Brain-Computer Interfaces: From Medical Miracles to Consumer Gadgets

[NightCrawler81]

In January 2024, a paralysed man named Noland Arbaugh became the first human recipient of a Neuralink N1 implant. Within weeks of his recovery he was playing online chess using only his thoughts, and within months he was streaming on social media about daily life with a chip in his motor cortex. That moment did for brain-computer interfaces roughly what AlphaGo's 2016 victory did for artificial intelligence, it converted a niche research field most people had only encountered in science fiction into something the broader public suddenly understood was actually happening. Two years on, the field looks nothing like a one company story, several rival approaches are producing genuinely comparable results in human patients, non-invasive versions of the same underlying technology are already sitting on consumer store shelves, and state legislatures across the US have begun passing the first laws in history specifically written to protect the privacy of a person's own brain activity. This piece looks at where the medical technology actually stands beyond Neuralink, what the non-invasive consumer version can and cannot currently do, and the genuinely hard ethical questions the whole field is now forced to confront in real time rather than in the abstract.

The medical frontier: it's not just Neuralink anymore

Neuralink's approach uses a robotic surgical system to insert 1,024 ultra thin electrode threads across 64 separate bundles directly into the brain's motor cortex, producing extraordinarily high resolution readings of individual neuron activity. By early 2026 the company's PRIME feasibility study had enrolled roughly 21 participants across the US, UK, Canada and the UAE, with sister trials underway targeting automatic arm control and speech restoration specifically. It remains the most heavily funded and most publicly recognised name in the field, but it is no longer the most clinically advanced by every measure, and several credible competitors are pursuing meaningfully different tradeoffs between signal quality, surgical risk and how quickly a treatment can realistically reach large numbers of patients.

Synchron has taken arguably the most pragmatic route, its Stentrode device is inserted through blood vessels using a catheter procedure similar to ones already performed daily in cardiology labs at hospitals worldwide, meaning it requires no open brain surgery at all. The tradeoff is real, Stentrode uses just sixteen electrodes compared to Neuralink's thousand plus threads, producing lower resolution signals and generally slower, less precise cursor control for patients. What Synchron gains in exchange is a dramatically better risk and scalability profile, no craniotomy means no risk of brain tissue damage during implantation, a recovery period measured in days rather than weeks, and a procedure that can plausibly be performed at the volume modern healthcare systems can actually support rather than requiring a specialised neurosurgical team. Synchron has implanted its device in patients across multiple countries with long term data showing the Stentrode remaining functional and stable for over two years without migration or signal degradation, and through 2025 the company expanded a partnership with Apple specifically to let Stentrode signals directly control an iPhone or Vision Pro. As of 2026, Synchron's ongoing pivotal trial represents the industry's first serious run at full FDA premarket approval for a permanently implanted communication BCI, a milestone that, if achieved, would arrive before Neuralink's own equivalent approval, which analysts do not expect before 2027 or 2028.

A third genuinely distinct approach comes from Precision Neuroscience, founded by a former Neuralink co-founder, which places a thin, flexible electrode array directly on the surface of the brain rather than penetrating into it, a technique known as electrocorticography. This surface based method sacrifices some of the fine grained resolution of penetrating electrodes but meaningfully reduces the risk of long term tissue damage, and in April 2025 Precision became the first BCI company in the current generation to receive FDA 510(k) clearance, for its Layer 7 cortical array in implantations lasting up to 30 days, a genuinely significant regulatory first even though it covers a narrower use case than a fully chronic implant. Paradromics received its own FDA investigational device exemption in November 2025 for a high data rate implant specifically targeting speech restoration for patients with locked in syndrome, with patient recruitment beginning in early 2026. Taken together, the field now has multiple credible, independently funded technical paradigms, penetrating threads, endovascular deployment, and cortical surface arrays, each targeting a genuinely different point on the tradeoff curve between signal fidelity, surgical risk and how many patients the treatment could eventually reach, which is a meaningfully healthier scientific and commercial landscape than a single company racing alone toward one specific technical bet.

Beyond the implant: what non-invasive BCI can actually do today

Everything described so far requires a genuine surgical procedure, but a parallel and much less discussed branch of this field has been quietly reaching consumers without any surgery at all. Non-invasive systems read electrical activity through the scalp using electroencephalography, commonly called EEG, packaged into headbands, headphones and other wearable form factors already sold directly to ordinary consumers rather than hospital patients. Companies including Emotiv, OpenBCI and Kernel, the latter using a more advanced technique called magnetoencephalography offering roughly 50 channels of coverage at around 1 centimetre spatial resolution, sell devices marketed around meditation feedback, sleep tracking and focus training. It is genuinely important to be clear about what these consumer devices can and cannot do, they cannot read specific thoughts, words or detailed intentions the way implanted systems increasingly can, the skull and scalp scatter and attenuate the brain's electrical signals enormously before they ever reach an external sensor, meaning non-invasive EEG produces a much blurrier, lower resolution picture limited mostly to broad states like relaxation, focus or drowsiness rather than anything approaching mind reading in the way people often imagine. Even so, that more limited signal has proven commercially valuable enough to build a genuine consumer product category, and the gap between what implanted systems can demonstrate in a lab and what a non-invasive headband can offer at home, while still very large, is one several major technology companies are actively working to narrow.

The ethics: what happens when your thoughts become data

The genuinely hard questions this field now faces are not really about whether the underlying technology works, the clinical results described above make clear that it increasingly does, they are about who gets access to the data these devices generate and what can legally and ethically be done with it once collected. A pivotal moment for this specific debate arrived in April 2024, when the Neurorights Foundation, an advocacy group focused on protecting people from neurotechnology misuse, published a survey of 30 consumer neurotechnology companies with products already available for purchase online. The findings were genuinely stark, 29 of the 30 companies surveyed had direct access to users' raw brain data and provided no meaningful limitations on that access internally, and almost all of them retained the ability to share that data with third parties. The report noted plainly that neural data is uniquely sensitive precisely because it can potentially reveal a person's memories, biases, emotional states and health conditions involuntarily, information a person may not even consciously realise their own brainwaves are exposing, in a way that is meaningfully different from other categories of sensitive personal data like a credit card number or a browsing history, both of which reflect a deliberate action rather than an involuntary internal state.

That report directly triggered a genuinely fast legislative response. Colorado became the first US state in history to pass targeted neural data protection when Governor Jared Polis signed HB 24-1058 into law on April 17, 2024, amending the state's existing privacy act to classify neural data as a distinct category of sensitive personal information requiring heightened consent and handling obligations, the law passed the Colorado legislature by a combined vote of 110 to 1 and took effect that August. California followed just months later, Governor Gavin Newsom signed SB 1223 in September 2024, extending the same sensitive data protections found in the California Consumer Privacy Act specifically to neural data, with particular significance given how many neurotechnology companies are headquartered in or operate out of California specifically. Montana and Connecticut have since passed comparable legislation, and Democratic senators on the US Senate Commerce Committee wrote directly to the Federal Trade Commission in April 2025 urging a formal investigation into whether neurotechnology companies were actively exploiting consumers' brain data given the Neurorights Foundation's findings. Colorado state senator Cathy Kipp, who sponsored her state's original bill, captured the core anxiety driving this legislative wave in a genuinely simple question, if you collect someone's neural data today, what might you be able to read from it five years from now, given how quickly the underlying decoding technology keeps improving. The spirit of these laws, according to the Neurorights Foundation's own medical director, is deliberately narrow, they aim to protect the neural data itself as a category, rather than attempting to regulate whatever algorithm or AI system might eventually be used to interpret it, a distinction that matters because the actual decoding capability is advancing considerably faster than any single piece of legislation can realistically anticipate.

Where this leaves us

The honest picture in 2026 is one of two genuinely different technologies converging toward the same difficult conversation from opposite directions. Implanted medical BCIs are demonstrating real, clinically meaningful capability, restoring communication and computer control to people with severe paralysis, with multiple companies now pursuing distinct regulatory pathways toward broader approval within the next few years. Non-invasive consumer BCIs remain far more limited in what they can actually detect, but are already shipping in commercial products with remarkably little oversight over what happens to the data they collect, exactly the gap the Neurorights Foundation's research exposed and state legislatures have only just begun addressing. The genuinely open question sitting underneath both branches of this field is not whether brain-computer interfaces will keep improving, the last two years make clear that they will, it is whether the legal, ethical and commercial norms governing who can access a person's neural data, and what they are allowed to infer from it, can be built out fast enough to keep pace with a technology whose entire value proposition is reading information more directly out of the human brain than any previous consumer technology has ever attempted.

[Tia88]

Genuinely excellent overview of the competitive landscape, the Synchron versus Neuralink tradeoff explanation is the clearest I've read anywhere, catheter procedure versus open brain surgery is such a stark difference in accessibility that it's honestly surprising Synchron doesn't get more mainstream attention given how far ahead its regulatory pathway apparently is. Where I think the essay slightly undersells the risk is on the non-invasive consumer side, the piece is careful and accurate to say these headbands can't read specific thoughts today, but the Neurorights Foundation quote about what might be readable from the same stored data five years from now is exactly the point that should worry people more than the essay's relatively reassuring current capability framing suggests, companies are retaining raw data indefinitely in many cases, and decoding capability is improving faster than most people realize, meaning today's meditation headband recording could become tomorrow's much more revealing dataset even without the device itself ever changing.