Dyslexia, Redefined: What Genetics and Neuroscience Are Changing About How We Understand Reading

The Society Journal | Research Insight | October 2026

Beyond letter reversals, IQ discrepancies and the search for a single dyslexia gene

Abstract

Dyslexia is among the most widely recognised learning differences, yet some of the ideas most strongly associated with it, seeing letters backwards, possessing a distinctive intelligence profile, carrying a particular “dyslexia gene”, or having an identifiable “dyslexic brain”, do not accurately reflect contemporary scientific evidence.

In October 2025, the International Dyslexia Association (IDA) adopted its first major revision of its dyslexia definition since 2002. The scientific rationale, published in Annals of Dyslexia in April 2026, reflects an important conceptual shift. Dyslexia is defined primarily through persistent difficulties in word reading and/or spelling involving accuracy, speed or both. Its causes are described as complex combinations of genetic, neurobiological and environmental influences interacting throughout development. Phonological and morphological processing difficulties are recognised as common, but not universal.

At the same time, genetics has moved from investigating a small number of candidate genes towards analysing populations of more than one million participants. A 2025 genome-wide study identified 80 independent loci associated with dyslexia and reading ability, yet polygenic indices derived from the study explained only a modest proportion of reading differences in an independent cohort.

Neuroscience presents a similarly complex picture. Brain differences associated with reading difficulty exist at the group level, but contemporary evidence has not produced a biological marker capable of diagnosing dyslexia in an individual.

The emerging scientific picture is therefore neither one of biological determinism nor of a single cognitive deficit. Dyslexia increasingly appears as a multifactorial developmental outcome produced through multiple interacting pathways.

This Research Insight examines what has changed, what older assumptions are being reconsidered, and what the latest research means for how dyslexia should be understood.

Keywords: dyslexia, reading, genetics, neurodevelopment, neuroscience, literacy, phonological processing, learning differences

Editorial approach

This article is a narrative research synthesis rather than a systematic or scoping review.

It draws primarily on the International Dyslexia Association's revised 2025 definition and its 2026 scientific rationale, recent large-scale genomic research, genetic reviews, and a 2026 interdisciplinary review of developmental language and reading disorders.

The purpose is not to propose a new diagnostic framework. It is to examine how advances in research are changing several common assumptions about what dyslexia is,and what it is not.

A definition changed because the science changed

On 22 October 2025, the International Dyslexia Association formally adopted a revised definition of dyslexia.

It was the organisation's first major revision since its influential 2002 definition.

The change followed an extensive process involving a steering committee, advisory committee, examination of contemporary research and more than 3,000 responses to a public consultation. The scientific justification was subsequently published by Catts, Haynes and Joshi in Annals of Dyslexia on 29 April 2026.

The resulting definition is significant not simply because some words changed.

It reflects a different scientific model.

Dyslexia is now described as a specific learning disability characterised principally by persistent difficulties in word reading and/or spelling, affecting accuracy, speed or both.

Those difficulties exist on a continuum of severity rather than forming a biologically neat binary category.

They can also vary across writing systems—or orthographies—because learning to read English is not neurologically or linguistically identical to learning to read Finnish, Arabic, Chinese or another writing system.

Perhaps the most important change concerns causation.

The revised definition states that dyslexia has complex causes involving combinations of:

genetic influences,

neurobiological influences,

and environmental influences

that interact throughout development.

That language represents a substantial departure from searching for one defining deficit.

Dyslexia is not simply seeing letters backwards

Few dyslexia myths are as persistent as the idea that dyslexic people see letters or words in reverse.

Letter reversals can occur during literacy development. But they are not a defining characteristic of dyslexia, nor does their absence rule dyslexia out.

The International Dyslexia Association describes the notion that people with dyslexia “read backwards” as a myth. Research surveying professionals involved in dyslexia assessment has also found belief in letter reversal to remain one of the most common misconceptions surrounding the condition.

The revised IDA definition does not define dyslexia through visual reversal.

Instead, the defining academic difficulty lies at word level: acquiring sufficiently accurate and/or fluent word reading and spelling.

That distinction matters.

Someone might reverse letters during writing and not have dyslexia.

Someone might have significant dyslexia and rarely reverse letters.

Reducing dyslexia to reversals can therefore delay recognition of the difficulties that actually matter.

It can also encourage interventions directed at visual symptoms rather than literacy processes.

Modern dyslexia science is much more concerned with how written language is acquired than with the idea that print is literally perceived backwards.

Phonological processing still matters; but it is no longer the whole explanation

Another important change concerns phonology.

Phonological processing refers broadly to how people represent, access and manipulate the sound structure of spoken language.

For decades, phonological difficulties have occupied a central position in theories of dyslexia, and the evidence connecting phonological awareness with reading development remains substantial.

The revised definition does not reject that evidence.

Instead, it adds an important qualification:

phonological processing difficulties are common but not universal among people with dyslexia.

The same definition also gives greater recognition to morphological processing—the ability to understand meaningful units within words, such as roots, prefixes and suffixes.

Why does this distinction matter?

Because if one underlying cognitive difficulty were both necessary and sufficient for dyslexia, it should reliably identify almost everyone who develops the condition.

Research does not show that.

The scientific rationale for the new definition discusses studies in which only around half of children who later developed dyslexia showed a significant early phonological-awareness deficit. Conversely, not every child with a substantial phonological-awareness difficulty subsequently developed dyslexia.

This does not diminish the importance of phonology.

It changes its status.

Rather than being the single cause of dyslexia, phonological difficulty becomes one important risk pathway within a larger developmental system.

Intelligence has also moved out of the definition

Historically, dyslexia was often described as an “unexpected” reading difficulty.

In practice, the concept of unexpectedness frequently became tied to intelligence.

A child whose reading attainment was substantially below what their IQ score appeared to predict might qualify for identification, while another child with equally serious reading difficulty—but a lower IQ score—might not.

This became known broadly as an IQ–achievement discrepancy model.

The revised IDA definition intentionally removes the reference to “other cognitive abilities” that appeared in the 2002 definition.

The scientific rationale is clear: persistent word-level literacy difficulty occurs across the range of intellectual ability, and general intelligence does not meaningfully distinguish the nature of dyslexic reading problems or predict who will respond to appropriate reading intervention.

This has significant consequences.

A person does not need unusually high intelligence for their reading difficulty to constitute dyslexia.

Nor should struggling readers with lower cognitive scores automatically be considered less deserving of effective literacy intervention.

The change shifts attention away from the question:

“Is this person's reading unexpectedly poor given their intelligence?”

towards the more direct question:

“Does this person have persistent difficulty acquiring accurate and fluent word reading and/or spelling?”

That is a profound conceptual difference.

Genetics is transforming dyslexia science

If dyslexia does not arise from one cognitive deficit, perhaps it arises from one genetic difference?

Modern genomics strongly suggests otherwise.

Reading ability has a substantial genetic component. Twin research has historically estimated the heritability of reading and dyslexia-related traits at approximately 40–60%, although heritability is a population statistic rather than a measure of how “genetic” one person's dyslexia is.

Heritability also does not imply the existence of a single gene.

Instead, reading ability appears to be highly polygenic.

Polygenic means that many genetic variants contribute small amounts to variation in a trait.

The scale of that complexity became much clearer in 2025.

More than 1.2 million participants; and still no “dyslexia gene”

Mountford and colleagues conducted the largest genetic analysis of reading-related phenotypes reported at the time.

The researchers combined information from two major genetic datasets:

a GenLang genome-wide analysis of word reading involving 27,180 participants

and a 23andMe dyslexia dataset comprising 51,800 cases and 1,087,070 controls.

Using a multivariate genome-wide method, they increased the effective sample size to approximately:

1,228,832 participants.

The analysis identified:

80 independent genome-wide significant loci.

Thirty-six genomic regions had not previously been reported as genome-wide significant in dyslexia research.

Thirteen represented novel regions with no previous significant association with dyslexia.

At first glance, that might sound as though scientists were closing in on a genetic test for dyslexia.

They were not.

The same study demonstrates why genetic findings must be interpreted carefully.

When researchers constructed polygenic indices from their results and tested them in an independent cohort of 6,410 people, those indices explained approximately 2.34% to 4.73% of variation in reading traits.

In other words, very large genomic datasets can reliably identify genetic influences at the population level while still being unable to predict an individual's reading ability with anything approaching diagnostic certainty.

This distinction is fundamental.

Population genetics can reveal biological architecture without producing an individual diagnostic test.

There is no single dyslexia gene

Another 2025 investigation illustrates the same point from a different direction.

Dobrynin and colleagues reviewed approximately four decades of research concerning the genetics of specific reading disability and reading-related processes.

They compiled 175 putative candidate genes reported across the literature.

But these should not be understood as 175 genes whose biological purpose is “reading”.

Reading is evolutionarily extremely recent.

Human beings have not evolved a dedicated reading organ or a set of genes specifically designed to decode written language.

Instead, literacy depends upon neural systems that evolved for other functions and are recruited during reading development.

The candidate genes identified by Dobrynin and colleagues were largely evolutionarily ancient. Their analyses linked different groups of genes with processes including early brain morphology and later synaptic signalling.

Their conclusion directly challenges the idea of reading-specific genes.

What genetics appears to influence are biological systems involved in brain development and function that, in combination with many other factors, contribute to how readily literacy skills develop.

There is therefore no known “dyslexia gene”.

There are genetic differences that contribute probabilistically to reading-related variation.

Those are very different claims.

Genetic risk is probabilistic, not destiny

This leads to another important distinction.

A genetic association does not mean a predetermined outcome.

The scientific rationale behind the revised IDA definition explicitly characterises genetic influences on dyslexia as probabilistic rather than deterministic.

A child may inherit combinations of genetic variants associated statistically with greater reading difficulty and nevertheless develop relatively strong reading skills.

Another child may carry a different combination of genetic susceptibilities and experience significant difficulty.

The reason is that genes do not operate independently of development.

Reading emerges through interaction among many systems:

spoken language,

phonological learning,

morphological knowledge,

attention,

memory,

instruction,

print exposure,

educational opportunity,

orthography,

brain development,

and other individual characteristics.

Genes influence parts of this developmental landscape.

They do not dictate a finished reading outcome.

Genes and environments are not opposites

One of the most important developments in contemporary developmental science is the recognition that asking whether dyslexia is “genetic or environmental” poses the wrong question.

Genes and environments interact.

A major 2026 review by van Bergen, Dawson, James and Krishnan examining developmental language and reading disorders estimates that roughly half of population variation in language and reading outcomes is associated with genetic differences.

But the authors emphasise that genetic and environmental influences are intertwined.

For example, children's language environments are partly shaped by characteristics of their parents.

Parents transmit genes.

They also create environments.

Children's own genetically influenced characteristics can influence which activities they enjoy, how other people respond to them, and how much they engage with language and print.

This is known broadly as gene–environment correlation.

There are also direct environmental influences.

Instruction matters.

Language exposure matters.

The writing system being learned matters.

Educational opportunity matters.

The developmental outcome is produced through their interaction.

The new dyslexia definition therefore describes genetic, neurobiological and environmental influences as interacting throughout development, rather than assigning causation exclusively to one domain.

The brain matters; but there is no diagnostic “dyslexic brain”

Neuroscience has also profoundly influenced understanding of dyslexia.

Neuroimaging research has identified average differences between groups of people with and without reading difficulties in brain systems involved in language, visual word recognition, phonological processing and learning.

Contemporary research also implicates broader neural systems rather than a single reading centre.

The 2026 review by van Bergen and colleagues notes evidence involving not only cortical regions traditionally associated with language and reading, but also subcortical systems involved in more general learning processes.

This strengthens the biological understanding of reading development.

But it does not mean that neuroscientists can place an individual in a scanner and determine whether that person has dyslexia.

The authors of the 2026 IDA scientific rationale are explicit:

current neuroscience has not identified definitive brain-based biomarkers capable of diagnosing dyslexia or directing intervention at the individual level.

That distinction mirrors the genetics.

Researchers can identify average differences between groups.

Individual brains overlap considerably.

A statistically significant group difference is not automatically a clinically useful diagnostic marker.

Brain differences may be causes, consequences; or both

There is an additional complication.

Learning changes the brain.

Reading is itself an acquired skill requiring extensive experience.

This means that when researchers find differences between the brains of stronger and weaker readers, an important causal question arises:

Did that neural characteristic contribute to the reading difficulty?

Did reduced reading experience contribute to that neural characteristic?

Or did both processes influence one another over development?

This is why longitudinal and intervention studies are particularly important.

The 2026 review of developmental reading and language disorders highlights research attempting to distinguish relatively stable neural characteristics from features that change following learning or intervention.

This makes simplistic brain images especially dangerous.

A coloured region on a neuroscience figure may represent a group-level statistical difference.

It does not necessarily represent a defect.

It does not demonstrate causation.

And it does not reveal everything about an individual's reading ability.

Reading itself reshapes development

The relationship also works in the opposite direction.

Dyslexia does not only affect the act of decoding a word.

Persistent difficulty with word reading can alter the amount and type of reading experience a person receives.

A child who reads less encounters fewer written words.

Reduced reading experience can subsequently affect:

vocabulary growth,

background knowledge,

reading comprehension,

written expression,

and broader academic learning.

The revised IDA definition explicitly recognises these secondary consequences.

This produces a developmental feedback loop.

Initial reading difficulty can reduce reading experience.

Reduced experience can then widen later differences in language and knowledge.

The developmental outcome therefore cannot be understood by examining only what was present before a child learned to read.

Dyslexia is simultaneously influenced by development and capable of influencing subsequent development.

Different writing systems matter

Another significant change in the 2025 definition is explicit recognition of orthography.

Writing systems differ dramatically.

In some alphabetic languages, relationships between letters and sounds are highly consistent.

English is considerably less consistent.

Other languages use writing systems organised in very different ways.

The manifestation of dyslexia can therefore vary according to what a reader is being asked to learn.

A difficulty with reading accuracy may be particularly obvious in one writing system.

In another, a person may achieve reasonable accuracy but remain markedly slow.

The revised definition acknowledges this directly by stating that reading and spelling difficulties can involve accuracy, speed or both and vary depending on the orthography.

This is another challenge to the idea of one universal dyslexic phenotype.

The underlying vulnerability develops within a linguistic and educational context.

Dyslexia exists on a continuum

The new definition also describes dyslexia as existing along a continuum of severity.

This reflects a broader development in psychological and genetic science.

Many human characteristics do not separate naturally into two biological populations.

Reading ability is continuously distributed.

At one end are people who learn to read extremely readily.

At another are people with severe and persistent difficulties.

Diagnostic categories remain useful because they can identify meaningful impairment and access to support.

But the underlying trait does not necessarily contain a biological cliff separating everyone with dyslexia from everyone without it.

This idea is compatible with modern genetics.

Polygenic influences affecting reading ability are distributed throughout populations.

The same general biological systems contribute to ordinary variation as well as more severe difficulty.

Dyslexia can therefore be both a meaningful clinical and educational construct and part of continuous variation in human reading development.

Those positions are not contradictory.

What does not diagnose dyslexia?

The emerging research allows several widely repeated claims to be rejected with reasonable confidence.

Letter reversals do not define dyslexia.

A large discrepancy between IQ and reading attainment is not required.

There is no single dyslexia gene.

There is no currently validated diagnostic dyslexia brain scan.

A single phonological, cognitive, genetic or neurological marker cannot independently confirm or exclude dyslexia.

What remains central is much less technologically dramatic:

persistent difficulty in accurate and/or fluent word reading and spelling, interpreted within the individual's developmental, educational and linguistic context.

In an era fascinated by biomarkers and artificial intelligence, that may seem surprisingly conventional.

Scientifically, however, it matters.

A sophisticated biological explanation does not necessarily require a biological diagnostic test.

What the evidence currently supports

WHAT WE KNOW

Dyslexia involves persistent difficulties with word-level reading and/or spelling affecting accuracy, speed or both.

Reading ability has a substantial genetic component, but the genetic architecture is highly polygenic.

Large-scale genomic studies now identify many associated loci rather than one defining gene.

Phonological difficulties are strongly associated with dyslexia but are neither present in every person with dyslexia nor sufficient by themselves to determine diagnosis.

Brain differences associated with reading difficulty can be detected at the group level.

None of these genetic, cognitive or neural measures currently functions as an independent diagnostic biomarker.

WHAT THE RESEARCH SUGGESTS

Multiple developmental pathways can lead to persistent reading difficulty.

Different individuals may arrive at a similar reading outcome through different combinations of genetic liability, language development, cognitive characteristics, neural development and environmental experience.

Genes associated with reading appear to participate largely in broader developmental and neural processes rather than forming a specialised genetic programme for literacy.

Reading disorders may therefore be better understood through multifactorial developmental models than through attempts to discover one core deficit.

WHAT WE DO NOT YET KNOW

We cannot predict an individual's eventual reading development accurately from their genome.

We cannot diagnose dyslexia from a brain scan.

We do not yet know precisely how thousands of genetic variants interact with brain development, cognition, language exposure and education to produce individual reading trajectories.

Nor have researchers identified one cognitive or biological subtype structure capable of explaining every person with dyslexia.

The biological architecture is becoming clearer.

Individual prediction remains much harder.

What should be studied next?

The next phase of dyslexia research needs to follow development rather than simply compare groups after difficulties are already established.

Large longitudinal studies can help determine which biological and cognitive differences precede reading acquisition and which emerge partly as consequences of reading experience.

Genomic research also needs greater representation across ancestries and languages. Large genetic datasets remain disproportionately derived from people of European ancestry, limiting the extent to which findings can automatically be generalised globally.

Cross-linguistic research is equally important.

A scientific theory of dyslexia should explain reading difficulty across different writing systems rather than assuming that findings obtained primarily in English-speaking populations represent literacy universally.

Neuroscience needs stronger links with intervention research.

If a brain measure changes after effective reading instruction, that change can reveal something about learning mechanisms rather than simply cataloguing difference.

And perhaps most importantly, genetic and neuroscience research must continue to resist the temptation to turn population-level associations into individual destiny.

A more complex definition can lead to a simpler educational principle

There is a paradox at the centre of modern dyslexia research.

The biology is becoming more complicated.

The practical message may be becoming simpler.

If reading difficulty cannot be reliably separated according to IQ...

if there is no diagnostic gene...

if there is no diagnostic brain scan...

and if children can reach persistent reading difficulty through different developmental routes...

then waiting for a perfect biological explanation before providing support makes little sense.

The 2025 definition emphasises early identification and targeted instruction while also recognising that appropriate intervention remains important at any age.

Scientific complexity should therefore not become educational paralysis.

It should make support more responsive to the individual.

Conclusion: dyslexia has been redefined because certainty has become more sophisticated

For much of the public, dyslexia remains associated with a remarkably simple collection of ideas.

Letters appear backwards.

A bright child inexplicably struggles to read.

A particular part of the brain works differently.

Somewhere, perhaps, there is a dyslexia gene waiting to be discovered.

Contemporary science tells a different story.

Dyslexia is real.

Its biological foundations are substantial.

But those foundations are distributed across many genetic variants, multiple neural systems, cognitive and linguistic processes, environmental experiences and developmental stages.

There is no single gene.

There is no single brain signature.

There is no universal phonological profile.

There is no intelligence threshold.

And there is no single developmental route.

That complexity is not a weakness in the science.

It is evidence of its maturation.

The revised definition of dyslexia reflects a broader transformation occurring across developmental neuroscience: away from searching for one defect that explains everyone, and towards understanding how multiple biological and environmental influences combine differently across individuals and over time.

The central question is therefore changing.

It is no longer simply:

“What is wrong in the dyslexic brain?”

A more useful question is:

“How do different biological, cognitive, linguistic and environmental pathways shape the development of reading—and why does learning to read remain persistently difficult for some people?”

That question is harder.

But it may bring us much closer to understanding dyslexia as it actually exists.

The Society Research Position

WHAT WE KNOW

Dyslexia is characterised principally by persistent difficulty developing accurate and/or fluent word reading and spelling. Genetic and neurobiological influences are substantial, but neither genes nor brain differences determine outcome in isolation.

WHAT THE RESEARCH SUGGESTS

Multiple developmental pathways can produce dyslexic reading difficulties, with genetic, neural, cognitive, linguistic and environmental influences interacting across development.

WHAT WE DO NOT YET KNOW

There is no validated genetic test, brain scan or single cognitive marker capable of independently diagnosing dyslexia or predicting an individual's reading trajectory.

WHAT SHOULD BE STUDIED NEXT

Research needs larger longitudinal, cross-linguistic and ancestrally diverse studies capable of connecting genetics, brain development, learning environments and intervention across time.

References

1. Catts HW, Haynes CW, Joshi RM. Defining dyslexia: 2025 revision. Annals of Dyslexia. Published 29 April 2026. doi: 10.1007/s11881-026-00363-4.

2. International Dyslexia Association. 2025 Definition of Dyslexia. Adopted by the IDA Board of Directors, 22 October 2025.

3. Mountford HS, Eising E, Fontanillas P, et al. Multivariate genome-wide association analysis of dyslexia and quantitative reading skill improves gene discovery. Translational Psychiatry. 2025;15:289. doi: 10.1038/s41398-025-03514-0.

4. Dobrynin P, Zeng Y, Norkina M, Fedorova A, Zhuk A, Grigorenko EL. Four Decades of Inquiry Into the Genetic Bases of Specific Reading Disability. Journal of Speech, Language, and Hearing Research. 2025;68(11):5158–5173. doi: 10.1044/2025_JSLHR-25-00050.

5. van Bergen E, Dawson N, James E, Krishnan S. Progress and perspectives on developmental language and reading disorders. Quarterly Journal of Experimental Psychology. Published online 28 August 2026. doi: 10.1177/17470218261486347.

6. Elliott JG, Grigorenko EL. Dyslexia in the twenty-first century: a commentary on the IDA definition of dyslexia. Annals of Dyslexia. 2024;74:363–377. doi: 10.1007/s11881-024-00311-0.

The Society Journal publishes research-led analysis across neurodivergence, science, health, society, education and culture. Research Insights distinguish established evidence from emerging findings, unresolved questions and popular claims that have moved beyond the science.