Security

Connecticut adds neural data to its privacy law

Connecticut's updated data privacy law came into effect on 1 July 2026, and it adds neural data to the state's list of sensitive personal information categories, alongside things like health records and biometric data. It is one of the first explicit moves by a US state to regulate data collected directly from the brain and nervous system.

Why now

Consumer neurotechnology, from meditation headbands to focus-tracking wearables, has moved from niche gadget to mainstream product category faster than most privacy law could keep up with. Connecticut's update is a sign that regulators are starting to treat neural signals with the same seriousness as other deeply personal biometric data, rather than as a novelty data type.

The immediate trigger was evidence that the sector was not self-regulating. A review by the Neurorights Foundation of consumer neurotechnology companies found that 29 of the 30 it examined, around 97%, placed no meaningful limits on what they could do with users' neural data, with most reserving broad rights to share it with third parties. That is the kind of finding that produces legislation.

Connecticut is not alone, and not first

JurisdictionApproachStatus
ColoradoAdded neural data to its privacy act's sensitive data categoryIn force since 2024
CaliforniaAdded neural data to CCPA sensitive personal informationIn force since 2025
ConnecticutNeural data added as sensitive personal informationIn force 1 July 2026
ChileConstitutional protection for neurorightsIn force since 2021
EULikely captured as biometric or health data under GDPRNo neural-specific rule

Colorado and California moved first, so Connecticut joins a small group rather than leading it. The pattern is familiar from earlier privacy categories: a handful of states establish a definition, the definition gets copied, and companies end up complying nationally because building a separate data pipeline per state is more expensive than applying the strictest rule everywhere.

What "sensitive" actually changes

Reclassification is not a cosmetic change. Under laws of this type, sensitive data generally requires opt-in consent before collection rather than an opportunity to opt out afterwards, triggers a data protection assessment before processing, and narrows the purposes for which it can be used or sold.

For a company selling a focus-tracking headband, that is the difference between burying neural data use in a privacy policy and having to ask, specifically and in advance, with a genuine option to decline.

What neural data can and cannot reveal

Precision matters here, because the topic attracts overstatement. Consumer devices read electrical activity at the scalp, and that signal is noisy and spatially crude. These products cannot read thoughts, reconstruct sentences or extract memories, and claims in that direction should be treated sceptically.

What they can infer is still genuinely personal: attention and distraction, cognitive load, fatigue, arousal, emotional valence, sleep architecture, and responses to specific stimuli. Research has also indicated that neural signals can carry markers associated with certain neurological and psychiatric conditions, which is the real privacy concern. A dataset that reliably indicates early signs of a condition, held by a company with no obligation to keep it, is a problem regardless of whether anyone can read your inner monologue.

The commercial incentive is the other half of it. Measuring exactly what captured someone's attention, and how strongly they responded, is valuable to advertisers in a way few other signals are.

Consumer headbands, not brain implants

A distinction worth drawing, because the two get conflated and the law treats them very differently in practice.

Implanted brain-computer interfaces, the kind used in clinical trials to restore movement or communication, are regulated as medical devices. They go through trial protocols, ethical review and device approval, and the data they generate sits inside a medical framework with established rules.

Consumer neurotechnology has none of that. A headband sold as a wellness product is a consumer electronic device, and until laws like this one it fell under general privacy rules written with names and email addresses in mind. The irony is that the unregulated category is the one with millions of users, while the heavily regulated one has a few dozen.

What companies now have to do

For anyone building in this space, the practical compliance burden has three parts. Collection requires affirmative opt-in consent rather than a notice, which means a consent flow that a user can genuinely refuse without losing the product. Processing requires a documented assessment weighing benefits against privacy risk before the processing begins, not retrospectively. And the permitted-use boundary tightens, so training a general model on neural data, or sharing it with advertising partners, needs its own legal basis rather than riding on a broad policy clause.

Retention deserves particular thought. Neural data is not resettable in the way a password is, so holding it indefinitely accumulates a liability that grows with the dataset rather than a benefit.

What it means for readers

If you use any consumer device that reads brainwave or nervous system activity, Connecticut residents now have stronger legal grounds to demand disclosure of how that data is collected, used, and shared. For everyone else, it is an early signal of where privacy regulation is heading next: expect more states to follow with similar categories over the next year or two.

Enforcement is the open question. Connecticut's law is enforced by the state attorney general under its unfair trade practices framework, with civil penalties reaching $5,000 per wilful violation, while California provides for penalties of $2,500 per violation and $7,500 for intentional ones. Those figures bite at scale but require an attorney general to act, and state privacy enforcement has so far been thin. Connecticut's law also offers no private right of action, so an individual who believes their neural data was misused cannot sue over it directly and must rely on the state taking an interest.

By comparison, GDPR allows fines up to 4% of global annual turnover, which is why European enforcement produces the headline numbers, as in Google's €403 million location data fine and France fining IQVIA over health data. If you want to exercise rights over your own data in the meantime, our guide to making data requests to companies covers the mechanics.

Key takeaways

  • From 1 July 2026, Connecticut treats neural data as sensitive personal information, generally requiring opt-in consent, a data protection assessment, and narrower permitted uses.
  • Colorado and California got there first in 2024, so this is a consolidating trend rather than a lone move, and companies will likely comply nationally.
  • The evidence base was poor industry practice: the Neurorights Foundation found 29 of 30 consumer neurotech firms, roughly 97%, imposed no real limits on sharing neural data.
  • Be precise about capability. Consumer EEG cannot read thoughts. It can infer attention, cognitive load, fatigue, emotional response and sleep patterns, and may carry markers of neurological conditions.
  • The advertising value of knowing exactly what held someone's attention is the commercial driver behind collecting it.
  • Enforcement is the weak point. Connecticut allows up to $5,000 per wilful violation and California $2,500 to $7,500, against GDPR's ceiling of 4% of global turnover.

More from Future Technology HQ