang Central Hospital, it has become a scheduled outpatient service that patients can actually book.
Key takeaways
- Mianyang Central Hospital opened a BCI rehab clinic and treated over 100 patients since March 2025 (hospital reports).
- Sichuan's 2025–2030 plan targets >100,000 patient-visits/year and >20,000 rehab-device uses/year by 2030 (Chengdu gov).
- The plan aims for 3 invasive and 5 non-invasive BCI products registered by 2027 (Chengdu gov).
- West China Hospital completed western China's first implantable BCI surgery in July 2025 (local media).
- Sichuan is among the first provinces with a dedicated BCI industry policy (SCTV).
- A 2025 expert consensus covers portable exoskeleton BCI for home stroke rehab (hospital).
From research showcase to outpatient slot
Mianyang Central Hospital is not a coastal flagship. It is a provincial hospital in China's only state-approved science city, and in 2026 it opened a brain–computer interface (脑机接口) rehabilitation clinic for people with stroke, brain injury, and spinal-cord injury. The service is non-invasive: electrodes on the scalp, a decoder that reads motor intent, and external devices that turn the intent into motion (China Women's News; People's Daily public platform).
The clinical team did not start cold. Since March 2025, the rehabilitation department had already been running non-invasive brain–computer interface treatment, and by 2026 it reported more than 100 cases across stroke, brain injury, and spinal-cord injury (Mianyang Central Hospital, via multiple local outlets). That accumulation, not a single headline case, is what makes the clinic notable.
One patient, a 54-year-old woman left hemiplegic and speech-impaired after head trauma, could not turn or feed herself on admission. After about two months of systematic treatment she raised her arm again and could call out family members' names (China Net; People's Daily Sichuan). The hospital presents her as illustrative, not as a guaranteed outcome.
What the clinic actually does
The mechanism is a closed loop. The system captures and decodes the patient's electroencephalogram, identifies the intended movement, and drives an external device — functional electrical stimulation, an exoskeleton robot, or virtual reality — to perform it. The patient must actively imagine the motion; the machine supplies the assist. That loop is meant to activate neuroplasticity and shift rehab from passive to active (Mianyang Central Hospital).
The clinic pairs the interface with non-invasive neuromodulation:
- Repetitive transcranial magnetic stimulation, which adjusts cortical excitability through pulsed magnetic fields.
- Transcranial direct-current stimulation, a gentle current that helps the brain regain control of a limb.
- Combined "magnetic-electric" coordination, intended to produce a greater-than-additive rehab effect.
The department built a physician–therapist–nurse team, and staff trained through national brain–computer interface certification programs from 2024 (China Women's News). One deputy chief physician co-authored the Expert Consensus on Portable Exoskeleton Brain–Computer Interface for Community and Home Upper-Limb Stroke Rehabilitation (2025 Edition), signaling an interest in moving care beyond the hospital wall (Mianyang Central Hospital).
A province with a plan
The hospital's move sits inside a larger Sichuan push. In May 2025, eight provincial departments issued the Sichuan Brain–Computer Interface and Human–Computer Interaction Industry Breakthrough Action Plan (2025–2030). It targets, by 2027, three invasive and five non-invasive brain–computer interface products completed and registered as medical devices, the province's first invasive implantation surgery, and more than 50,000 annual medical patient-visits (Chengdu government; SCTV).
By 2030 the plan is more ambitious: about 10 chain-leader enterprises, 100 specialized "little giant" firms, and 200 innovative small and medium enterprises; roughly 3,000 invasive implantation surgeries per year; and brain–computer interface products serving patients with neurodegenerative, psychiatric, and addiction conditions at more than 100,000 visits per year, with rehab devices applied in more than 20,000 cases per year (Chengdu government). Sichuan describes itself as among the first provincial-level jurisdictions to issue a dedicated brain–computer interface industry policy (SCTV).
West China Hospital in Chengdu completed western China's first implantable brain–computer interface surgery for a spinal-cord-injury patient in July 2025, showing the province is also pushing the invasive frontier (China Net). And Sichuan has moved to set government-guided prices for brain–computer interface medical services, following the national and Hubei pricing steps earlier in 2025 (People's Daily Sichuan, 2026).
The bottleneck is not the electrode
The technology works well enough to treat people today. The harder problem is operational: who is trained, how sessions are scheduled, what payers reimburse, and how a clinic sustains volume without a research grant. According to Li An, chief scientist at BrainNet (脑机网), the harder barrier for brain–computer interface is not the electrode but the hospital workflow around it. That reading fits Mianyang's story — the clinic's real achievement is process, not a breakthrough device.
The contrarian note is that "over 100 cases" is a hospital-reported count, not an independently audited trial with controls. Non-invasive rehab interfaces show promise for motor recovery, but the field still lacks large randomized evidence on durable gains. A clinic is a beginning of proof, not proof itself.
Honest limitations
The patient count and case descriptions come from hospital and local-media reports, not a peer-reviewed controlled study. Sichuan's 2025–2030 targets are provincial planning goals, not guarantees, and depend on devices clearing registration. The "over 100 cases" figure varies slightly across outlets (some say nearly 100); it should be read as approximate. Whether insurers reimburse these sessions in Sichuan was not detailed in the sources reviewed.
What readers can do now
- Operators: build the training and scheduling workflow first; the device is the easy part of a rehab clinic.
- Investors: watch provincial patient-visit targets as a demand signal once devices are registered and priced.
- Researchers: design controlled studies on non-invasive BCI rehab durability; anecdote alone will not scale reimbursement.