Risk decides everything
Medical-device regulation is risk-based. The US FDA sorts devices into three classes by the harm a failure could cause. Class I (e.g. a tongue depressor) needs only general controls. Class II (e.g. an EEG amplifier) adds special controls and usually a 510(k) clearance. Class III — devices that support or sustain life, or present a potential unreasonable risk — needs the most rigorous review. An implanted, restorative BCI is Class III. It is also, during a study, a significant-risk device, which triggers the full investigational machinery below. The whole regulatory pathway flows from that single classification.
The IDE: permission to touch a human
You may not implant an investigational significant-risk device in a person without an approved Investigational Device Exemption (IDE). The IDE is the FDA's formal permission to run a clinical study; the exemption is from the laws that would otherwise forbid distributing an unapproved device. Two bodies must both approve before anyone is enrolled: the FDA (via the IDE) and a local Institutional Review Board (IRB).
- Device description and manufacturing — what exactly is implanted, made how, sterilized how.
- Preclinical evidence — biocompatibility (ISO 10993), electrical safety, bench durability, and animal studies establishing that the risk is acceptable to begin in humans.
- Clinical protocol — the study design, endpoints, sample size and stopping rules.
- Informed-consent documents and a risk analysis, plus a plan for monitoring and adverse-event reporting.
Routes to market
An IDE lets you study the device; a separate authorization lets you sell it. For a Class III BCI the default route is Premarket Approval (PMA) — a full demonstration of safety and effectiveness, usually resting on a pivotal trial. Two other doors exist. De Novo classification is for genuinely novel devices of low-to-moderate risk with no existing predicate; it can create a new Class II category. 510(k) clearance rests on substantial equivalence to an already-marketed predicate — powerful, but rarely available to a first-of-its-kind implant, because there is nothing equivalent to point to.
To compress the timeline, the FDA's Breakthrough Devices Program grants priority interaction and faster review to devices addressing serious conditions with no adequate alternative — a designation many BCI programs have received. It does not lower the evidence bar; it speeds the conversation. Outside the US, the EU's Medical Device Regulation (MDR) routes the highest-risk devices through Notified Bodies rather than a single agency, so a global product faces several parallel, non-identical dossiers.
Consent, capacity and the ethics scaffold
Every study runs under Good Clinical Practice and IRB oversight, and consent is its beating heart. BCI makes consent unusually hard because the very populations who benefit most — people with advanced ALS, brainstem stroke, or locked-in syndrome — may have impaired ability to communicate consent, and sometimes impaired capacity. Informed consent under impaired capacity leans on surrogate decision-makers, assent, and staged re-consent as a condition changes. A recurring hazard is therapeutic misconception: a participant hearing 'trial' as 'treatment' and expecting a cure the study was never designed to deliver.