How Does Your Brain Actually Read a Word? Science Finally Has an Answer

You’re doing something extraordinary right now. Your eyes are landing on a word, and within a fraction of a second, your brain knows exactly what it says — in any font, any size, in uppercase or lowercase. How?

You can read the word READING just as easily as reading, or Reading, or even r-e-a-d-i-n-g with spaces between the letters. The shapes are completely different each time, yet you instantly recognize them as the same word. This ability — called invariant word recognition — is one of the most remarkable things the human brain does. Scientists have had theories about how it works for decades, but the underlying neural mechanism — how the brain’s visual circuits actually pull it off — remained elusive.

A 2024 study by Aakash Agrawal and Stanislas Dehaene at the Collège de France offers a compelling answer — one that not only explains the mechanism, but also accounts for a range of observations about reading and dyslexia that earlier theories couldn’t fully address.

The Problem: How Does Your Brain Tell FROM Apart from FORM?

Consider the words FROM and FORM. They contain exactly the same four letters. The only difference is where the letters sit within the word. Your brain must track not just which letters are present, but where each one belongs — and it has to do this no matter what font, size, or position the word appears in.

This is genuinely hard. Scientists had long debated how the brain pulls it off. One influential idea was that the brain encodes ordered pairs of letters (FO, OR, RM for FORM vs. FR, RO, OM for FROM). Another, which accumulated strong behavioral evidence, was ordinal position coding: the brain tags each letter with its position number in the word (F is 1st, O is 2nd in FORM; F is 1st, R is 2nd in FROM).

The ordinal position idea had the better evidence from reading experiments — but the question remained: how does the brain actually compute these position codes? What’s the neural mechanism that lets the brain figure out that O is in the 2nd slot, regardless of where the word sits on the page? That’s the question this new study answers.

The Discovery: Your Brain Uses Spaces as Anchors

The researchers trained an artificial neural network — a type of AI that processes images the way the visual brain does — to recognize written words in multiple languages. After training, they opened up the network to see how it was actually computing letter positions.

The answer lies in how the brain’s visual cortex processes images at different scales. Neuroscientists know that visual neurons respond to features at different spatial frequencies: some are tuned to fine details (the specific curves and strokes that distinguish one letter from another), while others respond to broad patterns (large regions of light or dark — including the blank space surrounding a word).

In the next stage of visual processing, neurons combine these two types of information. The result is cells that are sensitive to both a specific letter and whether there is blank space nearby. For example:

• A cell might respond to “the letter G, with blank space immediately to its left” — effectively detecting G at the beginning of a word
• Another might detect “the letter D, with blank space to its right” — encoding D at the end of a word

Higher-level circuits then pool across many of these detectors to extract each letter’s ordinal position (1st, 2nd, 3rd...) regardless of where the word sits on the page. In other words, the spaces around a word act as anchor points, and the brain uses them to compute the position codes that behavioral research had already predicted.

The Visual Word Form Area: Your Brain’s Letterbox

One of the most striking findings was that the AI spontaneously developed a small cluster of units dedicated specifically to word recognition — they responded strongly to written words but not to faces, objects, or other visual categories.

This mirrors a real region in the human brain called the Visual Word Form Area (VWFA), sometimes called the brain’s letterbox. Located in the left side of the brain’s visual cortex, the VWFA activates specifically when we read. It’s not something we’re born with — it develops as we learn to read, as general-purpose visual circuits get gradually repurposed for the special task of recognizing written words.

The VWFA is tuned to the writing system you learned. If you read English, it responds most strongly to English words. But it still responds somewhat to unfamiliar scripts — suggesting it’s built on a universal visual mechanism that gets fine-tuned by experience.

The Cambridge Effect, Explained

You may have seen this famous text floating around the internet:

Most people can read that scrambled text without much difficulty. But why? The mechanism described above offers a clear explanation.

The first and last letters of each word are the ones encoded most sharply by the brain, because they sit right next to a space. The brain’s space-detecting circuits lock onto these edge letters with high precision. Middle letters, on the other hand, are positioned relative to these anchors with fuzzier, less precise coding.

This means you can scramble the middle letters of a word and the brain still recognizes it — because the precisely coded first and last letters, plus the overall visual word representation, are enough to identify it. But swap the first or last letter, and the word becomes much harder to read.

The Cambridge effect isn’t just a quirky internet phenomenon. It’s a window into the actual architecture of your reading circuitry.

What This Means for Dyslexia

This model of how the brain computes letter positions doesn’t just explain normal reading — it also sheds light on why reading can go wrong in specific, predictable ways.

Letter-Position Dyslexia

Some readers frequently transpose letters within words — reading FORM as FROM, or DESTINY as DENSITY. This is called letter-position dyslexia. The model predicts exactly this pattern: if the brain mechanisms that anchor letters to their positions are imprecise (especially for middle letters, which are encoded with inherently lower precision than edge letters), transposition errors are a natural consequence.

Attentional Dyslexia

In attentional dyslexia, letters seem to migrate between neighboring words — a reader might see “win fed” but read it as “fin wed.” This also makes sense within the framework: if the brain’s space-detection machinery misfires — failing to properly assign letters to the word they belong to — letters can get bound to the wrong word while preserving their relative positions.

The model gives us a mechanistic explanation for why these dyslexias produce their specific error patterns, rather than random mistakes. (For a deeper look at these and other types of reading difficulty, see our guide: Not All Dyslexia Is the Same.)

Why Does Spacing Help Struggling Readers?

One of the best-documented findings in dyslexia research is that extra spacing between letters improves reading fluency for struggling readers. This finding had been somewhat mysterious — why should adding space between letters make words easier to read?

This model offers a clear answer. If the brain relies on spaces as anchor points to figure out where letters belong in a word, then making those boundaries cleaner and more visually prominent should strengthen the very signal the brain depends on. Wider letter spacing may make it easier for the brain’s space-sensitive detectors to do their job accurately.

The Bottom Line

Every time you read a word, your brain runs a remarkable computation: it detects the letters, uses the spaces around the word as reference points to assign each letter its position, and generates a unique pattern of neural activity — a kind of barcode — that identifies the word. This happens three or four times per second, automatically, without effort.

This isn’t magic. It’s a hierarchy of simple detectors, each doing a small job — spotting a letter near a space, or combining many such signals to build a code that works regardless of font or size — that together produce one of the most sophisticated visual capabilities in the animal kingdom.

Understanding how this system works, and how it can break down, is a crucial step toward better tools for identifying reading difficulties early and designing interventions that target the real source of the problem.