People living with severe paralysis from spinal cord injuries now have access to a commercially approved brain-computer interface for the first time. China’s National Medical Products Administration (NMPA) has granted Class III medical device registration to NEO, an epidural brain chip developed with Fudan University researchers and implanted at Huashan Hospital. The device restores hand movements and allows users to control soft robotic hands, and its clearance for sale outside clinical trials puts it ahead of every competitor, including Elon Musk’s Neuralink, which remains restricted to research settings.
Why China cleared the world’s first brain chip matters now
The NEO device is not simply another research prototype. Its NMPA registration as a Class III device, the highest-risk category under Chinese medical device law, means it can be legally marketed and prescribed to patients. That distinction separates it from brain-computer interfaces in the United States and Europe, where devices from companies like Neuralink are still limited to investigational trials with small participant pools and no commercial pathway in sight.
The practical effect is that Chinese hospitals can now offer the NEO system to eligible patients with severe paralysis, generating real-world safety and performance data at a pace that trial-restricted Western programs cannot match. Over the next two years, this head start could produce the largest body of post-market evidence for any brain-computer interface, effectively setting the reference standard that regulators in other countries will use when evaluating their own approvals. If the device performs well across a broader patient population, the existing device registration framework will have created a de facto global safety benchmark before the U.S. Food and Drug Administration or European Medicines Agency clears a comparable product for sale.
For patients, the difference between a trial slot and a registered product is not abstract. Trial participants face strict enrollment criteria, geographic constraints, and the possibility that the study ends before they see lasting benefit. A registered device, by contrast, can be prescribed by any qualified clinician at an approved facility, expanding access far beyond the handful of academic centers that typically run clinical studies. It also opens the door to more routine follow-up care, structured rehabilitation programs, and, potentially, reimbursement pathways within China’s health system.
NEO’s clinical record and the institutions behind the approval
The supporting clinical evidence for NEO comes primarily from a medRxiv preprint documenting nine months of continuous use by a patient with complete spinal cord injury. During that period, the epidural minimally invasive brain-computer interface enabled the participant to regain functional hand movements and operate a soft robotic hand for daily tasks such as grasping and manipulating household objects. The preprint, identified by DOI 10.1101/2024.09.05.24313041, remains the most detailed public dataset describing how the system performs outside a tightly controlled laboratory environment.
Fudan University confirmed that the implant surgery took place at Huashan Hospital, one of China’s leading neurosurgical centers. The university’s institutional release named the device NEO and described the functional gains achieved by the patient, anchoring the regulatory filing to a specific clinical site and surgical team rather than leaving the approval as an abstraction. That level of institutional attribution matters for a first-in-class device, because it ties the technology to surgeons and engineers whose outcomes can be independently followed as more patients receive implants.
Nature Biotechnology has independently reported on the approval and classified it as the world’s first commercial authorization for a brain-computer interface, distinguishing it from earlier regulatory actions that permitted only research or trial-stage implantation. In that coverage, the journal highlighted how NEO’s clearance contrasts with competitors like Neuralink, which, according to recent analysis, are still navigating early-stage clinical studies and have not yet demonstrated long-term performance in broader, real-world populations.
The technical approach itself is central to understanding the device’s risk–benefit profile. NEO sits on the surface of the dura mater, the tough membrane covering the brain, rather than penetrating brain tissue directly. This epidural placement reduces surgical risk compared to intracortical arrays that require electrodes to be inserted into the cortex. In theory, avoiding penetration lowers the chances of hemorrhage, inflammation, and long-term scarring around the electrodes.
The tradeoff is typically lower signal resolution, because the electrodes are farther from the neurons that generate the electrical activity being recorded. Yet the nine-month usage record described in the preprint suggests the system captures enough neural activity to drive meaningful motor restoration. The patient was able to execute grasp-and-release motions, coordinate finger movements, and control a soft robotic assistive device with sufficient precision to perform daily tasks, indicating that the decoding algorithms and hardware design can compensate for some of the signal loss inherent to epidural placement.
Gaps in the evidence and what to watch next
Several questions remain open despite the approval. The medRxiv preprint has not yet appeared in a peer-reviewed journal, which means the clinical data have not been subjected to independent expert scrutiny through formal publication. The preprint documents outcomes for a single patient over nine months, a timeframe and sample size that demonstrate feasibility but do not establish long-term durability or reveal how the device performs across different injury levels, ages, or body types. Key parameters such as stability of signal quality over years, infection risk, and hardware reliability remain largely uncharacterized.
The full text of the NMPA registration certificate and its attached technical requirements has not been publicly linked in any of the primary sources. Without that document, outside researchers and regulators cannot independently verify the specific indications, contraindications, or post-market surveillance conditions that China’s regulator attached to the approval. It is unclear, for example, whether the device is formally indicated only for upper-limb motor restoration in complete spinal cord injury, or whether the label is broader and could encompass other forms of paralysis.
No official NMPA announcement or decision document directly quoting the approval date or the precise approved indications has surfaced in the public record. Instead, the picture must be reconstructed from institutional press releases and secondary reporting. That lack of direct regulatory documentation complicates efforts to compare NEO’s approval conditions with those that might be imposed by the U.S. Food and Drug Administration or European regulators on similar devices.
Direct statements from the implanting surgical team or the hospital ethics committee regarding plans for broader commercial implantation are also absent from available reporting. That gap leaves open the question of how quickly Huashan Hospital or other facilities will begin prescribing NEO to new patients, and under what clinical protocols. It is not yet known whether early commercial use will be clustered at a few high-volume neurosurgical centers or distributed more widely across provincial hospitals.
The absence of a corresponding clinical trial registration number in the source trail adds another layer of uncertainty. Registered trial identifiers allow independent observers to track enrollment criteria, endpoints, and adverse event reporting over time. Without such a listing, external experts have limited visibility into how new patients will be selected, what functional outcome measures will be used, and how complications will be documented as the device moves from a single preprint case to routine clinical practice.
For now, NEO’s approval marks a turning point rather than a conclusion. China’s decision effectively shifts the frontier of brain-computer interface development from the lab to the clinic, where questions of equity, safety, and long-term benefit will be answered in real patients rather than animal models or short-term trials. The speed and transparency with which those answers emerge will determine whether NEO becomes a durable standard of care for paralysis-or a cautionary example of moving to market before the evidence base is fully mature.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
