- China approved YY/T 2029–2026 for EEG datasets used by AI-enabled BCI medical devices.
- The reported scope includes collection, processing, labeling, storage and access.
- The standard is scheduled to take effect on September 1, 2027.
The standard addresses the data behind the device
China's National Medical Products Administration has approved an industry standard covering the quality and evaluation of electroencephalography datasets used in AI algorithms for brain-computer-interface medical devices, TechNode reported, citing China Food and Drug News. The identifier is YY/T 2029–2026, with implementation scheduled for September 1, 2027.
The reported scope runs from collecting and processing EEG data through labeling, storage and access. The announcement concerns the evidence used to develop and evaluate a device's algorithm; it does not announce approval of a particular BCI product.
Why dataset quality is a separate question
An AI system can only be evaluated meaningfully if its test data represents the task it is supposed to perform. For a BCI dataset, that makes the circumstances of recording and the meaning of each label consequential. A number describing algorithm accuracy is hard to interpret when the recording procedure, intended user population or task definition is unclear.
A common framework can make those questions more explicit for developers and evaluators. It can also help distinguish a dataset suitable for a research experiment from evidence offered to support a medical-device application. Those uses may overlap, but one does not automatically establish the other.
What the announcement does not establish
The source summary does not provide the complete standard or its detailed evaluation thresholds. It would be premature to claim that a specific dataset or product already complies. Nor does a data-quality standard demonstrate a treatment's clinical benefit.
For teams following the transition, the practical next document is the full standard, followed by any regulator guidance on how it will be used in submissions and testing. The September 2027 date provides a planning reference, while the product-specific evidence requirements still need to be read in their applicable regulatory context.
Sources & context
Go to the original material. Company claims remain attributed to their sources.
01Updates & corrections
— Expanded with source reporting, context and a clearer distinction between announced plans and demonstrated results.
