Blog · Workflows

Reading research papers with dyslexia

11 min read · May 27, 2026

A research paper is the worst-case scenario for a dyslexic reader. Two narrow columns in 10pt Times New Roman, justified to ragged rivers of white space, broken up with superscript citations, in-line equations, and abbreviations the authors invented three paragraphs ago. Generic dyslexia advice - "use a sans-serif, increase the line height" - bounces off the format. The fix is a layered workflow, not a single setting, and it has to work on whatever PDF the journal hands you.

Why papers are uniquely hostile

It is not your imagination. Almost everything academic typography does is the opposite of what dyslexia research recommends.

Two-column layouts force the eye to perform a long, error-prone return sweep at the end of every line. For a dyslexic reader, that sweep is exactly where you lose your place - we wrote about this mechanic in dyslexia and eye tracking. Justified text creates uneven word spacing and visible rivers, which we covered in justified vs left-aligned text. The default body font is almost always Times New Roman, which we have a whole post on - why Times New Roman is hard for dyslexic readers. Footnotes and superscript citations interrupt every other sentence. And the entire thing is rendered as a fixed-layout PDF, which actively resists reflow.

None of these choices were made to hurt you. They evolved for paper publication in tight page-count economies. But the conventions outlived the medium, and you are reading them on a screen that could reflow the text in half a second if the format would let it.

The short answer

Get the paper out of its native PDF form before you try to read it. Reflow it into a single column, swap the font, widen the line spacing, then anchor on structure (abstract, headings, figure captions) before reading the body. Treat citations as visual noise to skip on first pass. Recover before the discussion section.

You will not turn a 12-page paper into a magazine article. You will turn it from "I bounced off after column one" into "I got through it and remember the argument."

The six-step workflow

1Reflow the PDF

This is the single biggest lever. A reflowed PDF stops being a fixed two-column page and becomes a single column of text you can resize, restyle, and scroll. There are three reliable ways to do it.

Open it in Chrome's PDF viewer and use the "Open in browser" reader extension of your choice. Most modern reader-mode extensions (Mercury Reader, Just Read, the built-in Edge Immersive Reader) now handle PDFs as well as web pages. They strip the layout, extract the text in reading order, and render it as a single column. The conversion is imperfect with multi-column papers - tables and figures sometimes land in odd places - but the body text comes out clean. See reader mode vs reading extensions for dyslexia for the trade-offs.

Upload it to a reflow service. Tools like Adobe Acrobat Reader's "Liquid Mode" (free, mobile and desktop) and the open-source pdftotext -layout followed by a markdown re-render do the same job. Liquid Mode is the easiest entry point: open the PDF in Acrobat, tap the droplet icon, and the paper reformats into a phone-friendly single column with adjustable font, size and spacing. It does not preserve the page numbers, so keep the original open in a second tab for citation work.

Use a research-specific tool. Zotero's built-in PDF reader, Readwise Reader, and Paper QA all render PDFs in a single-column view with font controls baked in. If you read papers regularly, picking one and living in it pays back the setup cost within a week.

The full workflow for stripping a PDF down to readable plain text is in our dyslexia-friendly PDF reading in Chrome piece. Everything in that post applies here; the only difference is that academic PDFs reflow worse, so expect to clean up a paragraph or two by hand.

2Swap the font

Once you have the paper in something that can render text, change the font. The defaults inside reader apps are usually a generic serif or sans that was not chosen for dyslexic readers. Pick from the small set that actually helps:

A good starting set for academic prose OpenDyslexic if letter rotation (b/d, p/q) is part of your experience.
Lexend if your trouble is reading rate and you want to plough through faster.
Atkinson Hyperlegible if your trouble is letter-by-letter disambiguation, especially with numerals and abbreviations - very common in scientific papers.

For papers heavy with abbreviations, units, and numbers - clinical trials, physics, chemistry, anything with tables of values - Atkinson Hyperlegible tends to outperform the others. It was designed for low vision, and the side effect is that "Il1", "O0", and "rn/m" stop blurring into each other. For long, prose-heavy papers in the humanities or social sciences, Lexend usually feels lighter. The full comparison is in Lexend vs Atkinson Hyperlegible.

If you are reading the paper in your browser - any preprint server (arXiv, bioRxiv, SSRN) lets you view the HTML version when one exists - LexiFont applies your chosen font with one click across the page, including the tables and the footnotes. That HTML version is almost always more readable than the PDF and tends to be the path of least resistance when it exists.

3Single-column it (even if reflow failed)

If reflow gave you a messy result and you have to read the PDF in its native form, force a single-column experience another way. The simplest is to zoom in until only one column fits on screen and scroll vertically through column one, then column two. It feels primitive but it eliminates the return sweep that costs you the most.

A better option on a wide monitor is to rotate the PDF 90 degrees and view two pages side by side at high zoom, so each "page" on screen is one column. Most PDF viewers (Preview, Acrobat, Foxit, Chrome) support this. It is slower than reflow but preserves the exact layout, which matters when you need to cite something by page and column.

Increase the line height in whichever viewer you end up using. The research on line spacing and letter spacing for dyslexia is consistent: pushing line height from the typesetter's default (usually 1.15) to 1.5 or higher dramatically reduces line-tracking errors. If your reader does not let you adjust line height directly, this is another argument for reflowing into a tool that does.

4Anchor on structure before reading the body

Research papers are written to a fixed scaffold: abstract, introduction, methods, results, discussion, references. The scaffold is the most useful part of the paper for a dyslexic reader, because every section telegraphs what the next one will say. Spending three minutes building a mental map of the paper before you read it cuts the cognitive load of the body text in half.

A working pre-read pass:

Read the abstract twice. Slowly. This is where the entire argument lives.
Read every section heading in order, without reading the sections.
Read every figure and table caption. Captions in good papers are self-contained mini-summaries.
Read the first sentence of the discussion and the last sentence of the conclusion.

By the time you start the body text, you already know what the paper claims, how it argues, and what the result was. Working memory capacity is the bottleneck for dyslexic readers on long, dense material (we covered the mechanism in dyslexia and working memory); pre-loading the structure frees up working memory you would otherwise spend trying to predict what is coming.

5Treat citations as visual noise on the first pass

Numbered references like "[14]" or author-year like "(Smith et al., 2019)" interrupt every sentence in a research paper. Each interruption is a chance for your eye to fall off the line. On a first read, ignore them completely. You are not chasing references on the first pass; you are building the argument in your head. The citations are there for the second read, if there is one.

If the paper format allows it, hide superscripts entirely. In a reflowed view, you can often select-and-delete them; in your browser, a small custom user-style ("sup, .citation { display: none; }") suppresses them for the session. For browser-side custom CSS, the same approach we describe in dyslexia-friendly news sites works on most journal sites.

6Recover before the discussion section

The discussion section is where the paper does its real argumentative work, and it is also the section you will be most fatigued at. Build in a recovery pause after the results section: stand up, look at something twenty feet away for two minutes, drink some water. Then read the discussion as if it were its own short paper - because functionally, it is. Most dyslexic readers can sustain about 20-30 minutes of dense academic reading at a stretch before fatigue starts to compound. The pattern we lay out in how to read long articles with dyslexia applies in miniature to each paper.

Working with terminology you have never seen

Specialist vocabulary is its own problem. Dyslexic readers have a harder time recognising unfamiliar words by shape, because the shape-recognition shortcut depends on having seen the word before. A paper that introduces "supramarginal gyrus", "heterogeneous nucleation", or "type II error" in the first paragraph will slow you down more than it would slow a non-dyslexic reader.

Two habits that help:

Pre-skim the methods section for jargon, then look up the three most-repeated unfamiliar terms before you read the body. Spending five minutes building a small glossary up front saves you twenty minutes of mid-read context switching. You can keep the glossary in a sticky note, a notes app, or annotate the PDF in the margin.

Read the term aloud once, then write it. The motor and auditory loops give your brain extra anchors for an unfamiliar string of letters. This is the same trick that helps with names: a word you have only ever seen is harder to recognise next time than a word you have heard yourself say.

A note on screen vs paper

A lot of academics still recommend printing papers to read. For dyslexic readers, that advice is almost always wrong. Print loses every adjustability that screen gives you: font, size, line height, background colour, dark mode, zoom. If your only complaint about screens is fatigue or glare, the answer is a matte screen filter, a warmer background colour, and a 20-20-20 visual break habit - not paper. We covered the background-colour question in detail in background colours for dyslexia.

There is one case where paper wins: skimming for structure, where you want to flip between sections rapidly. If you read most papers this way, print is fine. For careful reading of the body text, the screen with the right setup is the easier surface.

Tools that pull this together

You do not need a stack of five apps. A working minimum:

A reflow tool. Adobe's Liquid Mode is free and works on desktop and mobile. If you have a paid Acrobat license you also get the "Read Out Loud" voice, which some dyslexic readers use as a third channel alongside reading.
A font-override extension for the browser. LexiFont applies OpenDyslexic, Lexend, or Atkinson Hyperlegible across journal sites, preprint servers and HTML versions of papers. Pro adds the full font set and the per-site overrides that matter when you are switching between a JAMA HTML page, an arXiv preprint and a PDF preview in the same session.
A reference manager with a built-in PDF reader. Zotero is free; Readwise Reader is paid. Both reflow and let you set font and spacing. If you read more than five papers a week, the time saved on cleanup is real.
A text-to-speech voice you can stand listening to. macOS's built-in voices have improved dramatically; on Windows the Microsoft Natural voices in Edge are passable. Reading dense paragraphs aloud (or having them read aloud) gives you an auditory channel that takes some load off the visual one.

A common combined setup: open the paper's HTML version in your browser with LexiFont applying Atkinson Hyperlegible at 18pt, with Zotero open in a side panel for citation parking, and Microsoft Edge's Immersive Reader as a fallback for any paper that does not have a clean HTML version. That covers about 90% of papers without changing tools.

A realistic expectation

None of this turns a dense methods section into easy reading. Research papers are written for an audience expected to bring substantial prior knowledge and a high tolerance for friction; that friction is part of the genre. What the setup above does is remove the extra friction that the typography adds. You still have to do the cognitive work of understanding the paper. You just stop also doing the visual work of fighting the layout.

If you are a student, a clinician, a researcher, or anyone whose job requires reading the primary literature, the time you invest in this workflow pays back across every paper you read after it. The marginal cost of the second paper is much lower than the first; by the tenth, you have a habit.

Get LexiFont Pro - OpenDyslexic, Lexend, Atkinson Hyperlegible and Comic Neue for $14.99 one-time