In September 2024 BROADMIND exhibited as an Industrial Partner of the 9th Graz Brain-Computer Interface Conference. It is the room where the field argues with itself — and we wanted to be in it.
September 2024BROADMIND
BROADMIND in the Industrial Partners block of the official GBCIC2024 programme, page 20 (TU Graz).
A consumer neurotech company has an awkward relationship with the research community. We make claims that only that community can check. We ship hardware to people who will use it differently from how we tested it. And we do all of this in a category where the phrase "brain-computer interface" carries twenty years of academic expectation behind it.
So when the 9th Graz Brain-Computer Interface Conference came around, exhibiting was not a marketing decision. It was the closest thing to a public audit we could buy.
The conference
GBCIC2024 ran 9–12 September 2024 at TU Graz, on the Campus Neue Technik at Stremayrgasse 16, under the title "Join Forces – Increase Performance." It is organised by the Institute of Neural Engineering and chaired by Gernot R. Müller-Putz, and in 2024 it was a partnered event of the International BCI Society.
The scale is worth stating plainly, because BCI conferences are often imagined as small specialist gatherings. This one drew contributions from over 22 countries — close to 100 peer-reviewed papers from roughly 476 authors, each reviewed by at least two reviewers, with the proceedings published by Verlag der TU Graz (DOI 10.3217/978-3-99161-014-4-000). Five keynotes anchored the programme: Andrea Kübler, Jennifer Collinger, Camille Jeunet-Kelway, Nick Ramsey and Henri Lorach. Around them: satellite workshops before the main days, talks, poster sessions, an industrial exhibition, and — after a break of several years — a closing tour to the South Styrian vineyards.
The title was not decorative. The conference description makes the point directly: the invasive and non-invasive communities have increasingly stopped treating each other as rivals and started forming one field. That convergence is the most important thing happening in BCI right now, and Graz is where you can watch it happen in a corridor.
Where we stood
BROADMIND appears on page 20 of the official programme, the page headed "The conference organizers gratefully acknowledge the sponsors and supporters," inside the Industrial Partners block.
The companies in that block tell you a lot about the neighbourhood we were keeping:
g.tec medical engineering
MES Forschungssysteme
ANT Neuro
Brain Products
Bitbrain
mBrainTrain
BROADMIND
These are the vendors whose amplifiers most BCI papers in the last decade were recorded on. Being listed beside them is the specific thing we wanted, and it came with a specific obligation: nobody in that hall was going to be impressed by a render.
For context on the tiers — the award sponsors that year were NIRx (best poster) and Zander Labs (best talk), each putting up €500 prizes; keynote sponsorship came from BioTechMed Graz and the Initiative Gehirnforschung Steiermark; the opening cocktail was sponsored by NeuroXess; local support from TU Graz and the Stadt Graz. We were not in those tiers. We were in the exhibition, which is exactly where a hardware company should be.
What we were asked
The questions at Graz were different in kind from the ones at an XR expo or a general tech conference. Nobody asked what was exciting. They asked what was measurable.
"What is your sampling rate and reference scheme?" — asked before anyone asked what it was for.
"Can I get the raw signal, or only your features?" — the same question developers ask, but here it came with a plan to reanalyse the data and publish it.
"How do you handle artefacts when the subject moves?" — the question that separates a wearable from a toy.
"What happens for the users your decoder does not work on?" — the hardest one, and the one we think about most.
That last question is the field's own unresolved problem. The literature has acknowledged for years that a meaningful share of users — commonly cited at up to 30% — show poor or unreliable control under standard paradigms. It is not a defect in any one device; it is a mismatch between fixed decoding assumptions and brains that differ. For a research lab that number is an interesting result. For a consumer product it is a shipping blocker.
We do not claim to have solved it. What we can say is that the architecture has to assume it from the beginning: raw access, transparent processing, and hardware that does not hide the signal behind a proprietary layer. If a researcher cannot reproduce our numbers, they should at least be able to see exactly why.
Why the consumer/lab gap matters to us
There is a version of this industry where consumer EEG and academic BCI are two separate worlds that occasionally nod at each other. We think that version is worse for both.
Labs have spent two decades building the methods — artefact rejection, spatial filtering, transfer learning, user training. Consumer hardware has spent a decade building something the labs have not: volume. Thousands of devices in homes generate data no study will ever collect. The gap is not in capability; it is in the interface between the two, and it is mostly a documentation and data-format problem.
That is the part we can influence. It is also why the phrase in our own copy is lab-quality, not lab-equivalent. The honest claim is that a well-documented, raw-data-first consumer device can be useful to research — not that it replaces an amplifier that costs as much as a car.
The short version
Graz is the serious room: ~100 peer-reviewed contributions, ~476 authors, 22+ countries, five keynotes, a partnered BCI Society event.
BROADMIND was an Industrial Partner in the official programme, listed with the established EEG vendors of the field.
At a scientific conference you are not asked whether a product is exciting, but whether its claims survive reanalysis.
The field's open problem — decoders that fail a meaningful share of users — is the same problem that decides whether BCI ever becomes a consumer interface.
The gap worth closing is not hardware capability but the interface between consumer devices and research practice: raw data, open formats, reproducible numbers.
We went to Graz to be seen by the people who could call our bluff. The most useful outcome was not the foot traffic. It was coming home with a clearer list of questions our next datasheet has to answer.
Postscript. The 10th Graz BCI Conference took place in September 2026 — over 190 participants, more than 100 papers, and a closing panel titled "Science or Application?" That question is still the right one, and the fact that the community keeps asking it is why the field is worth building in.
BROADMIND builds non-invasive EEG hardware with unrestricted raw-data access and an open, developer-friendly software path. Researchers and developers: get in touch.
BROADMIND in the Industrial Partners block of the official GBCIC2024 programme, page 20 (TU Graz).