--- title: "Two standards in a single month: how China built the rulebook for brain–computer implants" date: 2026-10-04 category: Brain–Computer Interface site: NeuroAI canonical: https://neuroai.site/a/na-bci-china-first-bci-standards language: en --- # Two standards in a single month: how China built the rulebook for brain–computer implants > China's drug regulator published the country's first brain–computer interface (脑机接口) medical-device standards in September 2025, creating a fast-track pathway for implants to reach clinics. A researcher holds a wafer-thin electrode smaller than a fingernail. Across the city, an examiner runs a test that will decide whether the device is safe enough to enter a human skull. Until recently, no one in China had written down exactly how that test should work. That changed in 2025. Quietly, and faster than most observers expected, China's medical-device regulator turned a frontier technology into something with a paper trail: definitions, terminology, classification, and test methods. For a field where one wrong word in a filing can stall a product for years, the rulebook is the real bottleneck — not the science. ## Why a rulebook decides who gets treated Brain–computer interface (脑机接口) devices sit in an awkward legal space. A non-invasive headband that reads attention is a wellness gadget. The same headband used to treat a seizure is a medical device. An implant that writes to the motor cortex to let a paralyzed person move a cursor is, unambiguously, a high-risk implant. Without standards, every company argues its own case to every reviewer. Trials get redesigned mid-flight. Hospitals hesitate to adopt tools they cannot yet be sure are approved. A rulebook does not make the technology work — it makes the technology *legible* to the system that lets patients use it. ## The first two standards, both in September 2025 China's National Medical Products Administration (NMPA, 国家药监局) published two industry standards within weeks of each other: - **YY/T 1987—2025《采用脑机接口技术的医疗器械 术语》** (Terminology for medical devices using brain–computer interface technology), released on 16 September 2025. This is China's first brain–computer interface medical-device standard. It builds a systematic terminology framework — defining the technology's basic principles, paradigm types, signal forms, signal processing, and applications. China Daily described it as the country's "first standard for medical devices based on BCI technology." - A second standard, released on 28 September 2025 and effective from 1 October 2025, covers **《采用脑机接口技术的医疗器械 具备闭环功能的植入式神经刺激器 感知与响应性能测试方法》** (Test method for the sensing and feedback performance of interactive neurostimulators that use BCI technology). It is aimed squarely at the hardest category: closed-loop implantable neurostimulators that sense neural signals and automatically adjust stimulation. The method standardizes how to test sensing hardware (EEG, local field potentials, spikes), stimulation output, closed-loop response time, and the core intelligence algorithm. Notably, it draws on the U.S. FDA's guidance for physiologic closed-loop control. Two standards in a single month is a deliberate signal. Regulators did not wait for the market to mature; they built the track while the runners were still lining up. ## Classification came first Before the terminology standard, NMPA issued two guiding principles on 30 June 2025 — 《脑机接口医疗器械产品分类界定指导原则》 and 《脑机接口医疗器械通用名称命名指导原则》. These gave the first official definition of a brain–computer interface medical device and settled a question companies had been guessing at: - Non-invasive devices used for disease treatment or function compensation are managed as **Class III (三类)** medical devices — the strictest category. - Only a narrow exception qualifies as Class II: non-invasive devices used strictly for stroke limb rehabilitation, and only when they contain no artificial-intelligence algorithm. That single clause tells you the regulator's instinct: keep the bar high, carve out a thin, well-defined lane for simpler rehab tools, and push everything else into the most rigorous review path. ## The fast lane: innovative-device review Standards are necessary but not sufficient. A product still has to clear review. NMPA's **Innovative Medical Device Special Review Program (创新医疗器械特别审查程序)** lets promising domestic products get early, dedicated interaction with reviewers instead of a silent queue. The semi-invasive NEO system — co-developed by Tsinghua University and Bournaleaf (博睿康) — became the first brain–computer interface product to enter that program, in August 2024. Today, three invasive products from Shanghai alone are in the program. The point is not the count; it is that a defined on-ramp now exists, and companies are using it. ## Who actually runs the tests A standard is only as real as the lab that enforces it. The **Shanghai Institute of Medical Device Inspection (上海市医疗器械检验研究院)** handles more than 90% of China's implantable brain–computer interface type testing and led the compilation of the first national standard. In December 2025 it opened a full-chain brain–computer interface service platform covering micro-fabrication, long-term reliability, large-animal trials, and databases — a one-stop path from research to registration. This concentration is a feature, not an accident. Shanghai has openly built a "four-in-one" cultivation mechanism: a project-manager team, a priority task list, a future-industry fund, and a physical industry cluster. The standard is the legal keystone of that machine. ## What this means for a patient in 2026 For someone with spinal-cord injury, Drug-resistant depression, or epilepsy, the practical effect is speed and predictability: - A manufacturer knows which tests to pass before human trials, not after. - Hospitals can plan procurement around a recognized device class. - Investors get a clearer signal of which products can actually reach market. None of this makes an implant safe by itself. But it converts "maybe someday" into "here is the pathway, here are the milestones." ## Honest limitations This article covers *regulatory and standards architecture*, not clinical outcomes. It does not assess whether any specific Chinese brain–computer interface implant is safe or effective in the long term — those judgments require peer-reviewed trial data that, for most products, is still being collected. Figures on market size and financing are drawn from municipal and industry reports and should be read as directional, not audited. The NMPA framework described here applies to China; the U.S. (FDA) and EU (MDR/CE) pathways differ substantially, and a device approved under one is not automatically valid under another. I have not evaluated how aggressively the new closed-loop standard will be enforced, nor whether small entrants can afford compliance. ## What readers can do now - **If you follow medtech:** track NMPA's brain–computer interface standard releases as a leading indicator — each new standard (terminology, closed-loop, next: AI datasets, non-invasive general conditions) marks a product category moving from lab to clinic. - **If you are a researcher or clinician:** the June 2025 classification principles tell you exactly which brain–computer interface claims trigger Class III review — design studies and ethics protocols around that line early. - **If you are an investor:** treat "entered the Innovative Medical Device Special Review Program" as a more meaningful milestone than a press release, because it implies direct regulator engagement and a defined review path. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-bci-china-first-bci-standards Free to quote with attribution and a link to the original.