Attention is a mute button, not a spotlight

In June 2026, a Johns Hopkins lab published a result in Nature Communications that quietly rearranges how you should think about every screen you design. Ninad Kothari and Shreesh Mysore found a small cluster of inhibitory neurons deep in the mouse brainstem — a network they label the PLTi — whose whole job is to turn down the things the animal isn’t attending to. Silence those neurons and the mouse doesn’t go blind. It goes distractible. It can still see everything; it just can’t stop seeing everything. Reactivate them and focus snaps back, cleanly, with vision and movement untouched the whole time.

The surprising part isn’t that the brain has an attention circuit. It’s where this one sits and what it does. It lives below the prefrontal cortex — the “executive” we usually credit with concentration — and it’s evolutionarily ancient, older than the cortex by a few hundred million years and shared across vertebrates. (It’s a mouse study; the human version is inferred from that shared lineage, not yet shown, so hold it loosely.) And it doesn’t create focus by amplifying the thing that matters. It creates focus by suppressing the things that don’t.

That is the whole essay. Attention is not a spotlight you aim. It’s a mute button the brain holds down on everything else.

Builders get this backwards constantly. When we want someone to notice something, we make it bigger, brighter, redder, animated — we amplify the hero. But if attention is fundamentally a suppression system, then a screen with one loud element and nine slightly-less-loud elements is a screen where the user’s brain is quietly running nine suppression jobs to see one thing. You didn’t make the hero win. You made the fight harder and hoped it would.

The reframe is simple and it changes real decisions: focus is a subtraction problem, not an addition problem.

Concrete decisions this flips:

The empty state and the first screen — the fastest way to make the primary action obvious is to delete its rivals, not enlarge it. One clear call, nothing else competing for the suppressor. “Focus mode,” reading mode, do-not-disturb — these work not because they add focus but because they remove suppression load; design them as acts of removal and judge them by what you took away, not what you left. The red notification dot — a badge doesn’t add attention, it adds one more rival that every unrelated glance now has to suppress. A notification is a tax on the suppression circuit, and you’re charging it forever, on every screen, not only when the user opens the thing.

Even the von Restorff trick — make one item different so it pops — is really this mechanism wearing a costume. The odd item stands out because everything else is uniform and therefore cheap to suppress as a group. The moat isn’t the highlight. It’s the sameness around it.

Here’s the uncomfortable version for anyone in the habit of adding features: every element you ship is not just something a user might use. It’s something the user’s brain must actively silence to use anything else. The real cost of a control isn’t its pixels — it’s the suppression it demands from every unrelated task, for the entire life of the product.

The next time a stakeholder asks how to make the CTA pop, the honest answer is usually “remove three things next to it.” The brain already ships with world-class hardware for finding the signal. Your job isn’t to shout over the noise. It’s to give it less noise to mute.

The science, to look up: PLTi / brainstem inhibitory neurons and selective attention (Kothari & Mysore, Johns Hopkins, Nature Communications, 2026); attention as competitive selection and distractor suppression more broadly (biased-competition account, Desimone & Duncan, 1995). The mouse-to-human step is a conservation-based hypothesis, not a demonstrated human result — hedge it.

The news, to check: “Scientists discover ancient neurons that control attention,” Johns Hopkins Hub, June 22 2026; coverage at Neuroscience News and ScienceAlert (June 2026).