OCXLY
CXNeuro · Guide · 14 min read

Neuroinclusive UI design: a practical guide

Interfaces that respect attention, working memory, and sensory load — without asking users to trade their privacy for accessibility. Seven principles, with concrete UI examples for neurodivergent-friendly experiences.

Published · 10 July 2026 Division · OCXLY Neuro Labs Audience · Designers, PMs, Engineers Citations · 12
Reading mode

Neuroinclusive design is the discipline of building interfaces that work for people whose cognition sits outside the assumed neurotypical median — including those with ADHD, autism, dyslexia, dyscalculia, sensory processing differences, and acquired cognitive load from anxiety, fatigue, or trauma[1]. Together this population is not a niche: modern epidemiology puts it at roughly 15–20% of adults[2].

It is often framed as a subset of accessibility, and technically it is — but the WCAG guidelines that dominate accessibility work were written primarily around sensory and motor disability, with cognitive accessibility added later and still under-specified[3]. Neuroinclusive design fills that gap with patterns drawn from cognitive psychology, HCI research, and the lived experience of neurodivergent designers and users.

This guide is grounded in two commitments that shape OCXLY Neuro Labs' research: cognitive accessibility — the interface should not tax attention, memory, or sensory bandwidth for its own decoration — and digital sovereignty — the user, not the platform, should own their attention, their data, and the terms of engagement. The two are inseparable. An interface that hijacks attention through dark patterns is not accessible, no matter how many alt-tags it ships with.

What follows is seven principles, each with the mechanism it addresses, the UI patterns that violate or embody it, and concrete implementation notes.

01

Predictability over novelty

The principle

Neurodivergent users often rely on pattern-based navigation — the memorised location of a button, the shape of a form, the sequence of a checkout — because rebuilding a mental model of an interface on every visit is expensive. A/B tests that shuffle navigation for optimisation, "surprise and delight" onboarding, and gamified layout randomisation all trade a small conversion lift for a large orientation tax on the users who most need stability[4].

In practice

Avoid
  • Rotating hero content that changes on each visit
  • Nav items that reorder based on "recency" or ML personalisation
  • Onboarding that hides features until "unlocked"
  • Buttons that swap labels between "Save" and "Update" contextually
Prefer
  • Stable primary navigation with consistent order across sessions
  • Fixed positions for high-frequency actions (search, cart, account)
  • Explicit "What's new" panels instead of stealth UI changes
  • Consistent action verbs across similar surfaces
02

Motion as a budgeted resource

The principle

Involuntary motion in the periphery triggers the salience network, which redirects attention regardless of the user's intent[5]. For neurotypical users this costs a fraction of a second. For users with ADHD, autism, or vestibular sensitivity, it can end a reading session, trigger nausea, or provoke a migraine[6]. WCAG 2.3.3 (Animation from Interactions) codifies the baseline — but neuroinclusive design goes further: motion is opt-in for delight, opt-out for function.

~35%
of the general population reports vestibular or motion sensitivity strong enough to affect UI use; the figure is higher among autistic and post-concussion users.[6]

In practice

Honour @media (prefers-reduced-motion: reduce) everywhere, not just on decorative pages. Beyond that, ship an explicit calm mode — a single toggle that disables ambient motion, parallax, auto-playing media, and non-essential transitions — and persist it in localStorage so it survives navigation without a server call. This site's own calm mode is the reference implementation: one control, one storage key, no telemetry.

03

Reading as a first-class mode

The principle

Dyslexia affects roughly 10% of the population, and typographic choices measurably affect reading speed and comprehension for that group[7]. Yet most design systems ship a single typographic scale optimised for aesthetic density, not readability. The gap between "beautiful body copy" and "readable body copy" is where a dyslexia-friendly mode belongs.

In practice

Avoid
  • Justified text (creates uneven word spacing / rivers)
  • Line-heights below 1.5 for long-form content
  • Line-lengths above 90 characters
  • Pure white on pure black (halation for many readers)
Prefer
  • Left-aligned text with a 1.6–1.8 line-height for body
  • A dyslexia-friendly toggle that swaps to Atkinson Hyperlegible or similar
  • Line-length capped at 65–75 characters
  • Off-white on near-black (e.g. #eaeaf0 on #0a0a12)
04

Sensory load has a ceiling

The principle

Autistic users and users with sensory processing differences describe UI as "loud" or "quiet" — a metaphor that maps precisely to what cognitive load research measures[8]. Simultaneous channels of colour, motion, sound, and density compete for the same limited attentional pool. A single loud channel is manageable; four moderate channels stacked on one screen is not.

In practice

Audit each surface for its channel budget: how many things move, how many hues compete for salience, how many sounds can play unprompted, how dense the information grid is. Then cut. Provide a low-contrast option for users who find high-contrast palettes overwhelming (contrary to WCAG's asymmetric emphasis), and never autoplay audio without a persistent global mute — mirroring the sound control this site keeps visible in the top corner.

05

Working memory is small

The principle

The classical Miller estimate of "7 ± 2" chunks has been revised downward — current research puts functional working memory closer to 3–4 chunks, and lower still under stress or in ADHD[9]. Any interface that requires the user to remember a value across surfaces — a code from one screen typed into another, a filter set that vanishes on back-navigation, an error from step 4 that must be recalled at step 1 — is imposing an invisible tax.

In practice

Avoid
  • Multi-step flows without a persistent step indicator
  • Filters that reset on pagination
  • Error messages that name the field without linking to it
  • Confirmation codes shown on one screen, entered on another with no autofill
Prefer
  • Visible progress: "Step 2 of 4 — Shipping"
  • URL-persisted filter state and back-navigation preservation
  • Inline validation with scroll-into-view on error
  • OTP autofill via autocomplete="one-time-code"
06

Language is an interface

The principle

Autistic users, users with intellectual disability, non-native readers, and users under cognitive load all benefit from literal, plain-language copy. Idiom, metaphor, and clever product-marketing voice — the industry's default register — measurably reduce comprehension in these groups[10]. Copy is not decoration; it is the primary interface for many actions.

In practice

Write button labels as the verb they perform ("Delete account" not "Peace out"). Prefer literal error messages ("Password must contain a number") over cute ones ("Oops, that didn't work"). Avoid negation in confirmation dialogs — "Don't cancel" and "Cancel" is a classic double-negative trap. Aim for a reading age of roughly 12 for consumer surfaces; check with the Flesch-Kincaid or SMOG index in editing.

07

Sovereignty is accessibility

The principle

The final principle is the one usually left out of accessibility guides: an interface that manipulates the user is inaccessible to anyone whose cognitive resources are already stretched. Dark patterns — pre-checked opt-ins, obscured unsubscribes, engagement loops tuned by ML on vulnerable users — extract attention and consent from the people least able to defend against them[11]. Neurodivergent users, older users, and users in acute stress absorb this disproportionately[12].

In practice

Treat consent as a first-class UX surface: symmetrical "Accept" and "Reject" buttons in equal visual weight, granular category toggles, and a persistent way back to the settings after the first choice. Prefer local-first storage for user preferences (calm mode, dyslexia toggle, language) so they survive without an account and without server telemetry. Publish an honest data-sharing statement — including for AI training crawlers — in robots.txt and ai.txt. Sovereignty at the infrastructure layer is what makes sovereignty at the UI layer credible.

The bigger picture

Neuroinclusive design is not a specialisation; it is the direction general design is already moving, once the assumptions about a median user are unpicked. The seven principles above overlap heavily with "just good design" — clearer language, less noise, more control — and every one of them helps the neurotypical user too. The difference is who counts as a rounding error.

Three concrete next steps for teams:

1. Ship a calm mode and a dyslexia toggle as core UI, not settings-page trivia. Give them a visible entry point and persist them locally. Two toggles cover a large fraction of the population that most sites currently exclude.

2. Add a neurodivergent participant to every research round. Not a separate audit — one participant per study who self-identifies. The patterns above surface within a session.

3. Publish your data posture. A cognitively accessible interface built on opaque data extraction is a contradiction. If you want neurodivergent users to trust your product, treat their attention and their data as theirs.

OCXLY publishes tools, articles, and interfaces to this brief. If you're building along the same lines — or finding somewhere we're falling short of it — [email protected] is a real inbox read by a real person.

References

  1. Doyle, N. (2020). Neurodiversity at work: a biopsychosocial model and the impact on working adults. British Medical Bulletin, 135(1), 108–125. doi:10.1093/bmb/ldaa021
  2. Cortiella, C., & Horowitz, S.H. (2014). The state of learning disabilities: Facts, trends and emerging issues (3rd ed.). National Center for Learning Disabilities.
  3. World Wide Web Consortium (W3C). (2023). Making Content Usable for People with Cognitive and Learning Disabilities. w3.org/TR/coga-usable.
  4. Krug, S. (2014). Don't Make Me Think, Revisited: A Common Sense Approach to Web Usability (3rd ed.). New Riders. Chapter 6 on "conventions vs. innovation".
  5. Corbetta, M., & Shulman, G.L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3(3), 201–215. doi:10.1038/nrn755
  6. Golding, J.F. (2016). Motion sickness. Handbook of Clinical Neurology, 137, 371–390. doi:10.1016/B978-0-444-63437-5.00027-3
  7. Rello, L., & Baeza-Yates, R. (2013). Good fonts for dyslexia. Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '13). doi:10.1145/2513383.2513447
  8. Robertson, C.E., & Baron-Cohen, S. (2017). Sensory perception in autism. Nature Reviews Neuroscience, 18(11), 671–684. doi:10.1038/nrn.2017.112
  9. Cowan, N. (2010). The magical mystery four: How is working memory capacity limited, and why? Current Directions in Psychological Science, 19(1), 51–57. doi:10.1177/0963721409359277
  10. Nicolle, C., & Abascal, J. (Eds.). (2001). Inclusive design guidelines for HCI. CRC Press. Chapter on cognitive and language considerations.
  11. Gray, C.M., Kou, Y., Battles, B., Hoggatt, J., & Toombs, A.L. (2018). The dark (patterns) side of UX design. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. doi:10.1145/3173574.3174108
  12. Mathur, A., Acar, G., Friedman, M.J., Lucherini, E., Mayer, J., Chetty, M., & Narayanan, A. (2019). Dark patterns at scale: Findings from a crawl of 11K shopping websites. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW). doi:10.1145/3359183