Doubling the electrodes on a brain-computer interface does not double the meaningful information a person can send or receive, according to a new Perspective paper by Boxuan Jiang at arXiv (submitted 17 July 2026, q-bio.NC).
The research
Jiang's paper, More Electrodes, Faster Minds? Rethinking Bandwidth in Brain-Computer Interfaces, is a conceptual review rather than a new experiment. It examines how gains in human input and output actually scale with interface capacity — and draws a sharp line between four things that are often blurred together:
- Raw bandwidth — the sheer data rate of the electrode array.
- Decodable neural states — patterns a machine can reliably read out.
- Neural states — the full set of brain activity, decoded or not.
- Information a person can use, confirm, and express — the only layer that matters for communication.
The argument leans on decades of motor-control research: slowly updated task states — a goal, an intention, a target — unfold into rich behavior through the body, sensory feedback, the environment, and shared context. A person can play a fast piece of piano music without issuing hundreds of independent commands per second, because much of the complexity is handled below the level of conscious selection. On the output side, decodable activity can support prediction and control, but subject-level communication depends on selection, confirmation, and authorization — the user has to approve what gets sent. On the input side, stimulation may guide plasticity and speed learning, but embodied skills still arise from coordinated brain, body, and environment. Jiang concludes the scaling relationship is likely nonlinear: more capacity yields real gains up to a point, then hits constraints rooted in embodiment, learning, and personal expression.
Why it matters
BCI headlines often promise accelerated thought output, mind reading, and instant skill acquisition. This paper is a useful reality check — and it applies to ordinary cognition too. Your brain already runs on surprisingly low effective throughput if you measure only conscious, deliberate decisions per second. Most of your fluent reading, typing, and conversation runs on compiled habits and prediction, not moment-by-moment control. That is why chunking — grouping small units into larger ones — makes people faster at chess, music, and mental math. It also means cognitive training aimed at higher-level selection and confirmation may matter more than raw processing speed.
What you can do
- Practice chunking. Deliberately group items — digits, notes, chess patterns — into single units you can hold as one.
- Offload the low level. Automate routine decisions (checklists, templates, keyboard shortcuts) so conscious bandwidth goes to choices that need it.
- Train confirmation. Slow down before acting on fast predictions; verifying output is a trainable skill.
- Move to learn. Embodied practice — handwriting, playing an instrument, physical rehearsal — builds skills that pure screen time often misses.
None of this is about raising a fixed number. It is about using the bandwidth you already have more deliberately. If you want a baseline read on your own working memory, processing speed, and pattern recognition, a short adaptive test can give you a starting point.
Source: arXiv q-bio.NC
Curious about your own brain? Take our free adaptive IQ test or try 306 brain training levels.