China’s First Commercial Brain Implant Shows Why Neurotechnology Needs Governance

Back in March, the National Medical Products Administration (NMPA) of China approved a brain implant for people with spinal cord injuries for commercial sale. Shanghai-based Neuracle Medical Technology developed NEO, a coin-sized implant that sits on the dura mater – the outermost layer of the brain. It has eight sensors that decode neural signals in real time. The device is paired with a pneumatic robotic glove, allowing patients with C2-C6 cervical spinal cord injuries to regain some hand motor function. This is the first invasive motor brain-computer interface in the world approved for routine clinical use outside a clinical trial.

What happened next is arguably more significant than the approval itself. Within 48 hours, China's National Healthcare Security Administration (NHSA) assigned the device a reimbursement billing code. That speed highlights something broader than regulatory efficiency: China has been building the administrative infrastructure needed to commercialize neurotechnology long before the first product reached the market.

Why NEO reached the market so quickly

China's investment in neurotechnology reflects a broader industrial strategy to move up the value chain: from a manufacturing-led economy toward one driven by high-value innovation. BCIs align with several national priorities: improving care for an aging population, strengthening China's domestic medical device industry, and building capabilities in adjacent sectors such as AI, robotics, and advanced semiconductors. Against that backdrop, it is less surprising that China has also invested in the regulatory and reimbursement infrastructure needed to bring these technologies to market quickly.

NMPA's approval mechanism for speed comes from its Innovative Medical Device Special Review Procedure. To qualify, an applicant must hold a Chinese patent over the device's core mechanism and demonstrate that the product's primary working principle is first-in-China. Eligible devices get priority testing and evaluation. In 2024, innovative devices were approved through the procedure in an average of 180 days, roughly half the standard review timeline. That's faster than the equivalent US FDA mechanism even in its best case, which would average just over 300 days.

On the reimbursement side, NHSA didn’t just happen to develop billing codes a couple of days after NEO was approved; the agency built the pricing infrastructure before there was any commercial BCI product in the market (and a year before the NEO device was approved). Chinese medical service pricing is organized around technology and procedure categories rather than named products, and NHSA has a standing process for defining those categories through expert consultation even before any specific device exists to fill them. Once NMPA clears a product, it just needs to be slotted into an already-defined pricing category, which is why NEO could be assigned a billing code within 48 hours of approval. In other words, the approval was the final step in a reimbursement framework that had already been designed.

The US FDA regulatory approach

In the United States, the Food and Drug Administration (FDA) regulates medical devices, including BCIs. While there is no BCI-specific pathway, developers may receive expedited interaction with regulators through programs such as the Breakthrough Devices designation, which is intended for technologies addressing serious or life-threatening conditions. This grants more frequent, structured engagement with agency reviewers for technologies treating life-threatening or irreversibly debilitating conditions, resulting in shorter BCI approval timelines.

For example, in April 2026, CorTec received Breakthrough Device Designation for its Brain Interchange System, a closed-loop implant that reads and stimulates the brain to support motor recovery in chronic stroke patients. Breakthrough Device Designation isn't an approval nor is it a shortcut around the FDA's evidence standards. Devices still have to go through the same full 510(k), De Novo, or PMA review as anything else. What changes is the process around that review: designated devices get priority placement in the review queue, dedicated senior reviewers, and more frequent structured meetings with FDA staff throughout development, rather than the more limited, scheduled touchpoints companies get by default. Closer and earlier FDA engagement lets a small team catch a flawed trial design or an insufficient evidence package months before a rejection would otherwise surface those issues.

However, this only speeds up the device approval process. The US, unlike China, does not have a cohesive or efficient pathway to determine coverage for BCI medical devices. However, soon after the approval of and billing code assignment for NEO, the US Centers for Medicare and Medicaid Services and the FDA jointly announced something called the RAPID coverage pathway. It aims to synchronize FDA market authorization with Medicare national coverage determinations for Class II and Class III Breakthrough Devices.

Both countries have demonstrated that they can accelerate regulatory review for promising BCIs. The difference is that China also built much of the reimbursement infrastructure before the first commercial product existed. The US is now reactively building that infrastructure. Regulatory approval allows a device to be sold. Reimbursement determines whether hospitals adopt it and whether patients can realistically access it.

Beyond approvals – the governance questions that remain

Both approval processes answer one question: is this device safe and effective for its intended medical use? Neither touch on what distinguishes BCIs from most medical devices: their ability to collect and increasingly interpret neural activity.

As decoding algorithms improve, those signals may reveal increasingly rich information about movement intentions, emotional states, cognitive impairment or other aspects of mental activity. While today's clinical BCIs remain task-specific and cannot "read thoughts" in the science fiction sense, many researchers argue that neural data deserves special treatment because it is both biologically intimate and increasingly valuable as artificial intelligence models improve.

Neither approval process is designed to determine who controls neural data after collection; whether it may be reused to train AI systems; how long it may be retained; whether inferences derived from neural activity deserve protections beyond ordinary health information; or whether individuals should have distinct legal rights over their brain-derived data. Those questions largely fall outside the scope of traditional medical device approval and remain only partially addressed by existing privacy, data protection and health laws.

There is a growing conversation on neuro-rights and cognitive liberty that focuses not only on device safety, but also on frameworks that protect mental privacy. The OECD has argued that existing regulatory systems were built to evaluate medical devices rather than govern the broader societal implications of brain-computer interfaces, recommending a combination of standards, corporate governance, and international coordination while more comprehensive legal frameworks develop. Right now, there is some state legislation in the US that protects neural data. Otherwise, neurotechnology continues to be governed through a patchwork of medical device law, privacy law, cybersecurity requirements, research ethics, and general data protection legislation.

Closing that gap doesn't require waiting for a comprehensive federal neurorights statute, though. It needs infrastructure and coordination models regulators have already built for other purposes, earlier in the process than they've been willing to so far.

NHSA's approach to NEO worked because the pricing categories already existed before the device did. CMS could take the same posture toward neural-data-generating devices generally: defining technology and procedure categories in advance of specific product approvals, rather than building coverage pathways reactively once a device has already cleared FDA.

The harder gap is that FDA's device approval was never designed to touch data governance in the first place. However, there is precedent in place: FDA and CISA have coordinated on medical device cybersecurity through a formal agreement since 2018. That model could be widened rather than reinvented by bringing the FTC and HHS into the existing FDA-CISA structure, so that neural data reuse, retention, and AI training rights are governed by the same kind of standing coordination that already exists for device cybersecurity.

Consent is the piece closest to patients directly. Right now, a single signature at implantation typically covers clinical use, research use, and any downstream commercial or AI-training use of a patient's data; three decisions collapsed into one moment. Separating them would let patients consent to each independently. And because BCIs collect data continuously over months or years, unlike a single scan or blood draw, consent arguably shouldn't be a one-time event either; patients need some mechanism to revisit or revoke it as collection continues.

None of this requires Congress to first agree on what neural data is in a legal sense. Speed is not the part of this story the US (or other jurisdictions) need to copy. Rather, it’s foresight. China built reimbursement infrastructure before a product existed to use it. Regulators can do the same for the reimbursement and data questions, if they start now instead of after the next approval.