Skip to main content

Lab Notebook · Consumer BCI

A brain–computer interface for under £600 — no surgery, off published science

In April 2026 we designed a non-invasive, bidirectional BCI — EMG cursor control, EEG reading, and PEMF stimulation — for around £500–£606, grounded in two decades of published neuroscience. Design-first: eight review reports and the IP filed before code. Neuralink-class ambition at a fraction of the cost. An open-hardware, open-build report by choice.

JR
Jon RossFounder, vocabotics — 15 years building safety-critical systemsLab report · dated 10 April 2026
Verified by a human. Drafted with AI, verified by a human. Jon Ross, 10 Apr 2026
Living document. Reviewed 10 Apr 2026
Entry date
10 April 2026
Category
Hardware
Lead over the world
~1–2 years ahead
Access
🔓 Public
ResearchPartly provenDownload PDFPDF · 746 KB

Neuralink-class ambition, one-tenth the cost, zero surgery — a bidirectional brain–computer interface you could build for the price of a phone.

£56
hardware for the EMG cursor path
3-in-1
EMG control · EEG read · PEMF stimulate
8
design reports written before code

Honest evaluation

Partly proven

The EMG cursor-control path is proven on the bench; the full three-in-one integrated device is designed, not yet built.

What would prove or disprove it further

What would move this further toward fully proven: EEG reading and PEMF stimulation each getting their own working bench prototype, the same way EMG already has one, and then the three integrated into one device rather than three separate paths. What would disprove the wider claim: the £500-606 cost estimate or the "no surgery" premise failing to hold once EEG and PEMF are actually built and integrated rather than just designed and priced.

The evidence — full reasoning behind the verdict

Verdict: partly-proven. The narrowest claim — that EMG-based cursor control, built on published neuroscience, works on cheap hardware (ESP32-S3

  • ADS1299) at roughly £56 — is proven: working finger-test code exists and runs on the bench today. The wider claim — a three-in-one, non-invasive, bidirectional device (EMG plus EEG plus PEMF) for £500-606 — is not yet built; it's a rigorous, reviewed design (eight council reports, IP filed before disclosure) sitting ahead of the one path that's actually running.

It landed here because the project deliberately sequenced design before code on the safety-serious parts. EMG cursor control is the lowest-risk, best-precedented piece — motor cortex as a peripheral bus is decades-old published science — so it's the piece that got built first and is the piece that's proven. EEG reading and PEMF stimulation, especially the latter (writing back to a brain, even gently), were treated as needing review and IP protection before any code, which is exactly why they remain design rather than bench results at this point.

Runnable proof — see it work

The public conversation about brain–computer interfaces is dominated by one word: surgery. We wanted to know how far you could get without it — non-invasively, cheaply, and honestly — using science that has been in the literature for twenty years. This is an open report by design: the build belongs in the open, as a credibility-and-goodwill contribution, not a moat.

What it was

A non-invasive, bidirectional BCI for roughly £500–£606, combining three capabilities in one platform:

  • EMG-based cursor control — reading muscle signals to drive a pointer (about £56 of hardware for this path alone).
  • EEG reading — sensing brain activity with consumer-grade hardware.
  • PEMF stimulation — writing back, gently, with pulsed electromagnetic fields.

The animating thesis, drawn straight from the published work: the motor system is a peripheral bus the brain will learn to drive, so a "virtual muscle" controlling a cursor becomes, to the brain, a muscle.

What we built

A prototype and a rigorous design, in that deliberate order:

  • Working EMG finger-test code on ESP32-S3 + ADS1299 hardware.
  • Eight council design reports written before code, with the IP filed before disclosure.
Research

The measured, honest facts are about cost and scope, not a benchmark: a three-in-one, non-invasive, bidirectional device specified for the price of a phone, with a working EMG path proven on the bench. The claim we make carefully is integration and price, not a new sensor.

What we learned — including the honest edges

  • The novelty is the integration, not the science. Every underlying capability is grounded in decades-old, cited neuroscience (Carmena, Schwartz, Hochberg). The contribution is combining them, non-invasively, at this price — and being explicit that we did not invent the physics.
  • A prototype is not a product. Working EMG code is real; the fully integrated three-in-one device is not built yet, and we do not imply otherwise.
  • Stimulation is safety-serious. Writing to a brain, even gently with PEMF, is not a toy — which is exactly why it was treated design-first, with reviews and IP before any disclosure.

Where it went / status

Held as an open-hardware research prototype. It sits at the sensing end of the lab's "waves, not pixels" thesis — capture the world as waves and feed it to a wave model — and it makes a wider point the archive keeps returning to: the hardware floor for owned, private intelligence is far lower than the surgical-implant headlines suggest.

What is still open — kept visible

The honest edges, next to the wins. This is what turns 🔬 into 🟢 — honestly.

  • It is a prototype: working EMG finger-test code exists; the full integrated device does not yet.
  • The novelty is the integration and the price, not the sensors — the underlying science is two decades old and cited as such.
  • Neuro-stimulation (PEMF) is a safety-serious surface treated design-first, not a toy.

Proofs & sparks

We demonstrate rather than assert. Each ✅ proof is a visible result with a hard figure.

  • Sub-£600 bidirectional BCI~£500–606 bench builda bidirectional EMG interface (ESP32-S3 + ADS1299) with working finger-test code on the bench, integrated for ~£500–606 — needs the hardware to run.

Where this connects

Sources

  1. vocabotics project audit — BCI (EMG/EEG/PEMF, ~£500–606), Apr 2026vocabotics internal project history · as of April 2026
  2. Carmena et al., learning to control a brain–machine interface (motor cortex as a peripheral bus)the published neuroscience the design rests on · as of 2003

    We use cookies.