Neural implants: the regulatory road

The gap between a working prototype and an approved medical device is the part of neurotechnology that receives the least attention and determines the most.

Device classification

Regulators classify medical devices by risk. An implanted device that interfaces with the central nervous system sits in the highest category everywhere.

In the United States that generally means the premarket approval pathway rather than the lighter clearance route used for devices substantially equivalent to something already marketed.

Premarket approval requires demonstration of safety and effectiveness through clinical evidence, manufacturing inspection, and detailed labelling review. It is the most demanding pathway the agency operates.

The staged route

Preclinical work. Bench testing, biocompatibility, and animal studies establishing that the device does what is claimed and does not cause unacceptable harm.

For chronic implants this includes long-duration animal implantation, because the failure modes that matter are the slow ones.

An investigational device exemption, permitting human study of an unapproved device. This requires the preclinical package, a protocol, and institutional review board approval.

Early feasibility study. A small number of participants, primarily assessing safety and gathering enough information to refine the design.

This is where the current generation of implanted BCI efforts sits.

Pivotal trial. A larger study designed to demonstrate effectiveness against a prespecified endpoint.

Designing this for a BCI is genuinely difficult. What is the control condition? Sham surgery is ethically fraught. What is the endpoint — communication rate, independence, quality of life?

Approval, then post-market surveillance, which for a permanent implant continues indefinitely.

Why it takes years

The evidence that matters is long-term.

An implant that performs beautifully for six months tells you very little about a device intended to stay in a person's head for decades.

Chronic tissue response, electrode degradation, hermetic seal integrity, battery life and material breakdown all play out over years.

Which means the trial duration cannot be compressed by adding participants or funding. It is set by biology.

The specific safety questions

Surgical risk — infection, haemorrhage, seizure — quantifiable from neurosurgical experience and non-trivial.

Chronic infection risk, particularly for percutaneous designs where something crosses the skin. Fully implanted wireless systems avoid this and add complexity.

Explantation. An underdiscussed requirement. Devices must be removable, and tissue integration can make removal itself hazardous.

Regulators ask what happens if the company fails and support ends — a question the field has faced concretely, with participants in discontinued trials left with implanted devices no longer supported.

Software updates to an implanted device that adjusts neural stimulation raise questions the framework was not originally designed for.

Cybersecurity, now an explicit part of device submissions.

The ethical layer

Beyond safety, running alongside it.

Informed consent for a participant with severe disability who may see the trial as their only route to communication is a genuinely difficult problem. The therapeutic misconception — believing a study is treatment — is well documented.

Data. Neural recordings are intimate. Who owns them, who can access them, and what secondary use is permitted are unresolved questions, and several jurisdictions have begun legislating on neural data specifically.

Identity and agency. Reports from participants with deep brain stimulation for other conditions include changes in mood and sense of self. This is a real category of effect that requires monitoring.

Access. If these devices work, they will initially be extremely expensive. Who receives them is a distributive question that the approval process does not address.

Reimbursement, after approval

The step that decides whether an approved device reaches patients.

Approval establishes that a device may be sold. It does not establish that any health system will pay for it.

Coverage decisions rest on cost-effectiveness evidence gathered in larger populations over longer periods than a pivotal trial provides, and they can take years after approval.

For a high-cost implant serving a small population, this is a substantial commercial risk, and it is the stage at which several approved neural devices have failed commercially despite working clinically.

What this means for announcements

A first-in-human implant means an early feasibility study has begun. It does not mean a product is near.

A demonstration video from a participant is genuinely encouraging and is not evidence of effectiveness in the regulatory sense.

The meaningful milestones are: a pivotal trial commencing, its endpoints published, and peer-reviewed multi-year outcome data.

Those are the events worth waiting for, and they arrive quietly.