10/7/2016

1

State of the Science of Dyslexia From Genes to Brain to Social Context

Robin L. Peterson, PhD, ABPP

Children’s Hospital Colorado

Lauren M. McGrath, PhD

Assistant Professor

University of Denver

Disclosure

We receive royalties for contributions to:

Diagnosing Learning Disorders (2nd Edition). Bruce F. Pennington (2008). Guilford.

Acknowledgments

• Bruce Pennington

• Colorado Learning Disabilities Research Center (NIH grant P50 HD27802)

• Richard Olson • Erik Willcutt • Jan Keenan • Richard Boada • Anne Arnett • Brian Byrne • Stefan Samuelsson

Overview

• “Levels of analysis” approach

• Definitions

• Review of what is known across levels of analysis, highlighting some exciting recent (and not-so-recent) findings

• Discussion

Levels of Analysis Approach

Etiology Brain bases

Neuropsychology

Behavior Social Context

DEFINITIONS & EPIDEMIOLOGY

10/7/2016

2

Definitions

“Dyslexia is a specific learning disability that is neurobiological in origin. It is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities. These difficulties typically result from a deficit in the phonological component of language that is often unexpected in relation to other cognitive abilities and the provision of effective classroom instruction. Secondary consequences may include problems in reading comprehension and reduced reading experience that can impede growth of vocabulary and background knowledge.” --Lyon, Shaywitz, & Shaywitz (2003)

Definitions • Individuals with Disabilities in Education Act (IDEA, 2004) The term “specific learning disability” means a disorder in

one or more of the basic psychological processes involved in understanding or in using language, spoken or written, which disorder may manifest itself in the imperfect ability to listen, think, speak, read, write, spell, or do mathematical calculations.

Identified by response to intervention (RTI) model and/or

patterns of strengths and weaknesses (PSW) model

Definitions

• DSM-5 – Difficulties learning and using at least one

academic skill, including: • Inaccurate or slow and effortful word reading (e.g.,

reads single words aloud incorrectly or slowly and hesitantly, frequently guesses words, has difficulty sounding out words).

• Note: Dyslexia is an alternative term used to refer to a pattern of learning difficulties characterized by problems with accurate or fluent word recognition, poor decoding, and poor spelling abilities.

Lyon, Shaywitz, & Shaywitz

IDEA DSM-5

How severe? ? “Substantial impact”

“Significant interference”

How specific? IQ discrepant?

“Often unexpected in relation to other cognitive abilities”

Different for RTI vs. PSW model

IQ discrepancy not considered

Etiology

Not due solely to ineffective instruction

Not due to environmental disadvantage,

sensory deficits, or intellectual

disability (ID)

Not due to inadequate

instruction, sensory deficits, ID, second

language instruction, or psychosocial

adversity

Neuropsychology Phonologically based

Language based ?

How severe?

• Some disorders are categorical (like cystic fibrosis)

• Some disorders are continuous

• Dyslexia, like all behaviorally-defined disorders, is defined by a cut-point in a continuous distribution

2%

7%

10%

16%

25%

10/7/2016

3

How specific?

• Historical focus on extreme discrepancies

British Medical Journal (1896)

But what about kids who have trouble with reading and

math?

Why IQ discrepancies were abandoned

• Children meeting IQ-discrepancy criteria and age-discrepancy criteria show similar deficits in phonological processing, and

• Tend to respond to the same intensive, phonics-based instruction

What about “patterns of strengths and weaknesses”?

• Problems with using this approach to identify learning disabilities

– How exactly do we define PSW? Must put cut-points on multiple different measures

– Most children with reading difficulties do not meet PSW criteria

– A slight change in criteria results in large differences in which children are identified

Comorbidity: Another problem for specificity

Approximate comorbidity rates with dyslexia:

• Math disability: ~50%

• Language Disorder: ~50%

• ADHD: ~25%

Also comorbid with anxiety, depression, speech sound disorder, developmental coordination disorder. . .

10/7/2016

4

But surely there must be something specific about a “specific learning disability”!

• We are currently looking at this question in the Colorado Learning Disabilities Research Center

• Analyzed data from 822 children aged 8 to 18 who had completed multiple measures of reading, math, writing, and intelligence

• Sample is slightly overselected for dyslexia

Basic Reading

Reading Comp

Math

PIAT Read

PIAT Spell

TLWRT

TOWRE

GORT Fluency

WRAT Spell

QRI Read Quest

WJ Pass Comp

GORT Comp

Handwrite copy

WJ Write Flu

WJ Write Samp

Wechsler Arith

WRAT Math

PIAT Math

.85**

.72**

.87**

.78**

.81**

.75**

.50**

.81**

.55**

.52**

.71**

.52**

.63**

.68**

.72**

.33**

.31**

.39**

.18**

.37**

.43**

.40**

.26**

.47**

.47**

.37**

.43**

Peterson, McGrath, Willcutt, Keenan, Olson, & Pennington (under review)

PIAT Comp .21** .79**

Academic g

Model fit: Χ2(91) = 471.79, p<.001, CFI=.95, RMSEA=.071

Basic Reading

Reading Comp

Math

PIAT Read

PIAT Spell

TLWRT

TOWRE

GORT Fluency

WRAT Spell

QRI Read Quest

WJ Pass Comp

GORT Comp

Handwrite copy

WJ Write Flu

WJ Write Samp

Wechsler Arith

WRAT Math

PIAT Math

PIAT Comp

Academic g

Verbal Comp

Percept Org

Working Memory

Information

Similarities

Vocabulary

Comprehension

Picture Compl

Picture Arrange

Object Assembl

Sentence Span

Counting Span

Digits Backward

Symbol Search

Identical Pict

CO Percept Sp

Block Design

Cognitive g

Process Speed

Coding

r=1.0

Model fit: Χ2(406) = 1781.98, p<.001, CFI=.90, RMSEA=.064

So, yes, there are specific learning disabilities, but . . .

Epidemiology: Gender Differences

• Slightly more boys than girls with dyslexia in population, but many more boys than girls referred

Epidemiology: Cross-Cultural Findings

• Manifests across languages and cultures in largely similar ways

Paulesu et al. (2001). Science

A: Typical Readers B: Dyslexic Readers C: Group differences

10/7/2016

5

ETIOLOGY

Single Gene vs. Complex Disorders

• Many genes contribute

• Each has a small effect by itself

• Effects may depend on interaction with environment and with other genes

• Examples: Fragile X syndrome

Huntington’s Disease

• Single gene causes disease

Courtesy J. W. Smoller

Twin studies - Terms

• Heritability – the percent of observed differences between individuals that are attributable to genetic differences

• Shared environment – all non-genetic influences that make family members similar to each other

• Unique environment – all non-genetic influences that make family members different from each other (includes measurement error)

Plomin, 2013

Heritability

Unique environment

Shared Environment

Twin studies – a natural experiment

~100% genetically similar

100% of shared env is similar

0% of unique env is similar

~50% genetically similar

100% of shared env is similar

0% of unique env is similar

• Give all twins a reading test

• Assess how alike MZ twins are compared to DZ twins

• Quantify differences between MZ and DZ pairs = heritability estimate

Twin studies – a natural experiment

~100% genetically similar

100% of shared env is similar

0% of unique env is similar

~50% genetically similar

100% of shared env is similar

0% of unique env is similar

Isolate

genetics as

a factor

68% concordant 38% concordant

DeFries, 1987

• Example from a twin study of dyslexia

What does it mean if:

MZ > DZ

MZ=DZ

MZ < 100%

Key assumption:

Equal environments

(Kendler, 1993)

Twin study results in dyslexia

• Heritability = 50-60%

• Shared environment = 0-20%

• Unique environment = 5-20%

Willcutt et al., 2010

10/7/2016

6

Heritability does not tell you:

• About individual people

• The heritability estimate is a group average

• You CANNOT say that 50% of cases are genetic

• You CANNOT say that this child’s reading is 50% genetic

• You CANNOT say that there’s a 50% chance that this child

inherited the trait from their parents

• About individual genes

• You only know that genes contribute but not which genes, how

many genes, or what combination of genes

Heritability does not tell you:

• About the future

• The heritability estimate describes WHAT IS not WHAT WILL BE

• It does not tell you ANYTHING about whether this child will respond

well to intervention

• About group differences (gender, social class, ethnicity)

• A twin study might determine that heritability is 50% for math ability

• This DOES NOT mean that 50% of the differences between men

and women are genetic

• ALL of the group difference could be environmental, even though

50% of the differences between individuals are due to genetic

differences

Disorder/trait Approx. h2

Autism 70-90%

Schizophrenia 80%

Bipolar Disorder 80%

Attention Deficit Hyperactivity Disorder 77%

Reading Disability 50-65%

IQ 50-80%

Mathematics Disability 40-50%

Language Impairment 40-50%

Personality traits 30-50%

Major Depression 42%

Anxiety Disorders 30-45%

Note that not one of these disorders is 100% heritable

GENE-FINDING

10/7/2016

7

Emerging Stories

• DCDC2

• KIAA0319

• DYX1C1

Remember the context… • In complex disorders, genes are not causal

• Early identification with these genetic effects is not possible

• Here’s why:

• This is an example assuming that RD has a similar genetic

architecture to other well-studied behavioral disorders

• Dyslexia prevalence = 5%

• Risk allele prevalence = 20%

• Strength of effect = odds ratio of 1.2

Dyslexia No Dyslexia

Risk allele 4 60 6%

No risk allele 13 240 5%

Take-home

• There is no genetic test available for dyslexia.

• There isn’t likely to be one in the future because we are

talking about hundreds to potentially thousands of genes

all combining with each other and with the environment

• Genetic studies can still inform us about the underlying

neurobiology of dyslexia…

Generalist Genes

Kovas & Plomin, 2006

BRAIN BASIS

Brain Bases

From Peterson & Pennington (2015) Annual Review of Clinical Psychology

10/7/2016

8

• Many brain differences are evident before literacy instruction begins

• Brain response to speech sounds at 2 months of age differentiates children with good and poor reading fluency in 2nd grade!

Van Zuijen et al (2013) Developmental Science

Interesting new directions

• Using brain correlates to predict response to treatment

Hoeft et al., 2011

Working out the cause vs. consequences of reading difficulties

using neuroimaging

Olulade et al., 2013

NEUROPSYCHOLOGY

Neuropsychology

A “beautiful hypothesis” (Scarborough, 2005)

Ugly Fact #1

• Best predictors of later reading change with development

– Syntax and speech production (2.5-3 years)

– Syntax and vocabulary (3.5 -4 years)

– Vocabulary and phonological awareness (5 years)

10/7/2016

9

Ugly Fact #2

• Learning to read changes how the brain processes phonological information

– “Naturally illiterate” adults do not have explicit awareness of individual phonemes until after they are taught to learn to read

Peterson, Arnett, Pennington, Byrne, Samuelsson, & Olson (under revision).

A reinterpretation Ugly Fact #3

• Some children with phoneme awareness deficits don’t develop dyslexia

• Peterson, Pennington, Shriberg, & Boada (2009)

– Literacy development in children with speech sound disorder with or without comorbid language impairment

Ugly Fact #4

• As a group, individuals with dyslexia have difficulties in multiple cognitive skills

– Rapid naming

– Processing speed

– Verbal working memory

– Vocabulary

– Verbal reasoning

Moving from single to multiple deficit models (Pennington, 2006)

10/7/2016

10

Moving from single to multiple deficit models (Pennington, 2006)

PA PS Vocab RAN VWM

Moving from single to multiple deficit models (Pennington, 2006)

PA PS Vocab RAN VWM

Multiple deficit models

Phonological

Processing

Inhibition

Processing

Speed

Reading

Attention

Verbal WM

Naming

Speed

McGrath et al., 2011

Cognition Trait/Disorder

.11

.59

.25

Multiple deficit models

Phonological

Processing

Inhibition

Processing

Speed

Reading

Attention

Verbal WM

Naming

Speed

McGrath et al., 2011

Cognition Trait/Disorder

.37

.36

Multiple deficit models

Phonological

Processing

Inhibition

Processing

Speed

Reading

Attention

Verbal WM

Naming

Speed

McGrath et al., 2011

Cognition Trait/Disorder

.37

.36

.11

.59

.25 CFI = 0.955

RMSEA = 0.048

RMSEA 90% CI = 0.043 - 0.052

Chi-square/df = 2.385

Coding

CO perceptual speed

Identical pictures test

75%

35%

Multiple deficit models

Phonological

Processing

Inhibition

Processing

Speed

Reading

Attention

Verbal WM

Naming

Speed

Cognition Trait/Disorder

Full model:

CFI = 0.95

RMSEA = 0.05

Chi-square/df = 2.61

Math Verbal

Comp

Peterson, Boada, McGrath et al. (2015)

10/7/2016

11

Ugly Fact #5 • Great individual variability Pennington et al. (2012), Journal of Abnormal Psychology

About ¼ of cases best explained by a single deficit in PA, language, OR processing speed About ¼ of cases best explained by multiple deficits Nearly ¼ of cases don’t have any pronounced deficits Remaining cases show mixed evidence across different approaches

Summary of Neuropsychology

• Yes, phonological deficits are strongly linked to dyslexia, BUT – We cannot use a PA deficit (or any other specific cognitive

deficit) to rule dyslexia in or out in any particular case

– Many children with phonologically-based reading difficulties may also have more generalized learning weaknesses

– Cognitive skills that are general risk factors across developmental disorders may contribute to comorbidity

SOCIAL CONTEXT

Etiology

• <100% heritability means the environment also matters! But which environments?

-Tricky to study

• Families vary in many aspects of home language and literacy environment

• Of course also differences in classrooms, schools, curricula, neighborhoods. . .

• Not “nature or nurture”

• Not even “nature and nurture”

• A complex developmental process

– “gene-environment interplay:

• For example, preschool children at genetic risk for dyslexia avoid being read to and spend less independent play time looking at books

10/7/2016

12

INTEGRATING LEVELS OF ANALYSIS

From genes to brain to behavior

• Genetic neuroimaging

Imaging Genetics

Gene

Brain

Behavior

Darki, Peyrard-Janvid, Matsson, Kere, & Klingberg (2012). Biological Psychiatry

INTERVENTION

Research in people with dyslexia shows that effective intervention changes the brain

Gabrieli, J. D. (2009). Science.

And this is the case even into adulthood

Eden et al., 2004

10/7/2016

13

Summary

• Genetics: There is no genetic test for dyslexia (and there isn’t likely to be one in the future)

• Neuroimaging: making an “invisible” problem visible (Shaywitz) – Trying to understand cause vs. consequence of reading difficulties because learning to read

changes the brain

– Trying to use imaging and behavior to predict response to treatment (early days)

• Neuropsychology: PA plays an important role in dyslexia, but multiple deficit models are helping to understand the full range of presentations

• Social context: Gene-environment interplay magnifies the effects of each factor

• Intervention: Reading interventions DO change the brain – We already know this must be true if behavior changes, but sometimes the brain pictures can

be powerful motivators

www.shutterstock.com www.crazytownblog.com dglopa.com www.pbs.org faculty.washington.edu www.mondopub.com www.goredforwomen www.chronicle.bg www.dreamstime.com www.healthline.com www.twin-pregnancy-and-beyond.com www.whatsupfagans.com blog.kidtelligent.com bluepuzzle.org/iceberg personal.kenyon.edu joyfultoddlers.com www.youandmeeducation.com www.huffingtonpost.com www.sciencedaily.com eastbrooklyn.com slideplayer.com youtube.com

Image credits