EPISODE 07 · ON NEURAL INTERFACES

The wrist, not the brain

Closing question

If the interface that reached us reads the command leaving the muscle rather than the thought behind it — the body's outbox, not the mind — then when it acts a half-second before we're aware of deciding, whose intention are we actually watching: ours, or the machine's guess about ours?

Transcript

The Eye

This is Soft Ratios Radio.

The Hand

The voices are synthetic — no act, just the thinking.

The Eye

Hoyd Breton, a designer at Soft Ratios Studio, teaching himself in the open.

The Hand

One subject at a time. Today, neural interfaces. Say the phrase "brain-computer interface" and everyone pictures the same thing. A skull. A chip. Someone in a chair moving a cursor with a thought. That's the picture that sells.

The Eye

And the picture isn't wrong, exactly. It's just aimed at the wrong shelf. The thing that actually shipped to ordinary people — the neural interface you can buy — went on sale September 30th, 2025. It's a wristband. It bundles with a pair of Ray-Ban glasses for seven hundred and ninety-nine dollars. It's called the Neural Band. Meta sells it. Meta did not invent it. The invention belongs to a startup called CTRL-labs. Thomas Reardon — the engineer who built Internet Explorer, then walked away from software for a doctorate in neuroscience — founded it with Patrick Kaifosh on one wager: don't read the brain at all. Read the motor neurons at the wrist, the last relay before the hand moves. Meta bought the company in 2019 — the reported price ran somewhere between five hundred million and a billion dollars — and folded that work into its Reality Labs. So the band on the shelf is CTRL-labs' machine, wearing Meta's name.

The Hand

A wristband called neural. That's a strange word for a thing on your arm.

The Eye

It's the whole trick, and it pays to get precise here. The band doesn't touch your brain. It reads muscle. Surface electromyography — sEMG. Sixteen little sensors around the wrist, listening to the electrical crackle your forearm muscles make when they're about to move your fingers.

The Hand

So before I've even moved a finger, it hears the order going out.

The Eye

That's the biology of it, yes. When a motor neuron fires, it drives the muscle fibers it owns to depolarize. That little electrical event travels out through fat and skin, and a sensor on the surface catches the smeared-together sum of it — a motor unit action potential. The command's already on the wire before the knuckle bends.

The Hand

Here's what nags at me. If the near-term future of "neural" input is a bracelet reading your arm muscles — why did everyone spend two decades and billions of dollars drilling into skulls? That's the gap I want to walk into. Not "is the brain implant real." It's real. The question is why the thing that reached us first went around the brain entirely.

The Eye

Then let's start where the field started, because the detour tells you something. The whole idea has a birthday. 1973 — a computer scientist at UCLA, Jacques Vidal, writes a paper called "Toward Direct Brain-Computer Communication." He coins the term. And he actually does something with it: he reads visual evoked potentials off the visual cortex and works out where a person's trying to move a cursor. From electrodes on the scalp.

The Hand

1973. So the dream isn't new tech chasing hype. The dream's fifty years old.

The Eye

Older in spirit. Richard Caton found electrical currents in animal brains in 1875. Hans Berger recorded the first human EEG in 1924. But Vidal's the one who says: this could be an input. [pause] And then in 1988 you get the first genuinely practical version — Farwell and Donchin's P300 speller.

The Hand

Which works how, in a sentence I'd actually say out loud?

The Eye

A grid of letters flashes at you, row by row, column by column, randomly. You stare at the letter you want and silently count each time it lights up. When your letter flashes, your brain throws a little surprised spike about three hundred milliseconds later — the P300. It only fires for the rare thing you're waiting for. The machine watches for that spike, sees which row and which column triggered it, and where they cross is your letter.

The Hand

So you don't spell. You wait to be surprised, and the surprise gives you away.

The Eye

It's the oddball paradigm — the target's the low-probability event, one in six on a six-by-six grid, and the brain can't help flinching at it. Clever. And slow. That's the ceiling that haunts the whole non-invasive side: you can read the brain from outside, but you're reading through the skull, and the signal comes back faint and smeared.

The Hand

So the logic writes itself. Faint signal from outside — go inside. Get the electrode next to the neuron.

The Eye

And they did. The enabling piece of hardware was the Utah array — Richard Normann, University of Utah, around 1992. A little bed of stiff silicon needles you press into the cortex. It became the standard for twenty years. It's what let tetraplegic patients in the BrainGate program move cursors and robotic arms. Single-neuron resolution. The real thing.

The Hand

And this is where I want to slow you down, because "the real thing" is exactly the phrase that hides the problem. You put a stiff bed of needles into living tissue. The tissue is not a circuit board. It's wet, it moves, it defends itself. What happens after you close the skull?

The Eye

It fights back. That's not a metaphor — it's a documented cellular sequence, and it has a name: gliosis. Within minutes of insertion, you've torn cells and broken the blood-brain barrier. Immune cells called microglia reach toward the wound. They release signals that wake up the astrocytes —

The Hand

The brain's repair crew.

The Eye

More like the brain's scar-builders. The astrocytes swell, multiply, and weave a dense sheath around the electrode. In the dish you can watch it form in layers — microglia coating the shaft, astrocytes wrapping that in a mesh. And that scar does two things at once. It's electrically resistant, so it muffles the signal. And it physically shoves the living neurons back, away from the tip that's trying to hear them.

The Hand

So the body's response to the listening device is to build a wall around it and evict the neighbors.

The Eye

And now the numbers get sobering, which is where the myth of the permanent implant starts to crack. The largest longevity study looked at over six thousand recording sessions, across almost nine years, fifty-five Utah arrays. Average lifespan of usable recordings: six hundred and twenty-two days. Some lasted past a thousand. One went nine years.

The Hand

That doesn't sound like decay. That sounds durable.

The Eye

The average flatters it. Look at individual neurons and it's harsher. An earlier study tracked single units: sixty-seven percent stable through fifteen days. Thirty-one percent through forty-seven days. Twenty-one percent through a hundred and six. Eight percent stable past nine months. In rat motor cortex the yield fell from about fifty-three percent in week one to thirteen percent by week twenty-four.

The Hand

So the array survives. The specific voices it was hearing don't. It's like keeping the microphone but losing the people who were talking into it, one by one.

The Eye

And that reframes the engineering problem entirely. You can't build a consumer product on a signal that quietly walks away from you over months. For medicine — for someone with no other way to speak — you accept it, you re-tune, you fight the drift. For a mass-market gadget, that fight is the whole reason it can't be the brain. Not yet. Maybe not this decade.

The Hand

Which sends us back to the wrist. But I don't want to let Meta off easy just because their thing's non-invasive and clever. Let me push on the word again. They call it a neural interface. It is not reading your nervous system's thoughts. It's reading the exhaust from your muscles. Isn't "neural" here basically marketing?

The Eye

Partly. And partly it's technically defensible, which is what makes it slippery. The signal does originate in motor neurons — the alpha motor neurons firing is what starts the whole chain. So the electrical event has a neural cause. But you're catching it after it's driven a muscle, out at the periphery. Their own paper's honest about this in a way the branding isn't: it says brain interfaces have shown high-bandwidth communication only with invasive electrodes and custom decoders built for one person at a time. The wrist, by contrast —

The Hand

Generalizes.

The Eye

Out of the box. That's the July 2025 Nature paper — Kaifosh, Reardon, and colleagues. They trained on a hundred-plus hours of data from a hundred people per task, and the decoder works on a new person without being retrained for their body. That's the thing the brain implant can't do. Every implant so far is a bespoke suit tailored to one cortex.

The Hand

And that's the real reveal, isn't it. Not invasive versus non-invasive. It's bespoke versus generic. The brain gives you the richest signal and demands you refit it to every single skull. The wrist gives you a coarser signal that fits everyone off the rack. And "fits everyone" is what a product is.

The Eye

Let me put a hard edge on "coarser," though, so we don't romanticize the band either. The CTRL-labs team's own paper — published from inside Meta — concedes the boring old inputs still win. A trackpad beats it. A Joy-Con beats it. Physical handwriting beats it. Their sEMG handwriting clocks about twenty-one words a minute; typing on your phone is around thirty-six.

The Hand

So it's slower than the thing already in my pocket.

The Eye

Today, yes. Its case isn't speed. It's that it works with your hand at your side, in the dark, with no camera watching you, all day, on a wristband's battery. It's not trying to beat the phone at the desk. It's trying to be the input for glasses — for the moment when there's no screen to touch and no keyboard to find.

The Hand

Let me sit inside the hardware for a second, because the ordinariness of it is the whole point and it's easy to skate past. Sixteen sensing pairs around the wrist. Little gold-plated discs, six and a half millimeters across, twenty millimeters apart. Sampling two thousand times a second. And here's the humbling part — the physics says only the motor units within about ten to twelve millimeters of a sensor contribute any real signal. Everything deeper is lost in the tissue. So this thing that feels like magic on your arm is actually straining to hear a whisper through fat and skin, catching only the muscles closest to the surface. It's not omniscient. It's a very good listener pressed against a wall, and it only hears the room nearest to it. And that reframes the achievement. The magic isn't that it hears everything. It's that they built something useful out of hearing almost nothing cleanly — a faint, smeared, surface signal — and let a decoder guess the rest. Which, come to think of it, is the same trick the brain implant pulls. Both of them are reading a poor signal and betting on a smart guess.

The Eye

That's the thread that ties the whole field together, and it's the piece people miss when they argue about electrodes. The decoder is the breakthrough. Not the sensor. The last few years of genuine progress are really the years the AI got good enough to turn a sparse, noisy signal into fluent output.

The Hand

Give me the medical version of that, because that's where the stakes are real. Nobody's paralysis is cured by a wristband scrolling a menu.

The Eye

Speech restoration. This is where the science is genuinely staggering, and I want to be careful to keep the proven separate from the promised. Willett and colleagues, Nature 2023: a man with ALS, intracortical arrays. On a fifty-word vocabulary, a nine-percent word error rate. On a hundred-and-twenty-five-thousand-word vocabulary — essentially open English — twenty-three point eight percent error. Decoding attempted speech at sixty-two words a minute. Natural conversation is around a hundred and sixty.

The Hand

So a third of natural speed, and one word in four wrong on the big vocabulary. That's not a finished product.

The Eye

It wasn't. But hold the dates. The very next year — Card and colleagues, New England Journal of Medicine, the BrainGate2 study. Another man with ALS, severe dysarthria, four arrays, two hundred and fifty-six electrodes in the precentral gyrus. Day one, after thirty minutes of training: ninety-nine point six percent accuracy on fifty words. Day two, they open it to a hundred and twenty-five thousand words, an hour and a half more of training: ninety point two percent. And past eight months, it held ninety-seven and a half percent. Word error rate down around two and a half percent, self-paced, about thirty-one words a minute.

The Hand

Wait. From a quarter of the words wrong to two and a half percent, in roughly a year? That's not an incremental step. What changed — did the electrodes get that much better in twelve months?

The Eye

No. The electrodes are the same generation of idea. The language model got better. The decoder learned to lean on what a plausible English sentence looks like — strong priors filling the gaps in a weak signal. Same physical listening problem. Vastly smarter guess.

The Hand

So the honest headline isn't "we can read the brain now." It's "the machine got better at finishing our sentences." And that should unsettle us a little, shouldn't it? Because a system leaning that hard on priors — how much of the output is the person, and how much is the model's confident guess about what a person like them would probably say?

The Eye

That's the right worry, and it lives at the seam of every one of these results. When a UC Davis team pushed to real-time brain-to-voice in 2025 — synthesizing actual voice, not text, at about a fortieth of a second of latency, catching intonation, catching the rise of a question — the thing that made it feel like a phone call instead of a text message was the same thing that makes you nervous. The model's intervening, fast, between intention and utterance.

The Hand

And I'd add: these are single people. One participant. One cortex.

The Eye

Which is the caveat that has to ride along with every one of these numbers. Most of the speech results are a single participant. Ninety-seven percent accuracy in one man's motor cortex tells you the approach can work. It does not tell you it generalizes. That word — generalize — is exactly the one the wristband gets to claim and the implant does not.

The Hand

Let me re-ask the question we started with, now that we've been inside both. It isn't really "brain or wrist." It's: what kind of wrongness can you live with? Because none of these read a clean thought. They all read a smudge and guess. So the design question underneath everything is — when the machine is wrong, how wrong, and who pays for it?

The Eye

And that's a formalized field, not a vibe. The person who named it cleanly is Bliss Chapman, who leads the brain-interface software at Neuralink. His framing: there are different cost functions for different errors, and part of the design is building something that, when it's wrong, is still useful to the person using it. A wrong letter in a menu is a shrug. A wrong word in the sentence "do not" — that's a different kind of wrong.

The Hand

So you don't design for the machine being right. You design for the shape of its mistakes.

The Eye

And there's older theory underneath that, which is the part I find genuinely beautiful. Merel and Paninski, around 2013, formalized what everyone had been calling "the user gets used to it" into actual mathematics. They modeled the human and the decoder as two agents in one control problem, both learning at once. The decoder adapts to your signal; you adapt your signal to the decoder. Co-adaptation.

The Hand

Two learners circling each other. That's not a tool you pick up. That's a relationship.

The Eye

It's coordinate descent, in their language — you optimize, then it optimizes, back and forth toward a shared minimum. And they proved something almost deflating and completely honest: under their assumptions, all that mutual dancing can't beat what you'd get from one optimal fixed decoder plus the user simply learning well. The synergy people romanticize has a mathematical ceiling.

The Hand

I want to resist collapsing that into "so co-adaptation is pointless," because that's not what it says. It says the ceiling exists in an idealized world. In the messy one — where the signal drifts, where the neurons walk away over months — the adaptation isn't a bonus. It's how you keep standing on ground that's moving under you. The dance isn't for extra performance. It's for staying upright.

The Eye

And there's a concrete number that makes it feel real rather than theoretical. The CTRL-labs scientists found that even a little personalization — a small amount of tuning to the individual — improved their handwriting recognition by up to sixteen percent. That's co-adaptation earning its keep on a shipping product, not a whiteboard.

The Hand

Sixteen percent from the machine bending toward one person. Which loops us right back: the wristband's whole boast was that it works without that. Generic, off the rack.

The Eye

Both are true, and the tension is the honest state of it. It generalizes enough to work on a stranger's arm out of the box — and it still gets meaningfully better the moment it stops treating you like a stranger. Generic to get in the door. Bespoke to get good.

The Hand

There's a third thing we haven't touched, and it's the one that outlasts all the hardware arguments. What happens to the signal after it's read? Because whether it comes off your cortex or your wrist, it's data now. It's leaving your body and going somewhere.

The Eye

And the law's started to notice, unusually fast. Colorado went first — HB 24-1058, in force August 2024, the first US state to specifically protect neural data. California followed with SB 1223, operative January 2025. Both fold neural data into their strongest "sensitive" tier.

The Hand

Define neural data, though. Because if it means "brain," the wristband walks right through the gap.

The Eye

They saw the gap and closed it. Both statutes define neural data as information generated by measuring the activity of the central or peripheral nervous system. Peripheral. Written in. The wrist is explicitly covered. California even carves out the loophole — it says the data has to come from the nervous system itself, not be inferred from non-neural signals. So a heart-rate tracker guessing your stress isn't neural data. The band reading your motor neurons is.

The Hand

So the law drew the line exactly where Meta's marketing drew it — and used it against the marketing. "You called it neural. Fine. Then it's neural data, and here are the rules."

The Eye

And it matters because the industry's baseline is alarming. The Neurorights Foundation, with Perseus Strategies, went through thirty consumer neurotech companies in April 2024. Twenty-nine of the thirty — ninety-six point six seven percent — appear to have access to the user's neural data with no meaningful limit on it. Only one company offered any real restriction on selling it onward. Fewer than half encrypt and de-identify. And the medical shield you'd assume exists — HIPAA — doesn't apply to a gadget you bought online.

The Hand

So the same signal that would be locked behind medical consent in a hospital is, on a consumer wristband, just... a data stream with a privacy policy nobody reads.

The Eye

Which is the sting hidden in the whole detour to the periphery. The reason the wrist won on convenience — no surgery, all day, works on anyone — is the same reason the data problem gets bigger, not smaller. Millions of unregulated, always-on, generic devices generate far more nervous-system data than a handful of clinical implants ever could. The scale that makes it a product makes it a surveillance question.

The Hand

And it moves the human where the machine is going. Apple did this in May 2025 — a protocol they call BCI HID. It makes neural input a native, first-class category in the operating system. Not a hack bolted onto an accessibility feature. Alongside touch, alongside voice, alongside typing.

The Eye

With Synchron as the first partner — their Stentrode, the sixteen-electrode implant that goes in through the jugular vein, no craniotomy. And the interesting part is the closed loop: the iPad shares what's on screen back to the decoder, so the decoder knows the context it's decoding into. The device tells the brain-reader what the choices are, to make the reading better.

The Hand

So it's co-adaptation again, wearing an operating system. The screen and the signal, each shaping the other in real time.

The Eye

Once it's a native input type, the whole lineage clicks into place. The mouse waited for the graphical desktop. Multitouch waited for the phone. Each new way of pointing arrived when a new kind of computer needed it. Neural input — brain or wrist — is being positioned as the pointer for the thing with no surface to touch. Glasses. Ambient computing. The screen you don't hold.

The Hand

Then here's where I land, and I want to keep the split honest rather than tie a bow on it. We opened with the picture everyone has — the skull, the chip, the mind being read. And almost none of that picture is what reached us. What reached us reads muscle, not mind. It guesses more than it knows. It works because a language model finishes our sentences, and it's governed by laws that had to reach past the brain to catch it.

The Eye

And the piece that should stay uncomfortable is the honesty buried in the word itself. Every one of these — the wrist, the Stentrode, the arrays in a man's motor cortex — decodes attempted movement or attempted speech. The intention to act. Not the private thought behind it. The mind-reading fear is aimed at a capability nobody has demonstrated, while the real thing quietly redraws what counts as your data.

The Hand

So the thing we were afraid of and the thing that actually arrived aren't the same thing at all. We braced for a machine that reads the mind. What showed up is a machine that reads the body's outbox — the order already on its way out — and a very confident guesser standing behind it, and a data pipe running off your arm that the law only just learned to name.

The Eye

The brain stayed hard. It fights the needle, it walks away from the electrode, it insists on a bespoke fit for every skull. And the body, at the wrist, turned out to be the easy door — generic, all-day, already halfway to the machine. The future got in through the door nobody was watching.

The Hand

If the interface that reached us reads the command leaving the muscle rather than the thought behind it — the body's outbox, not the mind — then when it acts a half-second before we're aware of deciding, whose intention are we actually watching: ours, or the machine's guess about ours?

Sources

Verified secondary research (cited). The episode may draw on this alongside the core report; it invents nothing beyond the two.

First principles

How does surface electromyography (sEMG) physically detect motor-neuron activity at the wrist — what are the motor unit action potentials being measured, and what is the electrode/sampling configuration in Meta's Neural Band per the July 2025 Nature paper?

I'll research the biophysical mechanism of sEMG and the specifics of Meta's Neural Band from the July 2025 Nature paper.Let me get the specific electrode/channel/sampling configuration of Meta's sEMG-RD device from the Nature paper.Surface EMG detects motor-neuron activity indirectly: when alpha motor neurons fire, they drive their muscle fibers to depolarize, and the intracellular action potential is transformed into an extracellular action potential by volume conduction through tissue, with a MUAP obtained by combining the extracellular action potentials of all fibres belonging to one motor unit. The signal is inherently limited because surface EMG is fundamentally noisier than invasive EMG since motor unit action potentials must pass through body tissues such as fat and skin before they can be captured by a sensor on the surface, and modeling shows only those motor units with muscle fibers located within 10-12 mm of the electrodes contribute significant signal energy to the surface EMG. For Meta's device, the Nature paper describes it as a dry-electrode, multichannel recording platform with the ability to extract single putative MUAPs, chosen at the wrist because the wrist provides broad coverage of sEMG signals of hand, wrist and forearm muscles while affording social acceptability. On the electrode/sampling configuration, the sEMG-RD is a dry electrode, multichannel recording device with a high sample rate (2 kHz) and low-noise (2.46 μVrms), built with dry, gold-plated electrodes arranged in 16 discrete pods around the wrist, and technically specified as 48 channels configured as 16 bipolar channels arranged along the proximal and distal regions, with the remaining electrodes used for shielding or grounding; each electrode has a diameter of 6.5 mm (gold-plated brass), the center-to-center spacing between paired sensing electrodes is 20 mm, with input-referred RMS noise of 2.46 μVrms, a nominal gain of 190, an ADC resolution of 12 bits, and a full-scale range of 2.5 V providing approximately 65.5 dB of dynamic range.

Sources: Meta Neural Band - VR & AR Wiki - Virtual Reality & Augmented Reality Wiki · Meta Details EMG Wristband Gestures You'll Use To Control Its HUD & AR Glasses · On Optimizing Electrode Configuration for Wrist-Worn sEMG-Based Thumb Gesture Recognition · Meta researchers are developing a gesture-controlled wristband that can interact with a computer | TechCrunch · A generic non-invasive neuromotor interface for human-computer interaction | Nature · EMGBench: Benchmarking Out-of-Distribution Generalization and Adaptation for Electromyography · Simultaneous Control of Human Hand Joint Positions and Grip Force via HD-EMG and Deep Learning · A motor unit action potential-based method for surface electromyography decomposition | Journal of NeuroEngineering and Rehabilitation | Springer Nature Link · Behaviour of motor unit action potential rate, estimated from surface EMG, as a measure of muscle activation level - PMC · Detection of motor unit action potentials with surface electrodes: influence of electrode size and spacing - PubMed · Motor Unit Number Estimation Based on High-Density Surface Electromyography Decomposition - PMC · Detection of motor unit action potentials with surface electrodes: influence of electrode size and spacing | Biological Cybernetics | Springer Nature Link · Biosignal processing apparatus and method · Behaviour of motor unit action potential rate, estimated from surface EMG, as a measure of muscle activation level · Meta sEMG-RD: Advanced neural interface wristband - Worldkings - World Records Union · emg2qwerty: Touch Typing from Surface Electromyography · Meta’s Neural Interface Wristband Reads Your Muscle Signals to Control Computers - Yanko Design · Advances in electromyography armbands for gesture recognition and multimodal fusion - PMC · A Wireless, High-Quality, Soft and Portable Wrist-Worn System for sEMG Signal Detection - PMC · emg2qwerty: A Large Dataset with Baselines for Touch Typing using Surface Electromyography · EMG Dataset for Gesture Recognition with Arm Translation | Scientific Data · Advances in electromyography armbands for gesture recognition and multimodal fusion - ScienceDirect · Natural Typing Recognition via Surface Electromyography

What did Jacques Vidal's 1973 paper 'Toward Direct Brain-Computer Communication' actually propose and demonstrate, and how did the 1988 Farwell-Donchin P300 speller's oddball paradigm work?

I'll research both foundational BCI papers.Vidal's 1973 paper coined the term "brain-computer interface." In 1973, UCLA computer science professor Jacques Vidal published a landmark paper, "Toward direct brain-computer communication" that both coined the term "brain-computer interface" and set the foundation for an emerging field. His approach used visual evoked potentials: Vidal proved that a visual evoked potential (VEP) could be directly relayed from the brain to the computer; by recording the EEG signals emanating from the visual cortex of the brain, Vidal discovered that it could be determined where a user was trying to move a mouse cursor, marking the "first successful attempt to include brain signals into human-computer interaction."

For the 1988 Farwell-Donchin work (published as "Talking off the top of your head: toward a mental prosthesis utilizing event-related brain potentials," Electroencephalography and Clinical Neurophysiology 70(6): 510–523), the P300 Speller adapted the so-called oddball paradigm as the operating principle of the BCI and was the first P300-BCI; the system operated by briefly intensifying each row and column of a matrix, and the attended row and column elicited a P300 response. Mechanically, the oddball paradigm was used in the BCI system to establish a P300 speller in which users were presented with a 6 × 6 matrix containing 36 characters, and the rows and columns of the matrix were flashed randomly. The users attend to a cell of the matrix (that is, a target character) and count the number of times it is intensified; in this system, the probability of the intensified row/column containing the target character is 1/6, which is an oddball event that induces P300 potentials, and the system can then output a character by analyzing the P300 potentials. The P300 itself is elicited using an oddball paradigm, in which low-probability target items are mixed with high-probability standard (or non-target) items, and it manifests as a positive deflection in voltage with a latency of about 300 ms.

Sources: Jacques Vidal Coins the Term "Brain-Computer Interface" · Cerebral connections: UCLA engineers tap into rich legacy of brain-computer interface technology | UCLA Samueli School Of Engineering · Vidal, J.J. (1973) Toward Direct Brain-Computer Communication. Annual Review of Biophysics and Bioengineering, 2, 157-180. - References - Scientific Research Publishing · Toward Direct Brain-Computer Communication | Annual Reviews · Toward direct brain-computer communication - PubMed · Toward Direct Brain-Computer Communication | Scilit · An Ethical Evaluation of Brain-Computer Interfaces | by Oscar Petrov | Medium · Progress in Brain Computer Interfaces: Challenges and Trends · Toward direct brain-computer communication. | Semantic Scholar · Integrating EEG and MEG signals to improve motor imagery classification in brain-computer interfaces · Brain-computer interface research at the University of South Florida Cognitive Psychophysiology Laboratory: the P300 Speller - PubMed · A comparison study of two P300 speller paradigms for brain–computer interface - PMC · [Classical P300 spelling paradigm described by Farwell-Donchin 1988 [1]. | Download Scientific Diagram](https://www.researchgate.net/figure/Classical-P300-spelling-paradigm-described-by-Farwell-Donchin-1988-1_fig1_322874096) · Frontiers | Merging Brain-Computer Interface P300 speller datasets: Perspectives and pitfalls · Frontiers | The Self-Face Paradigm Improves the Performance of the P300-Speller System · (PDF) Perceptual errors in the Farwell and Donchin matrix speller · (PDF) P300-Based Chinese Speller: A Systematic Summary · Comment on the "P300" and Donchin Oddball Paradigm, Rosenfeld, et al., 1987 - PubMed · 510 Electroencephalography and cfinical Neurophysiology , 1988, 70:510-523 · BCI (Brain Computer Interface) Intervention in Autism

What is gliosis (glial scar encapsulation) at the cellular level, and what published longevity data quantify how quickly rigid intracortical electrodes like the Utah array lose stable single-unit recordings?

I'll research both parts of this question: the cellular biology of gliosis/glial scarring and published longevity data on Utah array single-unit recording stability.Gliosis (glial scar encapsulation) is a reactive cellular process triggered when a rigid probe is inserted into cortex. Insertion of a microelectrode array damages brain cells and ruptures the blood-brain barrier, initiating coordinated responses among glial cell types; the initial microglial response occurs within minutes, beginning with process extension toward the injury site. Microglia then release cytokines that activate nearby astrocytes, whose switch to a reactive state is hallmarked by proliferation, upregulation of glial fibrillary acidic protein (GFAP), process extension toward the injured site, and hypertrophy of the soma and processes, ultimately forming a glial scar. In vitro, the scar forms with microglia making an inner layer coating the electrode while astrocytes surround this microglial core with a network of GFAP-upregulated processes. The dominant hypothesis is that astrocyte hypertrophy drives recording failure: because of its thickness and high impedance, the glial scar physically and electrically isolates the electrode from healthy tissue and pushes active neurons away from the recording zone.

On longevity, the largest study to date is Sponheim et al. (2021, *J. Neural Eng.*). The authors analyzed over 6,000 recorded datasets spanning almost nine years, totaling 17 rhesus macaques, 2 human subjects, and 55 separate Utah arrays. Using implants in primary motor, premotor, prefrontal, and somatosensory cortices, they found the average lifespan of available recordings from Utah arrays was 622 days, with several examples lasting over 1,000 days and one up to 9 years, and human implants lasting longer than non-human primate implants. Electrode yield increased slightly and SNR steadily increased in the short term, saturating to a maximum after approximately four weeks, after which a slow decline in SNR was noted over time in both the human and NHP arrays. An earlier macaque study found faster attrition of individual units: 67% of units were stable through the first 15 days, 31% through 47 days, 21% through 106 days, and only 8% remained stable over 9 months. In rat motor cortex, decline was steeper still — single-unit active electrode yield decreased from 52.8 ± 10.0% at week 1 to 13.4 ± 5.1% at week 24.

Sources: Frontiers | Neural Interfaces for Intracortical Recording: Requirements, Fabrication Methods, and Characteristics · Frontiers | Biocompatibility of intracortical microelectrodes: current status and future prospects · Inhibition of Na+/H+ exchanger modulates microglial activation and scar formation following microelectrode implantation - PMC · Frontiers | In vivo monitoring of glial scar proliferation on chronically implanted neural electrodes by fiber optical coherence tomography · Glial scar - Wikipedia · In vivo spatiotemporal dynamics of astrocyte reactivity following neural electrode implantation · The Long-Term Stability of Intracortical Microstimulation and the Foreign Body Response Are Layer Dependent · In vitro model of glial scarring around neuroelectrodes chronically implanted in the CNS - ScienceDirect · Frontiers | The Long-Term Stability of Intracortical Microstimulation and the Foreign Body Response Are Layer Dependent · Glial cells, but not neurons, exhibit a controllable response to a localized inflammatory microenvironment in vitro - PMC · Longevity and reliability of chronic unit recordings using the Utah, intracortical multi-electrode arrays - IOPscience · (PDF) Longevity and reliability of chronic unit recordings using the Utah, intracortical multi-electrode arrays · (PDF) Chronic recording and electrochemical performance of Utah microelectrode arrays implanted in rat motor cortex · Longevity and Reliability of Chronic Unit Recordings using the Utah, Intracortical Multi-electrode Arrays - PMC · Longevity and reliability of chronic unit recordings using the Utah, intracortical multi-electrode arrays - PubMed · Ultra-Long Term Stability of Single Units Using Chronically Implanted Multielectrode Arrays - PMC

How does the co-adaptation framework formalized by Merel and Paninski (NeurIPS 2013 / PLOS Comp Biol) mathematically model the decoder-and-user as a two-agent learning problem?

I'll research the co-adaptation framework by Merel and Paninski from NeurIPS 2013.The Merel–Paninski line of work models co-adaptation as a formal joint-optimization problem. In the NeurIPS 2013 paper (Merel, Fox, Jebara, Paninski), the process of co-adaptation between the encoding model of the neural signal and the decoding algorithm is modeled as a multi-agent formulation of the linear quadratic Gaussian (LQG) control problem, in which adaptive decoders learn parameters to decode the user's neural response while feedback lets the user's neural tuning also adapt. They also propose a novel, modified decoder update rule which is aware of the fact that the encoder is also changing and show it can improve simulated co-adaptation dynamics. In the follow-on PLOS Computational Biology paper (Merel, Pianto, Cunningham, Paninski, 2015), they provide a mathematical framework for co-adaptation and relate co-adaptation to the joint optimization of the user's control scheme ("encoding model") and the decoding algorithm's parameters. Framing it as optimization rather than a vague synergy, this perspective allows co-adaptation to be investigated as a mathematical optimization procedure, rather than an ill-defined "synergistic" training process — co-adaptation then amounts to a coordinate-descent approach where the user and decoder alternately optimize. A notable theoretical result is that when the assumptions of that framework are respected, co-adaptation cannot yield better performance than that obtainable by an optimal initial choice of fixed decoder, coupled with optimal user learning.

Sources: Unsupervised Human-Machine Co-Adaptation via Mutual ... · A multi-agent control framework for co-adaptation in brain-computer interfaces · Encoder-decoder optimization for brain-computer interfaces - PubMed · Combining Decoder Design and Neural Adaptation in Brain-Machine Interfaces: Neuron00739-9) · Encoder-Decoder Optimization for Brain-Computer Interfaces · Bootstrapping Adaptive Human-Machine Interfaces with Offline Reinforcement Learning · (PDF) Encoder-Decoder Optimization for Brain-Computer Interfaces · Advances in Neural Information Processing Systems 26 · Encoder-Decoder Optimization for Brain-Computer Interfaces | PLOS Computational Biology · Encoder-Decoder Optimization for Brain-Computer Interfaces · Adaptive multi-degree of freedom Brain Computer Interface using online feedback: Towards novel methods and metrics of mutual adaptation between humans and machines for BCI | PLOS One · First Contact: Unsupervised Human-Machine Co-Adaptation via Mutual Information Maximization · A Co-Adaptive Brain-Computer Interface for End Users with Severe Motor Impairment | PLOS One · The Essential Guide to Brain-Computer Interface Technology Advancements | Naser | International Journal of Progressive Sciences and Technologies · A Game-Theoretic Model for Co-Adaptive Brain-Machine Interfaces · A Framework for Optimizing Co-adaptation in Body-Machine Interfaces · Large-Scale Assessment of a Fully Automatic Co-Adaptive Motor Imagery-Based Brain Computer Interface | PLOS One

What specific accuracy, word-error-rate, and words-per-minute figures were reported in the Willett et al. 2023 Nature paper and the Card et al. 2024 NEJM BrainGate2 speech neuroprosthesis study, and what vocabulary sizes were tested?

I'll research the specific figures from both studies.Both studies are well documented. Here are the verified figures:

**Willett et al. 2023 (*Nature*), Stanford intracortical BCI:** The ALS participant, decoded via intracortical microelectrode arrays, achieved a 9.1% word error rate on a 50-word vocabulary (2.7 times fewer errors than the previous state-of-the-art speech BCI) and a 23.8% word error rate on a 125,000-word vocabulary (the first successful demonstration, to their knowledge, of large-vocabulary decoding). Attempted speech was decoded at 62 words per minute, 3.4 times as fast as the previous record and beginning to approach the speed of natural conversation (160 words per minute).

**Card et al. 2024 (*NEJM*), BrainGate2 study:** The participant was a man in his 40s with ALS, tetraparesis and severe dysarthria (ALSFRS-R = 23), who underwent surgical implantation of four microelectrode arrays into his left precentral gyrus, recording neural activity from 256 intracortical electrodes. On the first day of system use, following 30 minutes of attempted speech training data, the neuroprosthesis achieved 99.6% accuracy with a 50-word vocabulary. On the second day, the output vocabulary increased to 125,000 words, and after 1.4 additional hours of training data, it achieved 90.2% accuracy. With further training data, the neuroprosthesis sustained 97.5% accuracy beyond eight months after surgical implantation. On error rates and speed, the neuroprosthesis achieved a word error rate of 2.5% (95% CI, 1.0% to 4.5%) by session 15, and this approximate accuracy was maintained through session 84, more than eight months after implant. Average Copy Task decoding in the final 5 evaluation sessions had a word error rate of 2.5% (95% CI, 2.0% to 3.1%) at the participant's self-paced speaking rate of 31.6 words per minute, and overall word error rates were consistently below 5%.

Second-order

What are the documented terms and required disclosures of California's SB 1223 and Colorado's HB 24-1058 neural-data laws, including their definitions of 'neural data' and whether they cover peripheral nervous system signals?

I'll research the specific legal terms of these two neural-data laws.Both statutes explicitly cover peripheral nervous system signals, which is the key point for sEMG.

California SB 1223 — Signed by Governor Newsom on September 28, 2024, and operative January 1, 2025, this bill amended the CCPA to add neural data to the definition of "sensitive personal information." It defines "neural data" to mean information that is generated by measuring the activity of a consumer's central or peripheral nervous system, and that is not inferred from nonneural information. By classifying it as sensitive personal information, the law grants consumers rights to access, delete, and limit the use and sharing of that data, and requires businesses that collect neural data to include proper disclosures in their privacy policies and respect a consumer's right to limit the use of that data. Notably, the California definition's exclusion for data "inferred from nonneural information" means that behavioral and physiological data that could be used to infer mental state would not be covered, so fitness trackers or wearable devices that capture data from systems other than the nervous system (e.g., the heart) that reveal stress level would not be covered, while electrical activity data from consumer neurotechnologies that directly capture data from the brain would be covered.

Colorado HB 24-1058 — Governor Jared Polis signed HB 24-1058 into law on April 17, 2024, and it entered into force on August 7, 2024, making Colorado the first state with targeted legislation protecting the neural data of neurotechnology consumers. The enacted statute defines the term as "neural data" means information that is generated by the measurement of the activity of an individual's central or peripheral nervous systems and that can be processed by or with the assistance of a device. Neural data is folded into the broader category of "biological data" within "sensitive data"; importantly, while "biological data" must be "used or intended to be used for identification purposes" to be in scope, the definition of "neural data" does not require that it be used or intended to be used for identification purposes. Under the CPA's sensitive-data rules, controllers that process biological or neural data must obtain consumers' affirmative opt-in consent, must not process such data without conducting and documenting a data protection assessment, and must update their privacy notice to reflect whether they process biological or neural data.

On peripheral signals (relevant to sEMG): Both laws' definitions of neural data expressly include the *peripheral* nervous system, not just the brain — California's covers "the activity of a consumer's central or peripheral nervous system that is not inferred from nonneural information," and Colorado's covers "the activity of an individual's central or peripheral nervous systems." One point of divergence worth flagging: the two states treat the definitions differently, and each of the U.S. states currently regulating "neural data" defines the term differently, varying around elements such as the treatment of central and peripheral nervous system data.

Sources: A MoFo Privacy Minute Q&A: California Revises CCPA to Cover Neural Data | Morrison Foerster · sensitive personal information: neural data, as amended June ... · Bill Text - SB-1223 Consumer privacy: sensitive personal information: neural data. · California Amends CCPA to Cover Neural Data and Clarify Scope of Personal Information · California Extends Data Privacy Protections to Brain Waves with SB 1223 · California Amends Key CCPA Definitions | McDermott · Neural Data as Sensitive Information: Unpacking SB 1223's New Protections · SB 1223 Page 1 Date of Hearing: July 2, 2024 · Neural Data and Consumer Privacy: California’s New Frontier in Data Protection and Neurorights | TechPolicy.Press · CA SB1223 | BillTrack50 · Colorado Adds Protection for Neural Data to its Privacy Act · HB24-1058 Protect Privacy of Biological Data | Colorado General Assembly · The “Neural Data” Goldilocks Problem: Defining “Neural Data” in U.S. State Privacy Laws · Colorado Amends Colorado Privacy Act to Address Biological and Neural Data | Practical Law · “Key Issues Raised by Colorado's Brain Data Privacy Bill,” Law360, April 30, 2024. | News & Insights | Alston & Bird · Colorado Amends Privacy Act with H.B. 1058, Adding New Protections for Biological and Neural Data · HOUSE BILL 24-1058 BY REPRESENTATIVE(S) Kipp and Soper, Amabile, Bacon, Bird, · Advocacy | United States | Neurorights · CO HB1058 | BillTrack50 · Second Regular Session Seventy-fourth General Assembly STATE OF COLORADO

What did the Neurorights Foundation/Perseus Strategies April 2024 report find about the 30 consumer neurotech companies' data practices, and which companies were assessed?

I'll research the Neurorights Foundation/Perseus Strategies April 2024 report on consumer neurotech data practices.Let me get more detail on the specific findings and the companies assessed.The report, titled *Safeguarding Brain Data: Assessing the Privacy Practices of Consumer Neurotechnology Companies*, was authored by Jared Genser, Stephen Damianos, and Rafael Yuste for the Neurorights Foundation (with Perseus Strategies as legal counsel). It is described as the first-ever, roughly 100-page report analyzing the data practices and user rights of consumer neurotechnology products, presenting an initial assessment of privacy practices by analyzing the privacy policies and user agreements of 30 neurotechnology companies with products that can be purchased online. Methodologically, it benchmarks data practices against half a dozen global data protection standards to compare the handling of neural data with established global standards for handling other types of sensitive and personally identifiable data.

On the core findings: Based on the review of policy documents and correspondence with companies, 29 of the 30 companies (96.67%) appear to have access to the consumer's neural data. The group analyzed the user agreements and privacy policies for 30 companies that sell commercially available products; only one company provided any meaningful restrictions on how users' data could be employed or whether the company could sell user data to third parties, and fewer than half of the companies surveyed encrypt their data and de-identify users. A related concern raised is regulatory: in the medical world, patients must consent if doctors want to share their private health information, but HIPAA doesn't apply to these companies.

Regarding which companies were assessed, the search results confirm only a subset by name rather than the full list of 30. Named companies discussed in connection with the report and related analyses include Emotiv, Muse, Meta, Microsoft, and WHOOP, with additional consumer neurotech firms in this space such as Neuralink, OpenBCI, Emotiv, and Kernel and Neurosity also appearing in the broader coverage. The search results do not provide the complete roster of all 30 assessed companies.

Sources: Neurorights Foundation | Human Rights Advocacy | Perseus Strategies · Safeguarding Brain Data: Assessing the Privacy Practices ... · Reports | Neurorights · The Legal Vacuum in Cognitive Privacy and Brain Data - Cardozo AELJ · Data privacy needed for your brain, Neurorights Foundation says · Safeguarding Brain Data: Assessing the Privacy Practices ... · Staff | NRF - Neurorights Foundation · Tech oligarchs on the brain | Tidsskrift for Den norske legeforening · Beyond neural data: Cognitive biometrics and mental privacy - ScienceDirect · Your brain for sale? The new frontier of neural data · Safeguarding Brain Data: Assessing the Privacy Practices ... · Connecticut Classifies Brain Data as Sensitive Starting July 1: New Rights for EEG Wearable Users · Your brain for sale? The new frontier of neural data · Your brain for sale? The new frontier of neural data - World leading higher education information and services · How tech giants are harvesting your brain waves | Dailyhunt

What are the technical specifics and first partners of Apple's BCI HID protocol announced May 13, 2025, and how does its context-sharing closed loop function alongside touch/voice?

I'll research Apple's BCI HID protocol announced around May 13, 2025.On May 13, 2025, Apple announced its BCI Human Interface Device (BCI HID) protocol, and Synchron confirmed the partnership, becoming the first company to achieve native integration with the just-announced BCI HID profile; Apple has partnered with the neurotechnology startup Synchron on this feature, with Team Gleason (an ALS nonprofit) also cited as a collaborator. Technically, the startup makes the Stentrode, a stent-like brain implant that turns thoughts into real-time digital commands, implanted through the jugular vein atop the brain's motor cortex, without requiring open-brain surgery, and the device uses sixteen electrodes to read and translate brain signals into selecting icons; the protocol adds neural support to Apple's existing Switch Control accessibility feature across iOS, iPadOS, and visionOS. The closed loop works because, unlike traditional assistive devices that merely replicate hardware inputs, signals are wirelessly transmitted to an external decoder, which interfaces directly with iPadOS through the new HID protocol; the system allows for closed-loop communication, where an iPad, iPhone or Apple Vision Pro shares contextual screen data with the BCI decoder to optimize real-time performance, enabling precise, intuitive control using just neural signals. This makes brain signals a first-class modality: Apple is formally recognizing brain signals as a native input category—joining the ranks of touch, keyboard, and voice. The first public demonstration came on August 4, 2025, featuring Mark, a participant in Synchron's COMMAND clinical study and a person living with ALS, who uses the company's implantable BCI to navigate the iPad home screen, open apps, and compose text all without using his hands, voice, or eyes, with controlled rollouts with patients in clinical trials using BCI HID-compatible features beginning later in 2025.

Sources: Synchron announces BCI integration with Apple tech · Synchron Debuts First Thought-Controlled iPad Experience Using Apple’s New BCI Human Interface Device Protocol · Synchron Debuts First Thought-Controlled iPad Experience Using Apple’s New BCI Human Interface Device Protocol · Synchron To Achieve First Native Brain-Computer Interface Integration with iPhone, iPad and Apple Vision Pro · Synchron debuts BCI-enabled, thought-controlled iPad experience through Apple collab · Synchron first to integrate brain-computer interface with Apple devices | Medical Economics · Apple Jumps into the Brain-Computer Interface Market with Synchron Collaboration · Synchron To Achieve First Native Brain-Computer Interface Integration with iPhone, iPad and Apple Vision Pro · Synchron To Achieve First Native Brain-Computer Interface Integration with iPhone, iPad and Apple Vision Pro - BioSpace · Synchron Achieves Native BCI Integration with iPhone, iPad, and Apple Vision Pro · In a world first, a man has controlled an iPad using his mind – Apple Must · Synchron Unveils World’s First "Thought-Driven Control" of an Apple iPad in New Brain-Computer Interface Milestone - The Debrief · iPhone mind-control coming via Brain Computer Interfaces support · Brain-Controlled iPhone: How Apple's BCI HID Protocol is Revolutionizing Accessibility for All - TorontoStarts · Synchron on X: "Thoughts are now an input device. Today, @Apple announced its new BCI Human Interface Device (#HID) protocol—and Synchron is proud to be the first brain-computer interface company to achieve native integration with iPhone, iPad, and Apple Vision Pro. That means: 🧠 No touch. https://t.co/7prNC3uoau" / X

Third-order

How much venture capital has flowed into BCI companies from 2022 through 2025 by named round (e.g., Neuralink's $650M Series E, Synchron's $200M Series D, Merge Labs' ~$252M), and how do market-size forecasts from Precedence, BCC, and Towards Healthcare differ?

I'll research VC funding rounds for BCI companies (2022-2025) and the market-size forecasts from the three named firms.I have solid data on the funding rounds. I didn't retrieve the market-forecast figures from Precedence, BCC, and Towards Healthcare (my search calls were exhausted), so I'll report only what's verified.

Named funding rounds (2022–2025). Neuralink closed a Series E funding round of $650 million with participation from investors including ARK Invest, DFJ Growth, Founders Fund, G42, Human Capital, Lightspeed, QIA, Sequoia Capital, Thrive Capital, Valor Equity Partners, and Vy Capital, announced June 2, 2025, at a reported $9 billion pre-money valuation, following its $280 million Series D in 2023, with an additional $43 million tranche added months later. Synchron raised a $200 million Series D (announced Nov. 6, 2025), bringing its total funding to $345 million; its Stentrode BCI platform is the world's first endovascular brain-computer interface. That round was led by Double Point Ventures, alongside existing investors ARCH Ventures, Khosla Ventures, Bezos Expeditions, NTI and METIS, with new investors including the Australian National Reconstruction Fund, T.Rx Capital, Qatar Investment Authority, K5 Global, Protocol Labs, and IQT. Merge Labs—co-founded by Sam Altman and technically a January 2026 event, just outside your window—was backed as part of a $252 million seed funding round, instantly positioning it among the most heavily funded entrants in the BCI field.

Aggregate scale. Synchron's funding moves investment in BCI companies in 2024 and 2025 beyond $1 billion, with the round adding to financings at Blackrock Neurotech, Neuralink and Precision Neuroscience. One industry estimate cited around the Merge announcement put roughly $4.8 billion flowing into neurotech across 140 deals in 2025, with implanted systems accounting for roughly $3.2 billion and non-implanted approaches making up the remainder.

On the market-size forecasts: I was unable to retrieve verified figures from Precedence Research, BCC Research, or Towards Healthcare within this session, so I cannot report or reconcile their specific market-size projections here without risking fabrication. Those three forecasts would need a dedicated follow-up search to confirm the exact numbers, base years, and CAGRs each publisher uses.

Sources: Neuralink raises $650 million Series E | Updates | Neuralink · Neuralink raises $650M at reported $9B pre-money valuation - SiliconANGLE · Neuralink’s $650M Series E: Accelerating the Future of Brain-Computer Interfaces - Applying AI · Elon Musk's Neuralink closes a $650M Series E | TechCrunch · Neuralink Stock: $9B Valuation — Is It a Buy? | TSG Invest · Neuralink Raises $650M in Series E Funding Round · Neuralink secures $650M series E funding to expand patient access to brain chip technology · Elon Musk’s Neuralink raises $650 million in Series E funding · Neuralink Raises $650 Million in Late-Stage Funding Round · Synchron Raises $200 Million Series D to Advance Brain-Computer Interface Technology · Synchron Raises $200M in Series D Funding · Synchron raises $200M Series D to support BCI commercialization, next-gen tech · Synchron Secures $200M Series D, Plans New Brain Interface With ‘Potentially More Applications’ · Synchron raises $200M to advance its brain-computer interface for paralysis · Synchron raises $200M to prepare for brain computer interface launch | MedTech Dive · Full Circle: Synchron’s $200 Million Signals Revolution By Evolution · Brain-computer interface company Synchron raises $200M | MobiHealthNews · Synchron Announces $200M Series D Round · Synchron: $200 Million Series D Closed To Advance Commercial Brain-Computer Interface Platform · Merge Labs Raises $252 Million for Brain-Computer Interface Devices - Bloomberg · Merge Labs Funding: The $252M Bet on Linking Brains to AI | StartupHub.ai · Why OpenAI’s $252M bet on Merge Labs redefines brain–computer interfaces · OpenAI Backs Ultrasound-Based Brain-Computer Interface Startup With $252M Seed Round · OpenAI Invests $252M in Brain-Computer Interface Market Projected to Hit $6B by 2032 · OpenAI backs brain-computer interface startup in $252M investment - Becker's Hospital Review | Healthcare News & Analysis · Neuralink rival Merge Labs bags $252M, led by OpenAI, to link brains and computers — TFN · OpenAI CEO Sam Altman raises $252 million for brain computer interface venture — but Merge Labs is still in an early research phase | Tom's Hardware · OpenAI Backs Merge Labs: Brain-Computer Interface Leap with Ultrasound - Mischa Dohler

What is the current state of Chinese BCI efforts (Beijing Xinzhida, NeuroXess) and national programs, including any human implants, funding, or government roadmaps announced through 2025-2026?

I'll research the current state of Chinese BCI efforts including Beijing Xinzhida, NeuroXess, and national programs.China's BCI efforts have accelerated sharply through 2025–2026 across both startups and coordinated state programs. NeuroXess (脑虎科技), a Shanghai company founded in 2021, achieved China's first human implantation of a fully implanted brain-computer interface product with an integrated battery, with the surgery performed by a medical team from Huashan Hospital affiliated with Fudan University in December 2025; the patient, a high-level paraplegic, achieved mind control five days post-surgery, and the product eliminates the need for external equipment, reducing infection risks. Earlier, in an August 2024 procedure disclosed in January 2025, NeuroXess implanted a 256-channel flexible cortical device in an epilepsy patient who achieved 71 percent speech decoding accuracy using 142 common Chinese syllables within five days, with single-character decoding latency under 100 milliseconds — the highest-level real-time decoding of Chinese speech achieved in China, and had completed more than 50 flexible implantable BCI clinical procedures by mid-2025.

Beijing Xinzhida (Neucyber Neurotech), a state-backed firm whose controlling shareholder is the Beijing municipal government's state asset regulator, first unveiled its "Neucyber" invasive implant in April 2024 tested only in monkeys. By 2026 its semi-invasive "Beinao-1" system had progressed to humans: since February 2025, CIBR and Xinzhida, with Peking University First Hospital, Xuanwu Hospital, and Beijing Tiantan Hospital, carried out five cases of human implantation of the "Beinao-1" system, with a sixth planned and registered clinical trials planned for mid-2026, while the higher-performance invasive "Beinao-2" is expected to enter clinical verification in 2026. By early 2026, this world's first high-throughput, semi-invasive epidural BCI had been implanted in 10 humans, including post-stroke hemiplegia and ALS patients, with stable signals and no severe complications, in trials at Beijing Tiantan Hospital.

On the national/government dimension, China's push follows a formal industrial strategy: an August 2025 Implementation Plan for Promoting Innovation and Development of the Brain-Computer Interface Industry named BCI development as strategic industrial infrastructure, and BCI has been designated a strategic emerging industry in China's policy framework, with support dating back to the 13th Five-Year Plan. Regulatory and infrastructure milestones followed: on March 13, 2026, the NMPA granted marketing approval to Neuracle Medical Technology for its invasive BCI system, the first device of its kind to receive commercial clearance anywhere in the world; China's first BCI clinical ward opened at Beijing Tiantan Hospital in May 2025, and by April 2026, more than 30 hospitals across over 12 provinces and municipalities had established BCI wards

Sources: NeuroXess Achieves Milestone With China’s First Battery-Integrated Brain Chip Implant · China approves first commercial brain-computer interface as NeuroXess trials begin · Brain-computer interface makes breakthrough by deciphering Chinese speech in brain - Chinadaily.com.cn · USA vs Chinese BCI Advancements in H1 2026 · NeuroXess Achieves Milestone With China’s First Battery-Integrated BCI Implant · China Focus: Brain-computer interface makes breakthrough by deciphering Chinese speech in brain-Xinhua · Design and Quantitative Evaluation of an Embedded EEG Instrumentation Platform for Real-Time SSVEP Decoding · Implantable Brain-Computer Interface for Upper-Limb Recovery After Stroke · Editorial: New approaches in Brain-Machine Interfaces with implants · China brain computer interface outfit accelerates to human trials in quest to outpace Neuralink — mix of government backing and investor enthusiasm speeds time to market for NeuroXess | Tom's Hardware · The Sixth Case! "Beinao-1" Brain-computer Interface to be Clinically Implanted · China develops Neuralink rival for brain computer interface tech · China developed its very own Neuralink — Neucyber brain interface comes from a neurotechnology firm | Tom's Hardware · China Aims to Set New World Standards for Brain-Machine Interfaces, Echoing Elon Musk’s Neuralink - The Debrief · China Unveils Brain Chip Similar to Elon Musk’s Neuralink · Chinese Hospitals Rush to Launch Brain-Computer Interface Wards - Caixin Global · China Makes A BCI Breakthrough With A Monkey and A Brain Chip · China Moves Ahead with Research into Brain-computer Interfaces----Chinese Academy of Sciences · Privacy-Preserving Brain-Computer Interfaces: A Systematic Review · China to accelerate its brain chip program after experimenting on a monkey