Dyslexia Beyond the Word: An Ecological Study of Specific ...

Brigham Young University Brigham Young University

BYU ScholarsArchive BYU ScholarsArchive

Theses and Dissertations

2020-06-01

Dyslexia Beyond the Word: An Ecological Study of Specific Dyslexia Beyond the Word: An Ecological Study of Specific

Reading Disorder Reading Disorder

Benjamin T. Carter Brigham Young University

Follow this and additional works at: https://scholarsarchive.byu.edu/etd

Part of the Family, Life Course, and Society Commons

BYU ScholarsArchive Citation BYU ScholarsArchive Citation Carter, Benjamin T., "Dyslexia Beyond the Word: An Ecological Study of Specific Reading Disorder" (2020). Theses and Dissertations. 9207. https://scholarsarchive.byu.edu/etd/9207

This Dissertation is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please contact [email protected].

http://home.byu.edu/home/

http://home.byu.edu/home/

https://scholarsarchive.byu.edu/

https://scholarsarchive.byu.edu/etd

https://scholarsarchive.byu.edu/etd?utm_source=scholarsarchive.byu.edu%2Fetd%2F9207&utm_medium=PDF&utm_campaign=PDFCoverPages

http://network.bepress.com/hgg/discipline/419?utm_source=scholarsarchive.byu.edu%2Fetd%2F9207&utm_medium=PDF&utm_campaign=PDFCoverPages

https://scholarsarchive.byu.edu/etd/9207?utm_source=scholarsarchive.byu.edu%2Fetd%2F9207&utm_medium=PDF&utm_campaign=PDFCoverPages

mailto:[email protected]

Dyslexia Beyond the Word: An Ecological Study of Specific Reading Disorder

Benjamin T. Carter

A dissertation submitted to the faculty ofBrigham Young University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Steven G. Luke, ChairC. Brock KirwanMichael BrownDerin CobiaJared Nielsen

Department of Psychology

Brigham Young University

Copyright © 2020 Benjamin T. Carter

All Rights Reserved

ABSTRACT


Benjamin T. CarterDepartment of Psychology, BYU

Doctor of Philosophy

This dissertation discusses the effects of dyslexia on reading behavior and cognition. Itdoes so by first outlining the overall incidence of dyslexia, providing current definitions,giving a history of scientific inquiry and discussing relevant contemporary research.Thirteen different analyses are then discussed (ten a priori and three post-hoc). Individualswith dyslexia were found to have increased fixation duration, first run dwell time, totaldwell time, and refixation probability. The dyslexia group was also highly sensitive tolexical predictability. Within the reading network, the BOLD response was depressed indyslexia during reading in the following regions: the left medial and inferior temporalgyrus, the left temporal pole, the right cerebellum, right occipital gyrus and the rightparahippocampal gyrus. A second regions of interest analysis in the reading networkrevealed dyslexia was associated with a depressed BOLD response to lexical predictabilityin the following regions: left supplementary motor area, posterior middle frontal gyrus, andthe left temporal pole. A regions of interest analysis in the oculomotor network revealed adepressed BOLD response in the following regions during reading: the left parietal eyefields and the cerebellum. One oculomotor region had a depressed BOLD response tolexical predictability due to dyslexia: the left frontal eye fields. This sensitivity to lexicalpredictability and depression in the BOLD response is suggestive of reduced input intohigher cortical areas. Future study should be focused on finding the common origin of thisbottom-up deficit.

Keywords: dyslexia, reading, predictability, paragraph

iii

TABLE OF CONTENTS

ABSTRACT ii

Dyslexia Beyond the Word: An Ecological Study of Specific Reading Disorder 1

Incidence & Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Early Theory 1

Modern Theory 3

The Phonological-Deficit Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . 4

The Multifactorial Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

The Neural Noise Hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Documented Deficits 7

Oculomotor Attention Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Linguistic Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Neurological Deficits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Proposal 12

Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Hypotheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Novelty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Methods 14

Participant Recruitment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

iv

Participant Characteristics 19

Hypothesis 1 - Confirming Previous Findings 20

Analysis 1 - Oculomotor Differences . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Analysis 2 - Hemodynamic Differences . . . . . . . . . . . . . . . . . . . . . . . . 21

Analysis 3 - ROI Analysis of the Reading Network . . . . . . . . . . . . . . . . . . 22

Analysis 4 - ROI Analysis of the Eye Fields . . . . . . . . . . . . . . . . . . . . . 22

Analysis 5 - ROI Analysis of the Caudate and Putamen . . . . . . . . . . . . . . . 23

Hypothesis 2 - Exploring Linguistic Predictability 23

Analysis 1: Predictability Effects on Oculomotor Behavior . . . . . . . . . . . . . 23

Analysis 2 - Group Differences in BOLD Response to Predictability . . . . . . . . 24

Analysis 3 - ROI Analysis of the Reading Network . . . . . . . . . . . . . . . . . . 24

Analysis 4 - ROI Analysis of the Eye Fields . . . . . . . . . . . . . . . . . . . . . 25

Analysis 5 - ROI Analysis of the Caudate and Putamen . . . . . . . . . . . . . . . 25

Post-Hoc Analyses 26

Factor Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Analysis 1 - a Second Look at the Reading Network . . . . . . . . . . . . . . . . . 27

Analysis 2 - a Second Look at the Eye Fields . . . . . . . . . . . . . . . . . . . . . 27

Analysis 3 - a Second Look at the Striatum . . . . . . . . . . . . . . . . . . . . . 27

Discussion 28

Oculomotor Behavior in Dyslexia . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

BOLD Response in Regions of Interest . . . . . . . . . . . . . . . . . . . . . . . . 31

Poor Input or Output? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Next Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Conclusion 39

v

References 47

vi

LIST OF TABLES

1 Participant demographics . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2 Cognitive test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

3 Student’s T-test results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4 Oculomotor profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

5 Significance of group differences . . . . . . . . . . . . . . . . . . . . . . . . . 41

6 First fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

7 First run dwell time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

8 Total dwell time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

9 Intercorrelations of cognitive subtest scores . . . . . . . . . . . . . . . . . . . 44

10 Factor loadings from Principle Components Analysis . . . . . . . . . . . . . 45

11 Changes in group assignment from factor analysis . . . . . . . . . . . . . . . 46

vii

LIST OF FIGURES

1 Regions of Interest and Gray Matter Mask . . . . . . . . . . . . . . . . . . . 40

2 Distribution of cognitive test scores . . . . . . . . . . . . . . . . . . . . . . . 42

3 Whole brain response to paragraphs by group. . . . . . . . . . . . . . . . . . 43

4 The effect of group and lexical predictability on first fixation duration . . . . 44

5 The effect of group and lexical predictability on first run dwell time . . . . . 45

6 The effect of group and lexical predictability on total dwell time . . . . . . . 46

ECOLOGICAL DYSLEXIA 1


Incidence & Definition

Reading is a critical skill necessary for survival in a world dominated by information

technology (Snow, Burns, & Griffin, 1998). Without it educational opportunities and

income potential are severely limited. Early reading achievement is positively associated

with educational status (Lee, Daniels, Puig, Newgent, & Nam, 2008), employment (Caspi,

Wright, Moffitt, & Silva, 1998) and wages earned (Arnbak, 2004; McLaughlin, Speirs, &

Shenassa, 2014). One obstacle faced by many students learning to read is dyslexia.

S. E. Shaywitz, Escobar, Shaywitz, Fletcher, and Makuch (1992) estimated dyslexia’s

prevalence to be 9% of boys and 6.0% of girls in the third grade while others estimate it

may have an incidence as high as 10% in the general populace (Siegel, 2006). Discrepancies

between estimates are probably heavily influenced by the definition of dyslexia used in each

study (Rodgers, 1983; B. A. Shaywitz, Fletcher, Holahan, & Shaywitz, 1992; Siegel, 2006).

The current definition of dyslexia from Lyon, Shaywitz, and Shaywitz (2003) is as

follows:

Dyslexia is a specific learning disability that is neurological 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 the growth of

vocabulary and background knowledge.

Early Theory

The definition of dyslexia has evolved over time. Dyslexia’s discovery was

concurrent with a rise in general literacy during the end of the 19th century. The German


ophthalmologist, Dr. Rudolf Berlin, first coined the term dyslexia in his book Eine

besondere art der wortblindheit (dyslexie), published in 1887. In it he described six peculiar

cases in which individuals struggled to read yet were of normal intelligence and without

ocular abnormalities. Berlin believed this condition was neurological in origin,

hypothesizing it was due to a depletion in the conductivity of the connecting fibers of the

lower parietal lobe. In support of his hypothesis Berlin reported lesions of the

supramarginal and angular gyri were discovered during postmortem examinations of

similarly afflicted individuals (Berlin, 1887). Shortly thereafter cases of dyslexia began to

appear in the medical literature. Multiple etiologies were proposed. For example,

Hinshelwood (1896) reported a case in The Lancet in which an alcoholic was cured of toxic

dyslexia after abstaining for a period of seven weeks and remained cured for the rest of his

abstinent life. Cases of congenital "word blindness" also began to appear. Morgan (1896)

observed a 14 year old boy who struggled to read and spell despite normal intelligence.

Other cases were recognized and it became apparent that dyslexia was more complicated

than previously thought.

Eventually acquired dyslexia (alexia) and developmental dyslexia diverged as two

separate phenomena, the former being the result of pathology and the latter being present

from a young age and without a medical explanation. Samuel T. Orton (1925) first made

this distinction after meeting a young girl who could not read and had symptoms similar to

stroke victims without a prior medical history of neurological pathology. He later

developed a theory of "strephosymbolia" (twisted symbols) describing individuals who

struggled to associate word phonology (sound) with orthography (symbology). Once this

distinction was made, researchers began to propose physiological causes. Orton proposed

dyslexia was a result of incomplete hemispheric dominance (1928). Other theories included

delayed myelination of the cerebral cortex (Sinclair, 1948) and differences in the

oculomotor centers (Jossmann, 1948).


Modern Theory

Clinically, dyslexia has been redefined as a specific learning disorder with a basis in

genetic, epigenetic and environmental factors and is characterized at the cognitive level by

difficulty with accurate or fluent word recognition, decoding, and poor spelling abilities.

The impairment must be present for at least 6 months, with academic performance

significantly below what would be expected according to intelligence and chronological age

and cannot be better accounted for by sensory deficit, disability or other disorder. Onset

typically occurs during formal schooling but may not be fully manifest until demands

exceed capacity (American Psychiatric Association, 2013). The ICD-10 refers to dyslexia

as specific reading disorder and gives it essentially the same definition as the DSM-V

(World Health Organization, 1992).

Some quantitative guidelines have been suggested. The ICD-10 recommends that

reading accuracy and/or comprehension must be at least 2 standard errors below what

would be expected based on age and intelligence or a history of reading difficulties with a

spelling test at least 2 standard errors below what would be expected. A minimum IQ of

70 on a standardized test is also recommended (see World Health Organization, 1993, pp.

144-145). The DSM-V gives similar recommendations (a minimun IQ of 70 and a score on

a standardized reading assessment that is 1.5 standard deviations below the population

mean for age). It also recommends that "[o]n the basis of clinical judgment, a more lenient

threshold may be used... when learning difficulties are supported by converging evidence

from clinical assessment, academic history, school reports, or test scores" (American

Psychiatric Association, 2013). Practically speaking, developmental dyslexia is diagnosed

by this standard in order to capture individuals with less severe degrees of reading

disability but who might otherwise go unnoticed. This enables clinicians to diagnose more

subtle cases of dyslexia that would only become apparent when the demands of the reading

task exceed capacity.

Research has primarily focused on psychological theories of reading to discern


possible mechanisms of dyslexia. Reading is thought to rely on two methods for word

identification and lexical access: direct access and phonologically mediated access. Direct

access occurs when a word is visually identified via the printed form without having to

convert the word into a phonological code to access meaning. Phonologically mediated

access occurs when a word is seen and meaning is accessed first by decoding the sound of

the word and then using this phonological recreation to access word meaning.

Theoretically, sensory or processing deficits impacting either method could result in

reading impairment. Both have been considered as causal candidates for dyslexia.

The Phonological-Deficit Hypothesis

Reading acquisition is highly reliant on phonological or phonemic awareness.

Phonological awareness can be described as an individual’s "knowledge of the internal

sound structure of spoken words" (Rayner, Foorman, & Perfetti, 2001, p. 37). Reading also

relies on other abilities including the ability to see and visually distinguish between

orthographic representations and overall intelligence.

Dyslexia is prominently thought to be based in a deficit in phonological processing,

termed the Phonological-Deficit Hypothesis (Castles & Friedmann, 2014). Dyslexic children

have been shown to have deficits in phoneme based tasks (Harm & Seidenberg, 1999;

Rayner et al., 2001). Phonemes are the simplest units making up words within a language.

Phoneme processing deficits impair the mapping of speech sounds to semantic meaning.

Impaired semantic representations then affect orthographic mapping, which results in

reading difficulty. This theory is not without its flaws. The base assumptions of the

hypothesis are that phonological deficits are the sole cause of dyslexia and therefore all

dyslexic children have a phonological deficit (Castles & Friedmann, 2014). This is not the

case. Not all children with dyslexia have been found to have a phonological deficit (White

et al., 2006). A number of different types of dyslexia have been suggested (Castles &

Friedmann, 2014).


The Multifactorial Hypothesis

Reading is a highly complex behavior, relying on the cooperation of multiple

systems within the brain for healthy function. There may be multiple points of failure

which could result in reading difficulty (Castles & Friedmann, 2014; Rayner et al., 2001).

This has resulted in a multifaceted approach in which phonological deficits are not viewed

as the sole cause of dyslexia, with some cases resulting from a failure of sensory integration.

For example, direct access is also impaired in some cases, as manifested by deficits in rapid

automatized naming (RAN) tasks. RAN tasks assess the ability to quickly name aloud

familiar visual stimuli such as shapes or single letters and is believed to be testing the

integrity of visual-verbal associations (Norton, Beach, & Gabrieli, 2015). RAN deficits were

initially proposed as both a core cause and an augmenting factor as part of the double

deficit hypothesis proposed by Wolf and Bowers (1999). Recent investigations have shown

deficits in RAN to more likely be moderating rather than causal factors. Catts, McIlraith,

Bridges, and Nielsen (2016) has proposed that there is no singular deficit that will yield a

dyslexic phenotype, but rather a combination of deficits (many of which are discussed

later).

Many subtypes of dyslexia have been proposed including surface dyslexia,

letter-position dyslexia, and attentional dyslexia (Castles & Friedmann, 2014). There is

debate as to whether these subtypes represent distinct forms of dyslexia or whether they

are simply the result of interacting risk factors with phonological deficits as the primary

cause of impairment (Castles & Friedmann, 2014; Catts et al., 2016). Testing for these

proposed subtypes has not been standardized and they are not utilized in clinical practice.

The Neural Noise Hypothesis

A third potential cause of dyslexia lies with changes in cortical circuitry and

neuronal migration. Gross neural activity is not an all or nothing response but is actually

quite chaotic and "noisy." It is a conglomeration of the activity of many neurons. Much of


this activity is random, secondary to aberrant neurotransmitter release, reduced

neurotransmitter clearance, and random fluctuations of ion concentrations across neuron

membranes. As a result any signal of significance must be both "loud" enough to be

detected and short enough to reduce its potential contribution to overall noise within the

system. This is accomplished via local excitation-inhibition circuits in which significant

signals are amplified and then cut off through a feedback-inhibition loop. Errors in

signaling can be a result of dysfunctional excitation-inhibition. Slowed inhibitory signals

result in increased signal duration and increased noise within a system. Early inhibition

results in reduced signal strength. This results in faulty signal transmission and

computation.

Errors in the timing of the excitation-inhibition feedback loop could explain some of

the features of dyslexia (Hancock, Pugh, & Hoeft, 2017). Deletions of the Dcdc2 gene were

found in 19% of individuals with developmental dyslexia in a study by Wilcke et al. (2009).

Dcdc2 deletions in the rodent model have been shown to have increased glutamate

signaling and expression of the NMDA receptor, NMDA excitability, increased spontaneous

activity and spike timing variability (Che, Girgenti, & LoTurco, 2014; Che, Truong, Fitch,

& LoTurco, 2016). These models have also demonstrated impaired rapid auditory

processing and memory (Centanni et al., 2016; Truong et al., 2014). These increases in

glutamatergic activity would result in increased baseline neural activity, resulting in

impairment in auditory processing. Magnetic resonance spectroscopy has likewise

substantiated this possibility. Increased concentrations of glutamate have been observed in

dyslexics in the occipital visual areas and are predictive of reading skill (Pugh et al., 2014).

Left lateralized gene expression for glutamatergic signaling in the superior temporal gyrus

and auditory cortex have also been observed (Karlebach & Francks, 2015). Therefore

excess glutamate activity may be a potential explanation.

Errors of neuronal migration can also give rise to increased neural noise via altered

circuit structure and atypical synaptic development. Both Dcdc2 and Kiaa0319 deletions


have been shown to alter the neural migration process in the animal model. Dcdc2 has

been linked to reduced dendrite growth and abnormal migration (Meng et al., 2005).

Kiaa0319, also linked to dyslexia (Paracchini D Phil et al., 2008), is associated with

GABAergic interneuron and pyramidal neuron migration (Szalkowski et al., 2013). Normal

cortical circuitry is unlikely in the face of these mutations in humans and would result in

changes in the timing of the excitation-inhibition cycle, resulting in neural noise that could

interfere with normal processing.

Documented Deficits

Oculomotor Attention Deficits

Eye movements during reading are a series of "stop and go" movements (Erdmann &

Dodge, 1898). Eye movements fall into two different categories – saccades and fixations.

Saccades are the ballistic movement of the eye from one visual target to the next.

Fixations are periods when the eye is not moving and is training the fovea on a visual

target. Visual information is gathered and processing is initiated during a fixation. The

next saccade is also planned. The temporal duration of a fixation is directly related to the

cognitive effort and attention required to process the target. Eye movements during

reading are composed of a series of fixations and saccades as the eye is directed from one

word to the next. Words such as "the" or "a" are often skipped. Words that are longer,

unfamiliar, or linguistically unpredictable receive longer fixations.

Eye movements were among the early features studied in dyslexia (Jossmann, 1948).

Eye tracking studies have been used to make inferences concerning cognitive and linguistic

processes in healthy individuals and have thus been applied to dyslexia. Early studies

demonstrated saccadic irregularity and frequent reversal errors compared to healthy

readers (Jossmann, 1948). More recent studies have confirmed and extended these findings.

Individuals with dyslexia skip fewer words, make multiple fixations on more words, regress

more frequently and fixate longer (Biscaldi, Gezeck, & Stuhr, 1998; Hawelka, Gagl, &


Wimmer, 2010; Hutzler & Wimmer, 2004; McConkie et al., 1991). There is also a marked

effect of word length (De Luca, Borrelli, Judica, Spinelli, & Zoccolotti, 2002; Hawelka

et al., 2010). Visuo-attentional deficits have also been identified in dyslexia and this may

be contributing to reading deficits by impairing orthographic and phonological processing

(Bellocchi, Muneaux, Bastien-Toniazzo, & Ducrot, 2013; Facoetti, Corradi, Ruffino, Gori,

& Zorzi, 2010). Word predictability has also been shown to have a potential effect.

Hawelka, Schuster, Gagl, and Hutzler (2015) found individuals with dyslexia had decreased

reading speed when word predictability was reduced, indicating impairment of linguistic

prediction. While informative, all of these studies utilized stimuli composed of single words

or isolated sentences to study eye movements and therefore have reduced ecological

validity. None have observed how an individual with dyslexia reads a paragraph of text –

the most common reading environment. More work must be done to define how these

behaviors extend to reading a page of text and the effects of text characteristics on reading

behavior and comprehension in dyslexia. Such information would allow for the creation of

guidelines to facilitate the creation of an orthographic environment to accommodate

individuals with dyslexia. It is also debatable as to whether dyslexia is purely a linguistic

deficit (De Luca, Di Pace, Judica, Spinelli, & Zoccolotti, 1999; Lukasova, Silva, & Macedo,

2016). Individuals with dyslexia appear to have impaired oculomotor performance even in

non-reading tasks. Lukasova et al. (2016) demonstrated increased latency in a saccadic

prediction task, which is not a language based task. Prediction deficits therefore may not

be isolated to language alone and have a basis within the executive control of oculomotor

centers. This may negatively impact learning to read and reading fluency.

The effect of an ecological reading environment on oculomotor attention and

behavior in dyslexia is not well studied. While earlier studies such as Jossmann (1948) did

utilize page-based reading measures, few since have utilized this method. Using the Web of

Science Core Collection (Reuters, 2012) and the terms "dyslexia" and "eye tracking" only

returns 56 studies (as of March 5, 2018). Refining these results with the terms "page,"


"paragraph," "connected text," or "continuous text" reveals zero, three, zero and one result,

only one of which directly addresses dyslexia in a paragraph based reading environment

(Rello & Baeza-Yates, 2017). This particular study focuses exclusively on the effect of

orthographic and format-related features of text on comprehension in dyslexia rather than

linguistic factors.

Linguistic Deficits

A hallmark of dyslexia is difficulty learning new words and grammar rules (Lyon

et al., 2003). Language learning heavily relies on statistical and procedural learning skill

(Christiansen & Chater, 2016; Dell & Chang, 2014; Krishnan, Watkins, & Bishop, 2016) so

procedural learning has been studied to determine its role in dylexia. Ullman and Pierpont

(2005) first proposed the involvement of the procedural learning system as a potential

explanation for language learning deficits. Serial reaction time (SRT) protocols have been

used to test procedural learning abilities in children with dyslexia in a number of studies.

A recent meta-analysis of these studies by Lum, Ullman, and Conti-Ramsden (2013) found

dyslexia had a significant effect (p < 0.001) on task performance, demonstrating that

individuals with dyslexia were noticeably impaired in the SRT paradigm, and thus had

impaired procedural learning. Statistical and grammar learning are also impaired.

Pavlidou, Kelly, and Williams (2010) found that children with dyslexia performed similarly

to age-matched controls in a memorization task, but struggled in an implicit artificial

grammar learning task, which indicates impaired rule abstraction. Gabay, Thiessen, and

Holt (2015) found that adults with dyslexia were impaired in a statistical learning task in

which they were asked to differentiate between two strings of syllables and determine which

string best matched a preceding third string. Performance was found to correlate with

reading ability. The impairment persisted even when non-linguistic stimuli were used,

demonstrating that it was not limited to language learning. Gabay has also demonstrated

that individuals with dyslexia are impaired in other probabilistic learning tasks such as


weather prediction (Gabay, Vakil, Schiff, & Holt, 2015) and auditory category learning

(Gabay & Holt, 2015). Sigurdardottir et al. (2017) likewise demonstrated statistical

learning impairment via the use of a shape memorization and recognition task.

Impairments of statistical and grammatical learning would all impact the ability to predict

features of upcoming text. Impaired statistical learning would result in difficulty making

inferences based on previous experience (i.e. what had been previously read). Impaired

grammatical learning would cause difficulty in predicting sentence structure, word order

and composition. This could result in an overall inability to predict linguistic features of

text and difficulty achieving normal reading fluency and comprehension. Changes in text

predictability would then serve to either facilitate or confound reading and be linked with

cognitive and attentional responses.

Neurological Deficits

No studies have been conducted to date examining the hemodynamic responses in

individuals with dyslexia using a paragraph–based reading environment. To confirm this

the terms "dyslexia", "fMRI", and "paragraph" or "continuous text" or "connected text" were

used to search the Web of Science Core Collection (Reuters, 2012) for relevant studies.

None were found using these terms-which was surprising. Paragraphs are highly common

and are regularly encountered in books, magazines, web pages, computer and mobile

applications. This makes paragraphs a natural choice as stimuli. Studying reading deficits

in this environment would give a holistic picture of the condition and as well as control for

any changes in cognition resulting from simplifying stimuli. This study aims to fill this gap

in the literature.

Studies of dyslexia have utilized single sentences as stimuli. Christodoulou et al.

(2014) studied reading fluency in developmental dyslexia by presenting sentences in a single

word format (i.e. each word was individually displayed in a serial manner). As presentation

rate increased, normal readers had significantly higher activations in the left prefrontal


cortex and left superior temporal cortex than dyslexic readers. While not a natural reading

environment this study does demonstrate reduced activation in the dyslexic model in

regions associated with the phonological component of language processing and represents

a step in the right direction. Sentences, even in a serial presentation paradigm, are more

ecologically relevant than a single word presented in isolation from linguistic context.

Other methodologies have previously been applied to isolate regions of interest.

Phonologically based tasks have successfully observed changes in the insular cortex

(Steinbrink, Groth, Lachmann, & Riecker, 2012), medial geniculate body (Diaz, Hintz,

Kiebel, & von Kriegstein, 2012), left inferior frontal gyrus (Dufor, Serniclaes,

Sprenger-Charolles, & Demonet, 2009; Karni et al., 2005; MacSweeney, Brammer, Waters,

& Goswami, 2009; Peyrin et al., 2012), left middle frontal gyrus, left precuneus, left inferior

parietal lobule, superior temporal gyrus (Pecini et al., 2011), and left occipitotemporal

areas (Conway et al., 2008; Karni et al., 2005; McCrory, Mechelli, Frith, & Price, 2005).

Procedural-based learning tasks have also identified regions of interest including the

corticostriatal circuits and the hippocampus (Krishnan et al., 2016). These regions may be

relevant to dyslexia. Procedural learning is central to learning phonology, grammar and

orthography (Christiansen & Chater, 2016; Dell & Chang, 2014). Deficits in these abilities

could contribute to impaired literacy by interfering with language acquisition. However,

their contribution to the dyslexia phenotype has not been fully elucidated and requires

additional study.

Unfortunately, serial presentation methods, single word presentation, phonological

tasks and procedural-based tasks fall short of ecologically valid stimuli. This creates a

deficit in validity and these findings may not be preserved in the face of a more natural

reading environment. Additional study is necessary to determine if and to what extent

these functional differences are preserved when engaged in normal reading.


Proposal

Goals

The goals of the study were as follows: (1) confirm whether previous findings of

hemodynamic function using single words or serial presentation paradigms are replicable

using paragraphs, and (2) characterize differences in attention and cognition in dyslexia

when reading and responding to text predictability. The following hypotheses and

experiments are proposed in support of these aims.

Hypotheses

Hypothesis 1: Can previous findings be duplicated using paragraph based

stimuli?

If dyslexia is neurological in origin then there should be a common behavioral and

physiological phenotype. Past use of simplified stimuli has revealed some differences. If

past findings are relevant to dyslexia then they will be present when using paragraph based

stimuli. These hypotheses will be tested using the following experiments.

Supporting Experiments.

1. Compare the oculomotor behavioral profile of a individuals with dyslexia with a

healthy control group while reading paragraphs, measuring: first fixation time, total

fixation time, regression probability, and refixation probability (see Hypothesis 1,

Analysis 1).

2. Compare the hemodynamic response of a dyslexic test group with a healthy control

group while reading paragraphs using whole brain analysis (see Hypothesis 1,

Analysis 2).

3. Test for differences in the BOLD response to paragraphs between groups in the

reading network (see Hypothesis 1, Analysis 3).


4. Test for differences in the BOLD response to paragraphs between groups in the eye

fields (see Hypothesis 1, Analysis 4).

5. Test for differences in the BOLD response to paragraphs between groups in the

striatum (see Hypothesis 1, Analysis 5).

Hypothesis 2: Can dyslexic and healthy cognition be characterized using

linguistic predictability?

If dyslexia impairs probabilistic learning then this impairment should generalize to

difficulty predicting and responding to linguistic features of text. Attention and cognition

supporting prediction will be measurably different from otherwise healthy cognition and be

directly associated with the linguistic features of text, i.e. lexical predictability.

Supporting Experiments.

1. Compare the oculomotor behavioral profile of a dyslexic group with a healthy control

group while reading paragraphs, measuring the influence of lexical predictability of

the currently fixated word on reading times (see Hypothesis 2, Analysis 1).

2. Test for differences in the BOLD response to linguistic predictability using whole

brain analysis (see Hypothesis 2, Analysis 2).

3. Test for differences in the BOLD response to lexical predictability in the reading

network (see Hypothesis 2, Analysis 3).

4. Test for differences in the BOLD response to lexical predictability in the eye fields

(see Hypothesis 2, Analysis 4).

5. Test for differences in the BOLD response to lexical predictability within the

striatum (see Hypothesis 2, Analysis 5).


Novelty

Functional studies of dyslexia have previously used limited stimuli including single

word or sentence reading via serial presentation paradigm. This is unfortunate as the most

pronounced effects of dyslexia become apparent in a normal reading environment which is

paragraph–based. This lack of data in the field of dyslexia makes this study a valuable

addition to current literature by determining if previous findings extend into ecological

reading environments and allows us to search for changes in oculomotor and hemodynamic

function specific to this environment. This study will also utilize some of the

recommendations put forth by Ramus, Altarelli, Jednorog, Zhao, and Scotto di Covella

(2017), including increased statistical power and the use of predefined regions of interest in

the functional and structural analysis.

Methods

Participant Recruitment

All participants were recruited from Brigham Young University, Utah Valley

University and the surrounding community. Inclusion criteria for the control group were as

follows: a minimum IQ score of 70>5 (American Psychiatric Association, 2013), right

handed, literate native English speakers, normal vision and no history of neurological

disorder, psychiatric disorder, reading disability, and 18-35 years of age. Participants could

not have any contraindications to an MRI study including pregnancy, metallic or

unidentified foreign objects within their body, non-MRI compatible prosthetic devices,

stents, shrapnel, metal fragments, or claustrophobia.

Participants in the dyslexic group met the same inclusion criteria as the control

group with the addition of a diagnosis of dyslexia or specific reading disorder. Diagnosis

was confirmed using the Gray Oral Reading Test V, Comprehensive Test of Phonological

Processing 2, and Weschler Adult Intelligence Scale 2. For the purposes of this study and as

suggested in the DSM-5, clinical judgement took precedence over the results of any single


test, thus allowing for the inclusion of participants with subtle impairment. Individuals

with dyslexia were compensated $100 for participation, and $50 for the control group.

Measures

All participants were tested using the Gray Oral Reading Test (GORT-5) (Hall &

Tannebaum, 2013). Participant intelligence was assessed via the Wechsler Abbreviated

Scale of Intelligence (WASI-II) (McCrimmon & Smith, 2013). Phonological ability was

assessed via the Comprehensive Test of Phonological Processing (Wagner et al., 1999). All

tests were administered by a trained research assistant and scored by a trained graduate

student.

Materials

Participants were presented with 54 paragraphs of text that are a subset of Luke

and Christianson (2016). These passages were drawn from a myriad of sources that include

news articles and works of literature in the public domain. These passages were

approximately 50 words in length and averaged 2.5 sentences per paragraph. Each word

received a predictability rating via a cloze procedure as outlined in Luke and Christianson

(2016) for lexical, syntactic and semantic predictability.

Apparatus

Stimuli were presented using a 24-inch BOLDScreen from Cambridge Research

Systems, an MRI-safe LCD monitor at a resolution of 1600×1200 pixels. Text were

displayed in Courier New font 35 pt, so that approximately 3.5 letters subtend to 1 degree

of visual angle. Eye-movements were recorded via a monocular SR Research Eyelink 1000

plus long-range MRI eye-tracker sampling at 1000Hz.


Procedure

Eye-tracking and fMRI data were obtained first in six 5.4 minute runs. Each

functional run included nine paragraphs and nine images. Participants were instructed to

read the paragraphs silently and to not make use of their mouths while reading to minimize

movement in the scanner. They were also told to read at a normal pace and that they are

not required to finish the paragraph within the allotted time and that it was expected they

may not finish it. They were instructed to look at an X in the bottom right-hand corner of

the screen should they finish the paragraph before time expires. Finally, participants were

instructed to read the paragraphs so they could recall them for later questioning. Within

each run, the order of paragraph presentation was randomized for each participant. Prior

to each paragraph a fixation cross was presented at the location of the first word of the

subsequent paragraph. Paragraphs were presented for 12 seconds, with a 6 second

inter-trial interval in which the participant was presented with another fixation cross.

Eye-movement Data Acquisition

Each functional run was preceded by a nine-point calibration to ensure accurate

measurement. Error during calibration was limited to an average of less than 0.49 visual

degrees of angle and a maximum error of 0.99 degrees of visual angle. Prior to each trial

and during the inter-trial interval a fixation cross was presented at the location of the first

word in the paragraph. Stimulus presentation and eye position recording were controlled

by software from SR Research.

fMRI Data Acquisition

A Siemens 3T Tim Trio was be used for this study with a 12-channel receive-only

head coil. Software version was Syngo MR B17 DHHS. Three consecutive blocks of

functional scans were performed using an interleaved T2* weighted echo-planar imaging;

slice number 43, orientation was transverse, phase encoding direction was anterior to


posterior, rotation was 0, FOV= 224x224, matrix 64x64, slice thickness 3.0 mm, TR 2500

ms, TE 28 ms, 134 repetitions, and a flip angle of 90°.

Structural Data Acquisition

A structural scan was performed using T1-weighted imaging; orientation was

sagittal, anterior to posterior phase encoding, FOV=218x250, matrix 256x256, slice

thickness 1.00 mm, TR 1900 ms, TE 2.26 ms, flip angle was 9°. Conventionally, structural

scans are performed prior to any functional scans. Functional scans were performed prior

to the structural scans in this study in order to compensate for the effects of participant

fatigue and somnolence.

Coregistration of Eye-movement and fMRI Data

The experiment was programmed so that the stimulus presentation did not begin

until the functional scan had begun. A timestamp for scan onset was received by

Experiment Builder software and stored in the data and used to compute trial and word

fixation onset times.

Analysis

Eye Tracking Analysis

DataViewer (SR Research Ltd, version 1.11.1) was used to view the eye tracking

data. Saccades were defined as eye movements as the eye traveled from one target to

another. Fixations were defined as occurring between saccades and without an intervening

blink. Fixations with a duration less than 50 ms and greater than 1500 ms were excluded.

Fixations were excluded from the study if they did not fall on a word. Saccades were

considered a regression if they end on a previously fixated word. Data was then exported

in a tab-delimited format and analyzed using R (R Core Team, 2016).


Group Structural Model Creation

Structural T1-weighted images were preprocessed with dcm2niix (Li, Morgan,

Ashburner, Smith, & Rorden, 2016) and aligned via the anterior and posterior commissures

by ACPCdetect (Ardekani & Bachman, 2009). Participant scans underwent bias field

corrections using N4BiasCorrection (Tustison et al., 2010) and normalized to MNI space

via the ICBM 152 template (Collins, Zijdenbos, Baaré, & Evans, 1999; V. S. Fonov, Evans,

McKinstry, Almli, & Collins, 2009; V. Fonov et al., 2011) via ANTs. Participant scans

were then used to construct a study specific template via buildtemplateparallel.sh

(Avants et al., 2011; Klein & Tourville, 2012). This was then segmented using the

Desikan-Killiany-Tourville protocol via antsJointLabelFusion.sh using the

OASIS-TRT-20 database of atlases (Klein et al., 2017; Klein & Tourville, 2012; Tustison

et al., 2014; Wang et al., 2013). A binary gray matter mask was then created by joining

labels via Convert3D (Yushkevich et al., 2006).

fMRI Analysis

Data were analyzed using the Analysis of Functional NeuroImages (AFNI) software

program (Cox, 1996) and Advanced Normalization Tools (Avants et al., 2011). Functional

imaging were converted from DICOM imaging format to native BRIK and HEADER

formats via to3d for analysis in AFNI. T1-weighted images were co-registered to the third

functional scan imaging using 3dWarp. Functional scans were then corrected for motion

and slice time corrections were applied using the 69th TR as a reference via 3dvolreg. The

first and second blocks were then corrected for motion relative to the third block, also via

3dvolreg. A mask was then created for each subject via 3dSkullStrip and used to

restrict the analysis to only brain matter. Individual input matrices were constructed and

deconvolution performed via 3dDeconvolve, with 6 regressors coding for motion (pitch, roll,

yaw, superior-inferior translation, left-right translation and anterior-posterior translation).

Regressors of interest were coded according to the hypothesis tested and the reader is


referred each experiment for greater detail. Deconvolution of the BOLD response was then

performed via 3dDeconvolve. A 5mm blur was applied via 3dmerge and individual

anatomical and statistical maps were projected using a study specific template aligned to

standard MNI_152 space (Collins et al., 1999; V. S. Fonov et al., 2009; V. Fonov et al.,

2011) via ants.sh (Avants et al., 2011). A random effects analysis comparing the healthy

and dyslexia groups was then performed via 3dttest++, and an anatomical mask was used

to restrict the analysis to gray matter. A voxel-wise threshold of p<0.001, cluster-size

threshold > 39, were then applied for an α < 0.05 as determined via the -Clustsim option.

Regions of Interest Selection

Regions of interest included the oculomotor network, reading network, and

prediction were identified using a large-scale automated meta-analysis of current fMRI

literature via Neurosynth (Yarkoni, Poldrack, Nichols, Van Essen, & Wager, 2011). The

eye field regions of interest were created using an automated meta-analysis using the term

"eye movement." The reading network regions of interest were created using the term

"reading." The prediction regions of interest were created using the term "prediction."

Association maps were created with a FDR of 0.01. These were then converted to labeled

areas of interest via 3dClusterize. Voxels were assigned to a cluster if they had a z-score

> 0.01, shared a face with a neighboring voxel (nearest neighbors = 1) with a minimum

cluster size of 40 voxels. Participant maximum activation for each a priori region was

calculated and extracted via 3dROIstats for both the reading and lexical predictability

conditions. This was then entered into a linear model with participant group as the

explanatory variable and maximum activation as the response variable.

Participant Characteristics

A total of 45 participants were recruited for this study. All were either current

students or former students of Brigham Young University or Utah Valley University. Both

groups were similar in terms of age, gender distribution, ethnicity and race (see Table 1).

http://www.neurosynth.org


Differences were present between groups when examining both gross intelligence (assessed

via the WASI-II), reading ability (assessed via the GORT-5 ) and phonological ability

(assessed via the CTOPP-2 ), see Table 2. For intercorrellations between the scores see

Table 9.

Table 1Participant demographics

level Control Dyslexian 21 24Gender (%) Female 14 (66.7) 14 (58.3)

Male 7 (33.3) 10 (41.7)Age (mean (SD)) 23.62 (3.43) 22.96 (3.88)Ethnicity (%) Hispanic or Latino 1 (4.8) 3 (12.5)

Not Hispanic or Latino 20 (95.2) 21 (87.5)Race (%) Hawaiian 1 (4.8) 0 (0.0)

More than one race 1 (4.8) 0 (0.0)White 19 (90.5) 24 (100.0)

University student (%) No 1 (4.8) 3 (12.5)Yes 20 (95.2) 21 (87.5)

Documentation (%) Maybe 0 (0.0) 5 (20.8)No 21 (100.0) 0 (0.0)Yes 0 (0.0) 19 (79.2)

Other learning disorders (%) No 20 (95.2) 18 (75.0)Yes 1 (4.8) 6 (25.0)

Other psychiatric disorders (%) No 16 (76.2) 10 (41.7)Yes, one or more of the above. 5 (23.8) 14 (58.3)

Note: Of the participants who were not currently attending a university, all had attended auniversity in the past.

Hypothesis 1 - Confirming Previous Findings

Analysis 1 - Oculomotor Differences

The purpose of this analysis was to examine the oculomotor profiles of each group

and determine whether group differences were present, even after controlling for the

characteristics of the text. If the findings of previous studies using simplified text are valid,

then they should extend to more ecologically valid and complex reading environments.


Analysis

Oculomotor data were analyzed in R (R Core Team, 2016). Summary statistics

were then computed for each participant and then compared between groups via Student’s

T-test.

Results

The dyslexia group displayed increased mean first fixation duration, first run dwell

time, total dwell time (total time viewing a word during a trial), as well as higher

refixation probability (see Tables 4 and 5).

Analysis 2 - Hemodynamic Differences

Analysis

This analysis searched for group differences in the BOLD response to paragraphs of

text. The purpose of this analysis was to determine whether previous findings of the

dyslexic response to text were replicable using paragraphs rather than simplified stimuli.

Ideal hemodynamic response functions were created using the appearance of a paragraph

on-screen as the onset of an interest period. Each interest period had a duration of 12

seconds to create a boxcar design. This was contrasted against periods when a fixation

cross was displayed. This allowed for the examination of the BOLD response to paragraphs

of text for each individual. Group maps were then constructed and compared via

3dttest++ to search for whole brain group differences. See the section fMRI Analysis for

more information on preprocessing.

Results

No significant differences emerged between groups.


Analysis 3 - ROI Analysis of the Reading Network

The purpose of this test was to determine whether the magnitude of the BOLD

response within reading network regions was associated dyslexia. After deconvolution (see

the previous section), individual participant activations within each region of the reading

network were extracted from the participant’s template aligned statistical map of their

response to reading via 3dROIstats. These were then entered into a linear model with

participant diagnosis as the explanatory variable and activation within the eye field as the

response variable.

Results

Dyslexia was found to be associated with reduced activation in five regions: the left

middle and inferior temporal gyrus, F(1,43) = 4.14, p = 0.0482, β=-0.656, R2 = 0.08776,

the right cerebellum, F(1,43) = 14.3, p = 0.000475, β = -0.4603, R2 = 0.2496, the right

occipital gyrus, F(1,43) = 6.14, p = -0.659, R2 = 0.1249, the left temporal pole, F(1,43) =

9, p = 0.00448, β = -0.572, R2 = 0.1731, and the right parahippocampal gyrus, F(1,43) =

4.85, p = 0.0331, β = -0.1602, R2 = 0.1013.

Analysis 4 - ROI Analysis of the Eye Fields


response within each of the eye fields was associated with reported diagnosis. To do this,

individual participant activations within each of the eye fields were extracted from the

participant’s template aligned statistical map of their response to reading via 3dROIstats.

These were then entered into a linear model with participant diagnosis as the explanatory

variable and activation within the eye field as the response variable.

Results

Two regions had decreased activation in association with dyslexia: the left parietal

eye fields, F(1,43) = 6.58, p = 0.0139, β = -0.607, and the cerebellum, F(1,43) = 12, p =


0.00124, β = -0.749.

Analysis 5 - ROI Analysis of the Caudate and Putamen


response within right and left caudate and putamen were associated with reported

diagnosis. To do this, individual participant activations within each of the eye fields were

extracted from the participant’s template aligned statistical map of their response to

reading via 3dROIstats. These were then entered into a linear model with participant

diagnosis as the explanatory variable and activation within the eye field as the response

variable.

Results

Activation during reading within the right and left caudate and putamen was not

found to be associated with dyslexia when reading.

Hypothesis 2 - Exploring Linguistic Predictability

Analysis 1: Predictability Effects on Oculomotor Behavior

The purpose of this analysis was to determine whether sensitivity to linguistic

predictability differed between groups. To do so, oculomotor data were exported via

DataViewer and analyzed in R (R Core Team, 2016), these were then combined with

lexical predictability data from Luke and Christianson (2016). Eye movements were then

entered into a linear mixed-effects model via lmer, with an oculomotor response variable

(e.g. first fixation duration, total dwell time). Explanatory variables included reported

diagnosis and the lexical predictability of the fixated word. Random effects were entered

for each participant. Statistical significance was then evaluated via lmerTest (Bates,

Mächler, Bolker, & Walker, 2015; Kuznetsova, Brockhoff, & Christensen, 2017).


Results

Lexical predictability had little effect on the duration of the first fixation. It did

however increase the duration of first run dwell time indicating it had a facilitating effect

on reading behavior for both groups. It also positively associated with increased total dwell

time (p < 0.001) and interacted with group (see Tables 6, 7 and 8), demonstrating an

increased influence in dyslexia relative to controls.

Analysis 2 - Group Differences in BOLD Response to Predictability

Analysis

The purpose of this analysis was to determine if any group differences existed in the

hemodynamic response to linguistic predictability. Participants were assigned to groups

based on reported diagnosis. Individual hemodynamic response functions were then created

using a parametric regressor coding for lexical predictability. A second parametric regressor

was then added encoding fixation duration. This resulted in an amplitude and wavelength

modulated hemodynamic response function. Group statistical maps were then constructed

and compared via 3dttest++.

Results

No significant differences were discovered between groups across any of the

regressors of interest.

Analysis 3 - ROI Analysis of the Reading Network

The purpose of this test was to determine whether the magnitude of BOLD

response to lexical predictability in each part of the reading network was associated with

diagnosis. After deconvolution, individual participant activations within each region of the

reading network were extracted from the participant’s template aligned statistical map of

their response to lexical predictability via 3dROIstats. These were then entered into a


linear model with participant diagnosis as the explanatory variable and activation within

each region as the response variable.

Results

Dyslexia was found to be associated with reduced activation in two regions: one

centered on the left supplementary motor area and posterior middle frontal gyrus, F(1,43)

= 6.101, p = 0.0176, β = -2.6823, R2 = 0.1242, and the left temporal pole, F(1,43) =

5.303, p = 0.0262, β = -2.0648, R2 = 0.1098.

Analysis 4 - ROI Analysis of the Eye Fields


response to lexical predictability in each of the eye fields was associated with diagnosis.

After deconvolution, individual participant activations within each region of the reading

network were extracted from the participant’s template aligned statistical map of their

response to lexical predictability via 3dROIstats. These were then entered into a linear

model with participant diagnosis as the explanatory variable and activation within each

region as the response variable.

Results

One region of interest had decreased activation in the dyslexia group: the left

frontal eye fields, F(1,43) = 4.846, p = 0.0331, β = -2.4093, R2 = 0.1013.

Analysis 5 - ROI Analysis of the Caudate and Putamen


response to lexical predictability in each caudate and putament was associated with

diagnosis. After deconvolution, individual participant activations within each region of the

reading network were extracted from the participant’s template aligned statistical map of

their response to lexical predictability via 3dROIstats. These were then entered into a


linear model with participant diagnosis as the explanatory variable and activation within

each region as the response variable.

Results

The BOLD response to lexical predictability within the right and left caudate and

putamen was not found to be associated with dyslexia.

Post-Hoc Analyses

While a prior results were promising, it must be acknowledged that the reliability of

participant knowledge concerning a diagnosis of dyslexia was at times doubtful. Therefore,

several secondary analyses were undertaken to reexamine the findings of the regions of

interest analyses. Rather than sort participants by reported diagnosis, we developed a

secondary measure of reading impairment via factor analyses. Participants were then

ranked and reassigned to a group based on this measure. This resulted in the reassignment

of two control participants to the dyslexia group and four dyslexia participants to the

control group. The BOLD response of each participant to paragraphs (and not lexical

predictability) were then extracted from the same regions and were entered into a linear

model, with the composite score as the explanatory and participant β values as the

response variable.

Factor Analysis

In order to create a latent variable that represented group differences, an

exploratory factor analysis was conducted. All participant scores from the following tests

were used: reading fluency and comprehension from the GORT-5; phonological ability and

rapid symbolic naming from the CTOPP-2. These were entered into a principle

components analysis with varimax rotation using princomp from R (R Core Team, 2016).

The largest factor, with an eigen value of 1.46, explained 53% of the variance.


Intercorrelations between these factors are reported in Table 9 and factor loading scores are

found in Table 10. Changes in group assignment can be found in Table 11.

Analysis 1 - a Second Look at the Reading Network

Of the 12 regions represented in the mask, four had a significant association with

the composite measure: the left inferior and middle temporal gyri (F(1,43) = 7.775,

p<0.00786) with an R2 of 0.1531 and β=-0.2947, another comprising the right lobe of the

cerebellum (F(1,43) = 10.02, p<0.00284) with an R2 of 0.4234 and β=-0.13632, the left

medial temporal pole (F(1,43)=9.636, p<0.0037) with an R2 of 0.6345 and β=-0.20030,

and the left inferior temporal gyrus (F(1,43)=5.007,p<0.0305) with an R2 of 0.1043 and

β=-0.05530. These findings are similar in terms of the regions associated with dyslexia and

the direction of the association to that of the a priori analysis. Only the parahippocampal

gyrus is absent.

Analysis 2 - a Second Look at the Eye Fields

Of the 12 regions tested, two had a significant association with the composite

measure: the left parietal eye fields (F(1,43) = 4.825, p < 0.0335) with an R2 of 0.1009,

and the cerebellum (F(1,43) = 7.272, p < 0.00996) with an R2 of 0.1447. Slow reading was

associated with a decrease in activation by -0.18025 in the left parietal eye fields, and a

decrease of -0.2079 in the cerebellum. This is also a similar result to the a priori analysis

as well, in that these same regions were negatively associated with dyslexia.

Analysis 3 - a Second Look at the Striatum

Participant activation within the caudate and putamen was not found to be

associated with the composite measure, the same as in the a priori analysis.


Discussion

This dissertation presents the results of an experiment with ten a priori and three

post-hoc analyses, examining reading behavior and cognition in individuals with dyslexia,

or specific reading disorder. Twenty-four participants were recruited for the dyslexia group

and twenty-one for the control group. Each underwent both fMRI and eye tracking while

reading paragraphs of text, followed by a single structural MRI. These data were then used

to compare the groups, searching for differences in both gross differences in oculomotor

behavior while reading, sensitivity to linguistic characteristics of text, and hemodynamic

fluctuation. In the sections that follow, the oculomotor results will first be considered,

followed by a discussion of the whole-brain analysis. This is then followed by a discussion

of the findings of the regions of interest analyses, organized according to region rather than

analysis. Theoretical implications of these findings are then outlined and future directions

of research are proposed.

Oculomotor Behavior in Dyslexia

Reproducing Previous Findings

Group oculomotor profiles were as expected. Overall the dyslexia group displayed

increased first fixation duration, first run dwell time, and total dwell time. This means

they spent more time viewing individual words (and therefore fewer words in a trial period)

than the control group. Additionally they were more likely to fixated upon a word multiple

times before leaving it as demonstrated by the increased refixation probability. This is a

general characteristic of readers who are impaired or unskilled relative to their peers and

would be expected in dyslexia.

Effect of Lexical Predictability

Lexical predictability is a measure of the likelihood of the occurrence of a word in

the context of the surrounding text. For example, how likely is it that the word car will


appear as the last word of the sentence "I want to drive the..." as compared to the word

airplane? Words with a higher lexical predictability are generally more expected in the

context of their text just as car is more expected than airplane. Words with higher lexical

predictability receive shorter and fewer fixations than those with lesser predictability.

In this experiment, participants in both groups were found to be sensitive to lexical

predictability. Individuals with dyslexia were found to have increased first fixation

duration, first run dwell time, and total dwell time relative to controls. Of these, first run

dwell time and total dwell time had a significant relationship with lexical predictability

(see Tables 7 and 8; Figures 5 and 6). Both had a negative relationship with lexical

predictability, indicating the dyslexia group’s reading times were strongly influenced by

lexical predictability. The interaction between predictability and group in the total dwell

time analysis indicates that dyslexic participants are even more sensitive to predictability

than controls.

This demonstrates that individuals with dyslexia are more dependent on predictions

of the text while reading than other readers. Impairment in decoding word meaning, either

phonologically or orthographically, results in degraded input to higher cognitive centers

and increases their reliance on predictions about the text, resulting in increased time

fixating upon old words.

What Happened to Group Differences in the BOLD Response?

The lack of group differences in the BOLD response to paragraphs and linguistic

predictability was perplexing but not inexplicable. For example, this may simply be due to

inadequate statistical power. Both groups have a small sample size (21 in the control and

24 in the test group). It should be noted that the activation maps for each group are

somewhat similar in both activation geography and magnitude (see Figure 3) despite

differences in their cognitive test scores (see Table 3). It may be that differences between

these groups are subtle, requiring a much larger sample size to be detected as a change in


the BOLD response. The fact that this difference appears in the regions of interest-based

analysis supports this explanation.

Recruitment for this group was difficult. During the first year of this study we relied

exclusively on recruiting through the University Accessibility Centers of Brigham Young

University and Utah Valley University and received only four referrals during that time. In

order to increase recruitment, we began to advertise the study in the community via fliers

and word of mouth. Recruitment increased substantially, but we became more reliant on

participants accurately reporting and understanding their diagnoses, which may have been

a source of confounding by including individuals in the study who did not have a true

diagnosis, misunderstood their diagnosis or who may not have been truthful.

Despite this, all participants were either current or former university students,

which could have also been a source of confounding. Our test group is unique in that they

were able to meet the requirements for admission at a University. They had previously

displayed adequate scholastic aptitude for enrollment, and may have developed their own

cognitive compensations for any impairment prior to admission and diagnosis, resulting in

increased similarity with normal readers. It is also possible that their individual

compensations introduced additional variability within the test group, weakening any

potential differences from the control group BOLD response.

It may also be possible that confounding was introduced by potential subgroups of

dyslexia. This study did not separate the test group into subgroups based on the degree or

absence of phonological impairment. Current evidence suggests individuals can have

homogeneous reading profiles and yet have distinct causes of cognitive impairment

(Zoubrinetzky, Bielle, & Valdois, 2014). This may be the case here. Additional study is

necessary to examine the test group for the presence of subgroups.

Finally, previous studies have demonstrated differences using simplified stimuli

(Christodoulou et al., 2014; B. A. Shaywitz et al., 2002) rather than full paragraphs of

text. This is the first study to utilize paragraphs of text as stimuli in a study of dyslexia.


It is possible the paragraphs themselves introduced some confounding in the overall BOLD

response of the test group. Paragraphs are semantically rich stimuli. One need not look at

or decode every word to understand the meaning of the whole text. In fact, no reader ever

directly gazes upon every single word in a body of text (unless they are an unskilled

reader). This study demonstrates this; even the control group regularly skipped words (see

Table 4). It is possible that the additional information provided by the paragraphs allowed

the dyslexic group to adequately comprehend the text and to appear similar to the control

group. This could explain the lack of differences between the groups, though additional

study is needed to be sure of this. Future studies should implement single word or rapid

serial visual presentation based paradigms as well as paragraph based text in order to draw

direct comparisons between each method and determine the appropriate application of

each paradigm.

Why did the ROI analyses work? A note must be said concerning the success

of the ROI analyses. Despite a lack of differences in the whole brain analysis, the ROI

analyses did reveal several regions to be associated with dyslexia. ROI based analyses offer

a specific advantage compared to a whole-brain analysis, the minimization of statistical

tests. Second level analyses are much less reliable compared to region of interest based

analyses. ROI based analyses offer greater statistical control by minimizing the number of

statistical tests performed and decreasing Type I error (Poldrack, 2007). Therefore, the

results of the ROI analyses are much more robust than the group level results.

BOLD Response in Regions of Interest

Left Temporal Lobe

Activation within the left temporal lobe was found to have a negative association

with dyslexia in both the a priori and post-hoc ROI analyses of the reading network. The

left temporal lobe plays an important role as part of the ventral stream in reading, and

activation is directly tied to reading ability (Noppeney, Price, Duncan, & Koepp, 2005),


therefore its relationship with dyslexia is not surprising. Structures in this region are

generally involved in the orthographic processing of text and integrating it with semantic

representations. Decreases in activation in this region is interesting and reflects impairment

in tying semantic knowledge with word orthography (the visual form of the word). This is

consistent with their sensitivity to lexical predictability. Decreases in lexical predictability

were associated with increased total dwell time (the sum of fixations upon a word during

the first pass and all subsequent passes through the word). This could be driven by an

inability to associate semantic knowledge with word orthography, driving the reader to

spend more time on that word during a trial in order to divine its meaning. Decreased

temporal activation could be a driver of this behavior, i.e. the lack of temporal activation

could indicate decreased retrieval, which could be due to decreased bottom-up input to

language and word specific cortical areas. This would impede orthographic processing.

This particular ROI was quite large as defined. It included much of the left

posterior middle temporal gyrus, left inferior temporal gyrus and left fusiform gyrus.

Therefore the BOLD response within this region includes functional regions such as the

visual word form area and fusiform face area. This is a rather wide net to cast and both of

these regions have been shown to be involved in linguistic processing. It is vital in follow

up studies to further segment this large region of interest into smaller functional regions.

This could be done via additional meta-analyses using Neurosynth (if using the same

dataset) or ideally a localizer task in a follow-up study. A localizer task would enable

researchers to find the exact location of each functional region for each participant,

allowing for greater precision for each participant.

Left Temporal Pole

The left temporal pole likewise had a lower activation in relation to dyslexia in the

reading contrast as well as in the lexical predictability contrast. Previously, this region has

been shown to be involved in accessing semantic information (Binder, Desai, Graves, &


Conant, 2009; Price, 2012), sensitive to semantic integration and semantic prediction in

normal readers (Weber, Lau, Stillerman, & Kuperberg, 2016; Willems, Frank, Nijhof,

Hagoort, & Van den Bosch, 2016). It has also been shown to be sensitive to lexical

predictability in skilled readers (Carter, Foster, Muncy, & Luke, 2019). Willems et al.

(2016) observed an association with word surprisal values, (a measure inversely correlated

with predictability, i.e. the harder it is to predict a word the higher the surprisal value). As

surprisal increased, temporal activation increased.

Additionally the temporal poles are thought to be associated with accessing

syntactic information (J. R. Brennan, Stabler, Van Wagenen, Luh, & Hale, 2016;

J. Brennan et al., 2012; Dronkers, Wilkins, Van Valin Jr, Redfern, & Jaeger, 2004;

Henderson, Choi, Lowder, & Ferreira, 2016) and assimilating that information with

semantic knowledge (Wilson et al., 2014). In a previous study on reading behavior (see

Carter et al. (2019)), the left temporal pole demonstrated activity that was uniquely tuned

to lexical predictability, rather than semantic or syntactic predictability and it is thought

to be dedicated to integrating whole word morpho-syntactic information with specific

rather than general semantic integration. This study found a negative association between

dyslexia and the BOLD response to lexical predictability.

The reduction in activation as a result of dyslexia is interesting and again could

indicate a reduction in input from bottom-up processes. This degraded input would result

in reduced activation of this region. This could lead to impaired integration of semantic

knowledge with word form and make it difficult to understand complex, unfamiliar or

infrequent words. Such individuals would be highly reliant on predictable words when

reading like this test group. This would also impair an individual’s ability to integrate

semantic and syntactic information about the text, leading to increased confusion about

what was being read. The dyslexia group did score lower on comprehension questions in

the GORT-5 (t(42.567) = 2.7248, p = 0.009298). This contributed to lower sums of scaled

scores for the GORT-5 (see Figure 2, Tables 2 and 3). It is likely this impaired


comprehension is a product of decreased activation of the left anterior temporal pole, and

may be driven by decreased input to that region from other lower level processes.

Left Supplementary Motor Area and Posterior Middle Frontal Gyrus

This region of the brain has been reported as being central to the covert production

of articulation (subvocalizing) (Martin, Schurz, Kronbichler, & Richlan, 2015), and the

generation of eye movements (McDowell, Dyckman, Austin, & Clementz, 2008; Müri &

Nyffeler, 2008). It has direct connections to oculomotor nuclei in the brainstem. Activity

in this region is positively associated with increased saccade frequency, and usually

increases with the cognitive load of the task (McDowell et al., 2008). This region has also

been shown to be sensitive to text (Henderson, Choi, Luke, & Desai, 2015). More generally

it appears to play a role in goal directed behavior. According to Stuphorn (2015), this

region may be involved in learning context-dependent rules in order to predict the outcome

of an oculomotor plan. In terms of a reading paradigm, the role of the SMA would be to

integrate input from current and previous fixations to predict the outcome of a planned

saccade. This would aid in the planning and execution of an effective reading behavior.

Lowered activation of this region is interesting in the setting of dyslexia. These

individuals generally engage in longer fixations and shorter saccades. The SMA would be

expected to have reduced activation as a result. In this study, we have also shown the

oculomotor behavior of the dyslexia group to be highly dependent on lexical predictabilty.

Considering with the finding that the SMA activation was associated with reduced

activation in lexical predictability, one begins to see more evidence of a importance of and

reliance on lexical predictability for the dyslexic reader.

Right Inferior Occipital Lobe

This particular region of interest covered a number of structures including the

middle occipital gyrus, inferior occipital gyrus and lingual gyrus. Sakurai (2004) observed

that lesions in this region seemed to impair tying word forms to phonology. Tan, Laird, Li,


and Fox (2005) likewise observed that it was associated with phonological processing.

Cattinelli, Borghese, Gallucci, and Paulesu (2013) found this region to be associated with

pseudoword reading via a meta-analysis, and proposed this was an early site for visually

processing words. Overall, this region appears to be the first region to attempt to associate

word form with phonology since it neighbors the primary visual cortex, thus it should be

expected that impairment and dysfunction here is associated with greater difficulty in

identifying novel words via phonological decoding. Likewise, we have also observed this

region is negatively associated with lexical predictability (Carter et al., 2019), and

therefore is more active as word predictability decreases. This would be expected as

cognitive load increases when greater attention must be given to decipher it. It is not

surprising that it should appear in this study.

This is not the first time this region has appeared in studies of dyslexia.

B. A. Shaywitz et al. (2002) observed a decrease in activation relative to normal readers

during the use of a semantic category judgement task. Our findings are similar in that

paragraph reading is a highly semantic task, requiring the integration of the meaning of

many different words into a single concept. Differences this early in the decoding process

would certainly result in impaired integration with semantic knowledge at higher levels of

cognition.

Right Parahippocampal Gyrus and the Hippocampus

The involvement of this region is interesting. This region was also reported by

Cattinelli et al. (2013) in their meta-analysis. In that particular analysis it appeared as a

non-specific cluster, meaning that it did not appear to be associated with any specific

reading task and only appeared to be generally involved. Willems et al. (2016) noted the

region appeared to respond to surprisal, meaning that it became more active as surprisal

increased. This being the inverse of predictability, it is interesting to note it was also

involved in this study, with activation being reduced in the dyslexic group. This particular


region is too far forward to be involved in early processing of orthographic or phonological

representations of text. Rather, since it is sensitive to surprisal, it is possible this region’s

reduced activity is due to decreased retrieval of semantic information during reading,

relative to normal readers. This may play a role in the dyslexia test groups propensity to

return to words after completing their first pass through. It could be the decreased

activation in this region results in delayed word recognition, or integration of semantic and

phonological information, thus driving a need to refixate upon it multiple times while

reading. The reasons for this reduction in activity could be similar to those mentioned

before, such as insufficient bottom-up input to correctly tie semantic and phonological

information.

The Cerebellum

The cerebellum is not a structure one would think has a direct association with

reading. However it is often activated in reading and reading related tasks (Stoodley &

Stein, 2011, 1). Mariën et al. (2014) proposed the cerebellum was central to the

development of many processes supporting reading including fixation accuracy, hand-eye

coordination, eye-voice coordination, sensori-motor-cognitive integration, speech

internalisation, processing speed, verbal working memory, word recognition, and

grapheme-phoneme isolation. Therefore it is not surprising this region appeared in as part

of the meta-analyses for the reading and eye fields.

There is even a "cerebellar deficit" hypothesis which proposes the cerebellum

contributes to the development of articulation and phonological ability and skill

automatisation (Baddeley, 2003; Marvel & Desmond, 2010; Nicolson, Fawcett, & Dean,

2001). This would also result in deficits in procedural learning, which has also been

observed (Pavlidou et al., 2010) with mixed success (Nigro, Jiménez-Fernández, Simpson,

& Defior, 2016; Schmalz, Altoè, & Mulatti, 2017). Therefore it is not surprising cerebellar

activation is negatively associated with dyslexia.


Left Parietal Eye Fields

The parietal eye fields are central to visual integration and attention. Their

proximity allows them to influence attention allocation earlier than higher cortical regions.

It’s primary role is control of the reflexive exploration of the visual environment, i.e. it

controls whether or not saccades continue to be executed or if they are inhibited to allow

the gaze to be focused on an item of interest (Müri & Nyffeler, 2008). A recent lesion study

(Valdois, Lassus-Sangosse, Lallier, Moreaud, & Pisella, 2019) in a patient with bilateral

loss of the parietal eye fields has raised the possibility that selective impairment of visual

attention could lead to dyslexia without phonological impairment.

Poor Input or Output?

The overall findings of an increased sensitivity to lexical predictability, and

reductions in the BOLD response relative to controls in the ROI analysis leads us to a

question why this is the case. Is this phenomenon driven by bottom-up errors or top-down

errors? Is the sensitivity to lexical predictability a result of poor input from poor attention

or poor phonological and/or orthographic decoding? Or is this due to faulty prediction

machinery? While this study cannot provide a definitive answer to these questions, some

clues are present. For example, the overall depression in the bold response suggests a

potential lack of stimulation to these regions relative to controls. This is consistent with a

deficit in bottom up processing rather than a top-down etiology. Activation within one

region of the cortex should be directly related to the degree of input received from other

regions according to basic Hebbian theory (Hebb, 1949). Additionally, this depression of

activity could indicate reading behavior is less goal oriented in dyslexia. Indeed, the frontal

eye fields and supplementary eye fields are responsible for developing an oculomotor plan

fitted to the objectives of the individual (Müri & Nyffeler, 2008). This would also explain

the sensitivity to lexical predictability. Dyslexic readers may essentially be "along for the

ride" when reading and without a definite strategy. A deficit in bottom-up information


would make it difficult to engage in goal oriented reading; one must have preliminary

information about the text in order to decide how to read it. This lack may be leaving

dyslexic readers unable to make effective oculomotor plans, leaving them at the mercy of

the text rather than the masters of it.

More study is necessary to explore this idea – that there is a bottom-up deficit in

dyslexia. Past studies have pointed towards potential causes such as a magnocellular deficit

(D’Mello & Gabrieli, 2018; Stein, 2019). The magnocellular pathway is an important

component of early visual processing beginning in the retina and extending through the

thalamus to the visual cortex. This particular pathway is central to the allocation of visual

attention, performing visual search tasks, determining the location of objects and their

relation to each other. Pre-readers later diagnosed with dyslexia have been found to have

deficits in recognizing individual letters (Ozernov-Palchik et al., n.d.) and sequencing them

correctly (Howard Jr, Howard, Japikse, & Eden, 2006). The magnocellular pathway

therefore is important for allocating visual attention in reading (Vidyasagar & Pammer,

2010). Such a lower level deficit could explain the lack of input to higher cortical centers in

this study. Follow-up is needed to confirm this.

Next Steps

A few recommendations can be made based upon this study. First, studies using

paragraph based stimuli in the study of reading need greater statistical power. Similar

studies in readers used much larger sample sizes to study skilled readers (Carter et al.,

2019; Choi, Lowder, Ferreira, Swaab, & Henderson, 2017; Henderson et al., 2016;

Henderson et al., 2015) and even larger samples may be necessary to detect potentially

subtle changes in dyslexia. Second, future studies of reading should include simplified

readings tasks such as single or rapid visual presentation paradigms, in addition to

paragraphs of text. This will enable researchers to draw specific conclusions concerning the

effectiveness and application of each method in studying reading and dyslexia. Such tasks


also have the added benefit of serving as a localizer task for more targetted region of

interest analyses. Third, individuals reporting a diagnosis of dyslexia are usually reliable in

their reporting (at least in our sample population). This is supported by the preservation

of the findings of the ROI analyses between the a priori and post-hoc analyses. However, it

is still wise to independently administer cognitive tests for reading and phonological ability

to ensure data integrity. Fourth, individuals with dyslexia appear to have an overall

depression in the BOLD response in higher cortical regions. Finally, future studies should

consider the possibility that deficits in bottom-up processes like the magnocellular pathway

are the primary contributors to dyslexia and implement this into their experimental design.

Stein (2019) describes several examples of experimental paradigms that could be used to

this effect.

Conclusion

This study confirms findings of previous studies of dyslexia, i.e. that there are

functional differences between dyslexic and otherwise normal individuals. The ROI analysis

revealed that individuals with dyslexia had reduced activation compared to the control

group. Word predictability was likewise associated with decreased total dwell time,

implying that individuals with dyslexia are more reliant on word predictability than

average readers in order to successfully decipher text. This is more evidence of a greater

reliance on lexical predictability in dyslexia, however more study is needed to determine

the cause–whether it is due to defective predictive machinery or whether this is due to

degraded input. Given reduction in the BOLD response in all studied regions of interest,

and the increased reliance on word predictability, it is proposed this is due to degraded

input to the higher cortical centers and other regions engaged in prediction in dyslexia

rather than defective predictions. A prime focus of future studies should include

components of the magnocellular pathway as described by Stein (2019).


Figure 1Regions of Interest and Gray Matter Mask

A. The eye fields. This includes the frontal eye fields, dorsolateral prefrontal cortex, parietaleye fields, supramarginal gyrus, visual cortex, cerebellum, and right putamen. B. Thereading fields. This includes the left inferior frontal gyrus, left temporal pole, left inferiortemporal gyrus, left fusiform gyrus, left parietal lobe, left posterior superior frontal gyrus,right parahippocampal gyrus, right fusiform gyrus, and right parietal lobe. C. Predictionfields. This primarily focused on the head of the caudate and part of the putamen. D.Whole brain mask. This mask was used to restrict the whole brain analysis to gray matter.


Table 2Cognitive test results

Test Measure Control DyslexiaWASI-II FSIQ-2 119.62 (8.8) 112.5 (6.47)GORT-5 Sum of Scaled Scores 22.86 (2.03) 16.62 (3.06)CTOPP Phonological Awareness 114.43 (8.21) 107 (9.35)

Rapid Symbolic Naming 105.62 (11.28) 83.38 (17.29)

Table 3Student’s T-test results

Test Measure T-statistic Degrees of Freedom p-valueWASI-II FSIQ-2 3.055 36.32 0.0042GORT-5 Sum of Scaled Scores 8.133 40.25 p<0.0001CTOPP Phonological Awareness 2.839 43.00 0.0069

Rapid Symbolic Naming 5.170 39.95 p<0.0001

Table 4Oculomotor profiles

Statistic Control DyslexiaMean (SD) Mean (SD)

First fixation duration 215.52 (92.01) 255.46 (144.05)First run dwell time 244.13 (123.85) 324.92 (217.96)Total dwell time 287.09 (180.24) 392.06 (273.75)Refixation probability 0.28 (0.08) 0.34 (0.11)

Table 5Significance of group differences

T-statistic Degrees of Freedom p-valueMean first fixation duration -5.030 42.97 p<0.0001Mean first run dwell time -5.291 38.75 p<0.0001Mean total dwell time -6.116 37.32 p<0.0001Refixation probability -5.532 42.46 p<0.0001

Table 6First fixation

Estimate Std. Error Degrees of Freedom T-statistic p-value(Intercept) 5.286 0.020 43.59 268.411 p<0.0001Dyslexia 0.149 0.027 43.87 5.579 p<0.0001Lexical Pred. -0.001 0.002 33482.83 -0.481 0.631Dyslexia * Lexical Pred. 0.000 0.003 33485.38 0.010 0.992


Figure 2Distribution of cognitive test scores

Distribution of the scores by group for the (A) Full Scale IQ-2 (WASI-II), (B) Sum of Scaled Scores(GORT-5), (C ) Phonological Awareness (CTOPP-2), and (D) Rapid Symbolic Naming (CTOPP-2).


Figure 3Whole brain response to paragraphs by group.

Whole brain responses to text by group: A. Control group. B. Dyslexia group. Voxelwisethreshold p<0.001, cluster threshold = 39 voxels, nearest neighbors = 1, overall α < 0.05.

Table 7First run dwell time

Estimate Std. Error Degrees of Freedom T-statistic p-value(Intercept) 5.392 0.021 43.56 250.796 p < 0.0001Dyslexia 0.232 0.029 43.88 7.947 p < 0.0001Lexical Pred. -0.005 0.002 33486.63 -3.540 p < 0.0001

Table 8Total dwell time

Estimate Std. Error Degrees of Freedom T-statistic p-value(Intercept) 5.520 0.023 43.41 241.43 p<0.0001Dyslexia 0.274 0.031 43.75 8.82 p<0.0001Lexical Pred. -0.006 0.002 33482.89 -2.64 0.008Dyslexia * Lexical Pred. -0.010 0.003 33485.94 -2.88 0.004


Figure 4The effect of group and lexical predictability on first fixation duration

Table 9Intercorrelations of cognitive subtest scores

1 2 3 41. Fluency (GORT-5) 1.00 0.44 0.44 0.582. Comprehension (GORT-5) 0.44 1.00 0.45 0.023. Phonological Awareness (CTOPP-2) 0.44 0.45 1.00 0.264. Rapid Symbolic Naming (CTOPP-2) 0.58 0.02 0.26 1.00


Figure 5The effect of group and lexical predictability on first run dwell time

Table 10Factor loadings from Principle Components Analysis

Factor 1 Factor 2 Factor 3 Factor 4Fluency -0.595 0.199 -0.363 0.689Comprehension -0.448 -0.619 -0.461 -0.451Phonological Awareness -0.510 -0.283 0.809Rapid Symbolic Naming -0.430 0.705 -0.563


Figure 6The effect of group and lexical predictability on total dwell time

Table 11Changes in group assignment from factor analysis

Assignment by diagnosisControl Dyslexia Total

Assignment by factorControl 19 4 23Dyslexia 2 20 22Total 21 24 45


References

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental

disorders (DSM-5®). American Psychiatric Pub.

Ardekani, B. A. & Bachman, A. H. (2009). Model-based automatic detection of the

anterior and posterior commissures on mri scans. Neuroimage, 46 (3), 677–682.

Arnbak, E. (2004). When are poor reading skills a threat to educational achievement?

Reading and Writing, 17 (5), 459–482.

Avants, B. B., Tustison, N. J., Song, G., Cook, P. A., Klein, A., & Gee, J. C. (2011). A

reproducible evaluation of ants similarity metric performance in brain image

registration. Neuroimage, 54 (3), 2033–2044.

Baddeley, A. (2003). Working memory: Looking back and looking forward. Nature Reviews

Neuroscience, 4 (10), 829–839.

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models

using lme4. Journal of Statistical Software, 67 (1), 1–48. doi:10.18637/jss.v067.i01

Bellocchi, S., Muneaux, M., Bastien-Toniazzo, M., & Ducrot, S. (2013). I can read it in your

eyes: What eye movements tell us about visuo-attentional processes in developmental

dyslexia. Res Dev Disabil, 34 (1), 452–60. doi:10.1016/j.ridd.2012.09.002

Berlin, R. (1887). Eine besondere art der wortblindheit (dyslexie). Bergmann.

Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009). Where is the semantic

system? A critical review and meta-analysis of 120 functional neuroimaging studies.

Cerebral Cortex, 19 (12), 2767–2796.

Biscaldi, M., Gezeck, S., & Stuhr, V. (1998). Poor saccadic control correlates with dyslexia.

Neuropschologia, 36 (11), 1189–1202. doi:10.1016/S0028-3932(97)00170-X

Brennan, J. R., Stabler, E. P., Van Wagenen, S. E., Luh, W.-M., & Hale, J. T. (2016).

Abstract linguistic structure correlates with temporal activity during naturalistic

comprehension. Brain and Language, 157, 81–94.

https://dx.doi.org/10.18637/jss.v067.i01

https://dx.doi.org/10.1016/j.ridd.2012.09.002

https://dx.doi.org/10.1016/S0028-3932(97)00170-X


Brennan, J., Nir, Y., Hasson, U., Malach, R., Heeger, D. J., & Pylkkänen, L. (2012).

Syntactic structure building in the anterior temporal lobe during natural story

listening. Brain and Language, 120 (2), 163–173.

Carter, B. T., Foster, B., Muncy, N. M., & Luke, S. G. (2019). Linguistic networks

associated with lexical, semantic and syntactic predictability in reading: A

fixation-related fmri study. Neuroimage, 189, 224–240.

Caspi, A., Wright, B. R. E., Moffitt, T. E., & Silva, P. A. (1998). Early failure in the labor

market: Childhood and adolescent predictors of unemployment in the transition to

adulthood. American Sociological Review, 424–451.

Castles, A. & Friedmann, N. (2014). Developmental dyslexia and the phonological deficit

hypothesis. Mind & Language, 29 (3), 270–285. doi:10.1111/mila.12050

Cattinelli, I., Borghese, N. A., Gallucci, M., & Paulesu, E. (2013). Reading the reading

brain: A new meta-analysis of functional imaging data on reading. Journal of

Neurolinguistics, 26 (1), 214–238.

Catts, H. W., McIlraith, A., Bridges, M. S., & Nielsen, D. C. (2016). Viewing a

phonological deficit within a multifactorial model of dyslexia. Reading and Writing,

30 (3), 613–629. doi:10.1007/s11145-016-9692-2

Centanni, T. M., Booker, A. B., Chen, F., Sloan, A. M., Carraway, R. S., Rennaker, R. L.,

. . . Kilgard, M. P. (2016). Knockdown of dyslexia-gene dcdc2 interferes with speech

sound discrimination in continuous streams. J Neurosci, 36 (17), 4895–906.

doi:10.1523/JNEUROSCI.4202-15.2016

Che, A., Girgenti, M. J., & LoTurco, J. (2014). The dyslexia-associated gene dcdc2 is

required for spike-timing precision in mouse neocortex. Biol Psychiatry, 76 (5),

387–96. doi:10.1016/j.biopsych.2013.08.018

Che, A., Truong, D. T., Fitch, R. H., & LoTurco, J. J. (2016). Mutation of the

dyslexia-associated gene dcdc2 enhances glutamatergic synaptic transmission between

https://dx.doi.org/10.1111/mila.12050

https://dx.doi.org/10.1007/s11145-016-9692-2

https://dx.doi.org/10.1523/JNEUROSCI.4202-15.2016

https://dx.doi.org/10.1016/j.biopsych.2013.08.018


layer 4 neurons in mouse neocortex. Cereb Cortex, 26 (9), 3705–3718.

doi:10.1093/cercor/bhv168

Choi, W., Lowder, M. W., Ferreira, F., Swaab, T. Y., & Henderson, J. M. (2017). Effects of

word predictability and preview lexicality on eye movements during reading: A

comparison between young and older adults. Psychology and Aging, 32 (3), 232.

Christiansen, M. H. & Chater, N. (2016). The now-or-never bottleneck: A fundamental

constraint on language. Behav Brain Sci, 39, e62. doi:10.1017/S0140525X1500031X

Christodoulou, J. A., Del Tufo, S. N., Lymberis, J., Saxler, P. K., Ghosh, S. S.,

Triantafyllou, C., . . . Gabrieli, J. D. (2014). Brain bases of reading fluency in typical

reading and impaired fluency in dyslexia. PLoS One, 9 (7), e100552.

doi:10.1371/journal.pone.0100552

Collins, D. L., Zijdenbos, A. P., Baaré, W. F. C., & Evans, A. C. (1999).

ANIMAL+INSECT: Improved cortical structure segmentation. Biennial International

Conference on Information Processing in Medical Imaging.

Conway, T., Heilman, K. M., Gopinath, K., Peck, K., Bauer, R., Briggs, R. W., . . .

Crosson, B. (2008). Neural substrates related to auditory working memory

comparisons in dyslexia: An fmri study. J Int Neuropsychol Soc, 14 (4), 629–39.

doi:10.1017/S1355617708080867

Cox, R. W. (1996). Afni: Software for analysis and visualization of functional magnetic

resonance neuroimages. Computers and Biomedical Research, 29, 162–173.

De Luca, M., Borrelli, M., Judica, A., Spinelli, D., & Zoccolotti, P. (2002). Reading words

and pseudowords: An eye movement study of developmental dyslexia. Brain and

Language, 80 (3), 617–626.

De Luca, M., Di Pace, E., Judica, A., Spinelli, D., & Zoccolotti, P. (1999). Eye movement

patterns in linguistic and non-linguistic tasks in developmental surface dyslexia.

Neuropschologia, 37 (12), 1407–1420. doi:10.1016/S0028-3932(99)00038-X

https://dx.doi.org/10.1093/cercor/bhv168

https://dx.doi.org/10.1017/S0140525X1500031X

https://dx.doi.org/10.1371/journal.pone.0100552

https://dx.doi.org/10.1017/S1355617708080867

https://dx.doi.org/10.1016/S0028-3932(99)00038-X


Dell, G. S. & Chang, F. (2014). The p-chain: Relating sentence production and its

disorders to comprehension and acquisition. Philos Trans R Soc Lond B Biol Sci,

369 (1634), 20120394. doi:10.1098/rstb.2012.0394

Diaz, B., Hintz, F., Kiebel, S. J., & von Kriegstein, K. (2012). Dysfunction of the auditory

thalamus in developmental dyslexia. Proc Natl Acad Sci U S A, 109 (34), 13841–6.

doi:10.1073/pnas.1119828109

D’Mello, A. M. & Gabrieli, J. D. (2018). Cognitive neuroscience of dyslexia. Language,

Speech, and Hearing Services in Schools, 49 (4), 798–809.

Dronkers, N. F., Wilkins, D. P., Van Valin Jr, R. D., Redfern, B. B., & Jaeger, J. J. (2004).

Lesion analysis of the brain areas involved in language comprehension. Cognition,

92 (1-2), 145–177.

Dufor, O., Serniclaes, W., Sprenger-Charolles, L., & Demonet, J. F. (2009). Left premotor

cortex and allophonic speech perception in dyslexia: A pet study. Neuroimage, 46 (1),

241–8. doi:10.1016/j.neuroimage.2009.01.035

Erdmann, B. & Dodge, R. (1898). Psychologische untersuchungen über das lesen auf

experimenteller grundlage. Books on Demand.

Facoetti, A., Corradi, N., Ruffino, M., Gori, S., & Zorzi, M. (2010). Visual spatial attention

and speech segmentation are both impaired in preschoolers at familial risk for

developmental dyslexia. Dyslexia, 16 (3), 226–239. doi:10.1002/dys.413

Fonov, V. S., Evans, A. C., McKinstry, R. C., Almli, C. R., & Collins, D. L. (2009).

Unbiased nonlinear average age-appropriate brain templates from birth to adulthood.

Neuroimage, 47. doi:10.1016/S1053-8119(09)70884-5

Fonov, V., Evans, A. C., Botteron, K., Almli, C. R., McKinstry, R. C., Collins, D. L., &

Brain Development Cooperative, G. (2011). Unbiased average age-appropriate atlases

for pediatric studies. Neuroimage, 54 (1), 313–27.

doi:10.1016/j.neuroimage.2010.07.033

https://dx.doi.org/10.1098/rstb.2012.0394

https://dx.doi.org/10.1073/pnas.1119828109

https://dx.doi.org/10.1016/j.neuroimage.2009.01.035

https://dx.doi.org/10.1002/dys.413

https://dx.doi.org/10.1016/S1053-8119(09)70884-5

https://dx.doi.org/10.1016/j.neuroimage.2010.07.033


Gabay, Y. & Holt, L. L. (2015). Incidental learning of sound categories is impaired in

developmental dyslexia. Cortex, 73, 131–43. doi:10.1016/j.cortex.2015.08.008

Gabay, Y., Thiessen, E. D., & Holt, L. L. (2015). Impaired statistical learning in

developmental dyslexia. J Speech Lang Hear Res, 58 (3), 934–45.

doi:10.1044/2015_JSLHR-L-14-0324

Gabay, Y., Vakil, E., Schiff, R., & Holt, L. L. (2015). Probabilistic category learning in

developmental dyslexia: Evidence from feedback and paired-associate weather

prediction tasks. Neuropsychology, 29 (6), 844–54. doi:10.1037/neu0000194

Hall, A. H. & Tannebaum, R. P. (2013). Test Review: JL Wiederholt & BR Bryant.(2012).

Gray Oral Reading Tests-Fifth Edition (GORT-5). Austin, TX: Pro-Ed. Sage

Publications Sage CA: Los Angeles, CA.

Hancock, R., Pugh, K. R., & Hoeft, F. (2017). Neural noise hypothesis of developmental

dyslexia. Trends Cogn Sci, 21 (6), 434–448. doi:10.1016/j.tics.2017.03.008

Harm, M. W. & Seidenberg, M. S. (1999). Phonology, reading acquisition, and dyslexia:

Insights from connectionist models. Psychological review, 106 (3), 491.

Hawelka, S., Gagl, B., & Wimmer, H. (2010). A dual-route perspective on eye movements

of dyslexic readers. Cognition, 115 (3), 367–79. doi:10.1016/j.cognition.2009.11.004

Hawelka, S., Schuster, S., Gagl, B., & Hutzler, F. (2015). On forward inferences of fast and

slow readers. An eye movement study. Sci Rep, 5, 8432. doi:10.1038/srep08432

Hebb, D. O. (1949). The organization of behavior: A neuropsychological theory. J. Wiley;

Chapman & Hall.

Henderson, J. M., Choi, W., Lowder, M. W., & Ferreira, F. (2016). Language structure in

the brain: A fixation-related fmri study of syntactic surprisal in reading. Neuroimage,

132, 293–300.

Henderson, J. M., Choi, W., Luke, S. G., & Desai, R. H. (2015). Neural correlates of

fixation duration in natural reading: Evidence from fixation-related fmri.

NeuroImage, 119, 390–397.

https://dx.doi.org/10.1016/j.cortex.2015.08.008

https://dx.doi.org/10.1044/2015_JSLHR-L-14-0324

https://dx.doi.org/10.1037/neu0000194

https://dx.doi.org/10.1016/j.tics.2017.03.008

https://dx.doi.org/10.1016/j.cognition.2009.11.004

https://dx.doi.org/10.1038/srep08432


Hinshelwood, J. (1896). A case of dyslexia: A peculiar form of word-blindness. The Lancet,

148 (3821), 1451–1454. doi:10.1016/S0140-6736(01)60603-2

Howard Jr, J. H., Howard, D. V., Japikse, K. C., & Eden, G. F. (2006). Dyslexics are

impaired on implicit higher-order sequence learning, but not on implicit spatial

context learning. Neuropsychologia, 44 (7), 1131–1144.

Hutzler, F. & Wimmer, H. (2004). Eye movements of dyslexic children when reading in a

regular orthography. Brain and Language, 89 (1), 235–242.

doi:10.1016/s0093-934x(03)00401-2

Jossmann, P. B. (1948). Dyslexia. Archives of neurology and psychiatry, 60 (2).

Karlebach, G. & Francks, C. (2015). Lateralization of gene expression in human language

cortex. Cortex, 67, 30–6. doi:10.1016/j.cortex.2015.03.003

Karni, A., Morocz, I. A., Bitan, T., Shaul, S., Kushnir, T., & Breznitz, Z. (2005). An fmri

study of the differential effects of word presentation rates (reading acceleration) on

dyslexic readers’ brain activity patterns. Journal of Neurolinguistics, 18 (2), 197–219.

doi:10.1016/j.jneuroling.2004.11.002

Klein, A., Ghosh, S. S., Bao, F. S., Giard, J., Häme, Y., Stavsky, E., . . . Neto, E. C., et al.

(2017). Mindboggling morphometry of human brains. PLoS Computational Biology,

13 (2), e1005350.

Klein, A. & Tourville, J. (2012). 101 labeled brain images and a consistent human cortical

labeling protocol. Frontiers in Neuroscience, 6.

Krishnan, S., Watkins, K. E., & Bishop, D. V. (2016). Neurobiological basis of language

learning difficulties. Trends Cogn Sci, 20 (9), 701–14. doi:10.1016/j.tics.2016.06.012

Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest package: Tests

in linear mixed effects models. Journal of Statistical Software, 82 (13), 1–26.

doi:10.18637/jss.v082.i13

Lee, S. M., Daniels, M. H., Puig, A., Newgent, R. A., & Nam, S. K. (2008). A data-based

model to predict postsecondary educational attainment of low-socioeconomic-status

https://dx.doi.org/10.1016/S0140-6736(01)60603-2

https://dx.doi.org/10.1016/s0093-934x(03)00401-2

https://dx.doi.org/10.1016/j.cortex.2015.03.003

https://dx.doi.org/10.1016/j.jneuroling.2004.11.002

https://dx.doi.org/10.1016/j.tics.2016.06.012

https://dx.doi.org/10.18637/jss.v082.i13


students. Professional School Counseling, 11 (5), 306–316. Retrieved from

https://www.lib.byu.edu/cgi-bin/remoteauth.pl?url=http://search.ebscohost.com/

login.aspx?direct=true&db=aph&AN=32697994&site=ehost-live&scope=site

Li, X., Morgan, P. S., Ashburner, J., Smith, J., & Rorden, C. (2016). The first step for

neuroimaging data analysis: Dicom to nifti conversion. Journal of Neuroscience

Methods, 264, 47–56.

Lukasova, K., Silva, I. P., & Macedo, E. C. (2016). Impaired oculomotor behavior of

children with developmental dyslexia in antisaccades and predictive saccades tasks.

Front Psychol, 7, 987. doi:10.3389/fpsyg.2016.00987

Luke, S. G. & Christianson, K. (2016). Limits on lexical prediction during reading. Cogn

Psychol, 88, 22–60. doi:10.1016/j.cogpsych.2016.06.002

Lum, J. A., Ullman, M. T., & Conti-Ramsden, G. (2013). Procedural learning is impaired

in dyslexia: Evidence from a meta-analysis of serial reaction time studies. Res Dev

Disabil, 34 (10), 3460–76. doi:10.1016/j.ridd.2013.07.017

Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of dyslexia. Annals of

Dyslexia, 53 (1), 1–14.

MacSweeney, M., Brammer, M. J., Waters, D., & Goswami, U. (2009). Enhanced

activation of the left inferior frontal gyrus in deaf and dyslexic adults during

rhyming. Brain, 132 (Pt 7), 1928–40. doi:10.1093/brain/awp129

Mariën, P., Ackermann, H., Adamaszek, M., Barwood, C. H., Beaton, A., Desmond, J., . . .

Kueper, M., et al. (2014). Consensus paper: Language and the cerebellum: An

ongoing enigma. The Cerebellum, 13 (3), 386–410.

Martin, A., Schurz, M., Kronbichler, M., & Richlan, F. (2015). Reading in the brain of

children and adults: A meta-analysis of 40 functional magnetic resonance imaging

studies. Human Brain Mapping, 36 (5), 1963–1981.

Marvel, C. L. & Desmond, J. E. (2010). Functional topography of the cerebellum in verbal

working memory. Neuropsychology Review, 20 (3), 271–279.

https://www.lib.byu.edu/cgi-bin/remoteauth.pl?url=http://search.ebscohost.com/login.aspx?direct=true&db=aph&AN=32697994&site=ehost-live&scope=site

https://www.lib.byu.edu/cgi-bin/remoteauth.pl?url=http://search.ebscohost.com/login.aspx?direct=true&db=aph&AN=32697994&site=ehost-live&scope=site

https://dx.doi.org/10.3389/fpsyg.2016.00987

https://dx.doi.org/10.1016/j.cogpsych.2016.06.002

https://dx.doi.org/10.1016/j.ridd.2013.07.017

https://dx.doi.org/10.1093/brain/awp129


McConkie, G. W., Zola, D., Grimes, J., Kerr, P. W., Bryant, N. R., & Wolff, P. M. (1991).

Children’s eye movements during reading. Vision and Visual Dyslexia, 13.

McCrimmon, A. W. & Smith, A. D. (2013). Review of the wechsler abbreviated scale of

intelligence, (wasi-ii). Sage Publications Sage CA: Los Angeles, CA.

McCrory, E. J., Mechelli, A., Frith, U., & Price, C. J. (2005). More than words: A common

neural basis for reading and naming deficits in developmental dyslexia? Brain,

128 (Pt 2), 261–7. doi:10.1093/brain/awh340

McDowell, J. E., Dyckman, K. A., Austin, B. P., & Clementz, B. A. (2008).

Neurophysiology and neuroanatomy of reflexive and volitional saccades: Evidence

from studies of humans. Brain and Cognition, 68 (3), 255–270.

McLaughlin, M. J., Speirs, K. E., & Shenassa, E. D. (2014). Reading disability and adult

attained education and income: Evidence from a 30-year longitudinal study of a

population-based sample. J Learn Disabil, 47 (4), 374–86.

doi:10.1177/0022219412458323

Meng, H., Smith, S. D., Hager, K., Held, M., Liu, J., Olson, R. K., . . . O’Reilly-Pol, T.

(2005). Dcdc2 is associated with reading disability and modulates neuronal

development in the brain. Proceedings of the National Academy of Sciences of the

United States of America, 102 (47), 17053–17058. doi:10.1073/pnas.0508591102

Morgan, W. P. (1896). A case of congenital word blindness. British Medical Journal,

2 (1871), 1378.

Müri, R. M. & Nyffeler, T. (2008). Neurophysiology and neuroanatomy of reflexive and

volitional saccades as revealed by lesion studies with neurological patients and

transcranial magnetic stimulation (tms). Brain and Cognition, 68 (3), 284–292.

Nicolson, R. I., Fawcett, A. J., & Dean, P. (2001). Developmental dyslexia: The cerebellar

deficit hypothesis. Trends in Neurosciences, 24 (9), 508–511.

https://dx.doi.org/10.1093/brain/awh340

https://dx.doi.org/10.1177/0022219412458323

https://dx.doi.org/10.1073/pnas.0508591102


Nigro, L., Jiménez-Fernández, G., Simpson, I. C., & Defior, S. (2016). Implicit learning of

non-linguistic and linguistic regularities in children with dyslexia. Annals of Dyslexia,

66 (2), 202–218.

Noppeney, U., Price, C. J., Duncan, J. S., & Koepp, M. J. (2005). Reading skills after left

anterior temporal lobe resection: An fmri study. Brain, 128 (6), 1377–1385.

Norton, E. S., Beach, S. D., & Gabrieli, J. D. (2015). Neurobiology of dyslexia. Curr Opin

Neurobiol, 30, 73–8. doi:10.1016/j.conb.2014.09.007

Orton, S. T. (1925). Word-blindness in school children. Archives of Neurology &

Psychiatry, 14 (5), 581–615.

Orton, S. T. (1928). Specific reading disability–strephosymbolia. Journal of the American

Medical Association, 90 (14), 1095–1099.

Ozernov-Palchik, O. et al. (n.d.). Longitudinal stability of pre-reading skill profiles of

kindergarten children. Developmental Science, 20.

Paracchini D Phil, S., Steer, C. D., Buckingham, L.-L., Morris, A. P., Ring, S.,

Scerri D Phil, T., . . . Golding, J., et al. (2008). Association of the kiaa0319 dyslexia

susceptibility gene with reading skills in the general population. American Journal of

Psychiatry, 165 (12), 1576–1584.

Pavlidou, E., Kelly, M. L., & Williams, J. M. (2010). Do children with developmental

dyslexia have impairments in implicit learning? Dyslexia, 16, 143–161.

doi:10.1002/dys.400

Pecini, C., Biagi, L., Brizzolara, D., Cipriani, P., Di Lieto, M. C., Guzzetta, A., . . .

Chilosi, A. M. (2011). How many functional brains in developmental dyslexia? When

the history of language delay makes the difference. Cognitive and Behavioral

Neurology, 24 (2), 85–92.

Peyrin, C., Lallier, M., Demonet, J. F., Pernet, C., Baciu, M., Le Bas, J. F., & Valdois, S.

(2012). Neural dissociation of phonological and visual attention span disorders in

https://dx.doi.org/10.1016/j.conb.2014.09.007

https://dx.doi.org/10.1002/dys.400


developmental dyslexia: Fmri evidence from two case reports. Brain Lang, 120 (3),

381–94. doi:10.1016/j.bandl.2011.12.015

Poldrack, R. A. (2007). Region of interest analysis for fmri. Social Cognitive and Affective

Neuroscience, 2 (1), 67–70.

Price, C. J. (2012). A review and synthesis of the first 20 years of pet and fmri studies of

heard speech, spoken language and reading. Neuroimage, 62 (2), 816–847.

Pugh, K. R. [K. R.], Frost, S. J., Rothman, D. L., Hoeft, F., Del Tufo, S. N., Mason, G. F.,

. . . Fulbright, R. K. (2014). Glutamate and choline levels predict individual

differences in reading ability in emergent readers. J Neurosci, 34 (11), 4082–9.

doi:10.1523/JNEUROSCI.3907-13.2014

R Core Team. (2016). R: A language and environment for statistical computing. R

Foundation for Statistical Computing. Vienna, Austria. Retrieved from

https://www.R-project.org/

Ramus, F., Altarelli, I., Jednorog, K., Zhao, J., & Scotto di Covella, L. (2017).

Neuroanatomy of developmental dyslexia: Pitfalls and promise. Neurosci Biobehav

Rev. doi:10.1016/j.neubiorev.2017.08.001

Rayner, K., Foorman, B. R., & Perfetti, C. A. (2001). How psychological science informs

the teaching of reading. Psychological Science in the Public Interest, 2 (2), 31–74.

doi:10.1111/1529-1006.00004

Rello, L. & Baeza-Yates, R. (2017). How to present more readable text for people with

dyslexia. Universal Access in the Information Society, 16 (1), 29–49.

Reuters, T. (2012). Web of science. Retrieved from https://login.webofknowledge.com/

Rodgers, B. (1983). The identification and prevalence of specific reading retardation.

British Journal of Educational Psychology, 53 (3), 369–373.

Sakurai, Y. (2004). Varieties of alexia from fusiform, posterior inferior temporal and

posterior occipital gyrus lesions. Behavioural Neurology, 15 (1-2), 35–50.

https://dx.doi.org/10.1016/j.bandl.2011.12.015

https://dx.doi.org/10.1523/JNEUROSCI.3907-13.2014

https://www.R-project.org/

https://dx.doi.org/10.1016/j.neubiorev.2017.08.001

https://dx.doi.org/10.1111/1529-1006.00004

https://login.webofknowledge.com/


Schmalz, X., Altoè, G., & Mulatti, C. (2017). Statistical learning and dyslexia: A

systematic review. Annals of Dyslexia, 67 (2), 147–162.

Shaywitz, B. A., Fletcher, J. M., Holahan, J. M., & Shaywitz, S. E. (1992). Discrepancy

compared to low achievement definitions of reading disability: Results from the

connecticut longitudinal study. Journal of Learning Disabilities, 25 (10), 639–648.

Shaywitz, B. A., Shaywitz, S. E. [Sally E], Pugh, K. R. [Kenneth R], Mencl, W. E.,

Fulbright, R. K., Skudlarski, P., . . . Lyon, G. R., et al. (2002). Disruption of posterior

brain systems for reading in children with developmental dyslexia. Biological

Psychiatry, 52 (2), 101–110.

Shaywitz, S. E. [S. E.], Escobar, M. D., Shaywitz, B. A., Fletcher, J. M., & Makuch, R.

(1992). Evidence that dyslexia may represent the lower tail of a normal distribution

of reading ability. The New England Journal of Medicine, 326 (3), 145–150.

Siegel, L. S. (2006). Perspectives on dyslexia. Pediatrics & Child Health, 11 (9), 581–587.

Sigurdardottir, H. M., Danielsdottir, H. B., Gudmundsdottir, M., Hjartarson, K. H.,

Thorarinsdottir, E. A., & Kristjansson, A. (2017). Problems with visual statistical

learning in developmental dyslexia. Scientific Reports, 7 (1), 606.

doi:10.1038/s41598-017-00554-5

Sinclair, A. H. H. (1948). Developmental aphasia: Also known as congenital word-blindness

and sometimes referred to as alexia or dyslexia. British Journal of Opthalmology,

32 (9), 522–531. doi:10.1136/bjo.32.9.522

Snow, C. E., Burns, M. S., & Griffin, P. (1998). Preventing reading difficulties in young

children. Washington D.C.: National Academy of Sciences – National Research

Council, Commission on Behavioral, Social Sciences, and Education.

Stein, J. (2019). The current status of the magnocellular theory of developmental dyslexia.

Neuropsychologia, 130, 66–77.

Steinbrink, C., Groth, K., Lachmann, T., & Riecker, A. (2012). Neural correlates of

temporal auditory processing in developmental dyslexia during German vowel length

https://dx.doi.org/10.1038/s41598-017-00554-5

https://dx.doi.org/10.1136/bjo.32.9.522


discrimination: an fMRI study. Brain Lang, 121 (1), 1–11.

doi:10.1016/j.bandl.2011.12.003

Stoodley, C. J. & Stein, J. F. (2011). The cerebellum and dyslexia. Cortex: A Journal

Devoted to the Study of the Nervous System and Behavior, 47.

Stuphorn, V. (2015). The role of supplementary eye field in goal-directed behavior. Journal

of Physiology-Paris, 109 (1-3), 118–128.

Szalkowski, C. E., Fiondella, C. F., Truong, D. T., Rosen, G. D., LoTurco, J. J., &

Fitch, R. H. (2013). The effects of kiaa0319 knockdown on cortical and subcortical

anatomy in male rats. Int J Dev Neurosci, 31 (2), 116–22.

doi:10.1016/j.ijdevneu.2012.11.008

Tan, L. H., Laird, A. R., Li, K., & Fox, P. T. (2005). Neuroanatomical correlates of

phonological processing of chinese characters and alphabetic words: A meta-analysis.

Human Brain Mapping, 25 (1), 83–91.

Truong, D. T., Che, A., Rendall, A. R., Szalkowski, C. E., LoTurco, J. J.,

Galaburda, A. M., & Holly Fitch, R. (2014). Mutation of dcdc2 in mice leads to

impairments in auditory processing and memory ability. Genes Brain Behav, 13 (8),

802–11. doi:10.1111/gbb.12170

Tustison, N. J., Avants, B. B., Cook, P. A., Zheng, Y., Egan, A., Yushkevich, P. A., &

Gee, J. C. (2010). N4ITK: improved N3 bias correction. IEEE Transactions on

Medical Imaging, 29 (6), 1310–1320.

Tustison, N. J., Cook, P. A., Klein, A., Song, G., Das, S. R., Duda, J. T., . . . Gee, J. C.,

et al. (2014). Large-scale evaluation of ants and freesurfer cortical thickness

measurements. Neuroimage, 99, 166–179.

Ullman, M. T. & Pierpont, E. I. (2005). Specific language impairment is not specific to

language: The procedural deficit hypothesis. Cortex, 41 (3), 399–433.

doi:10.1016/s0010-9452(08)70276-4

https://dx.doi.org/10.1016/j.bandl.2011.12.003

https://dx.doi.org/10.1016/j.ijdevneu.2012.11.008

https://dx.doi.org/10.1111/gbb.12170

https://dx.doi.org/10.1016/s0010-9452(08)70276-4


Valdois, S., Lassus-Sangosse, D., Lallier, M., Moreaud, O., & Pisella, L. (2019). What

bilateral damage of the superior parietal lobes tells us about visual attention

disorders in developmental dyslexia. Neuropsychologia, 130, 78–91.

Vidyasagar, T. R. & Pammer, K. (2010). Dyslexia: A deficit in visuo-spatial attention, not

in phonological processing. Trends in Cognitive Sciences, 14 (2), 57–63.

Wagner, R., Torgesen, J., Rashotte, C., Pearson, N. A., Wiig, E., & Secord, W. (1999).

Ctopp-2: Comprehensive test of phonological processing–second edition. Pro-ed

Austin, TX.

Wang, H., Suh, J. W., Das, S. R., Pluta, J. B., Craige, C., & Yushkevich, P. A. (2013).

Multi-atlas segmentation with joint label fusion. IEEE Transactions on Pattern

Analysis and Machine Intelligence, 35 (3), 611–623.

Weber, K., Lau, E. F., Stillerman, B., & Kuperberg, G. R. (2016). The yin and the yang of

prediction: An fmri study of semantic predictive processing. PLoS One, 11 (3).

White, S., Milne, E., Rosen, S., Hansen, P., Swettenham, J., Frith, U., & Ramus, F. (2006).

The role of sensorimotor impairments in dyslexia: A multiple case study of dyslexic

children. Developmental Science, 9 (3), 237–269. doi:10.1111/j.1467-7687.2006.00483.x

Wilcke, A., Weissfuss, J., Kirsten, H., Wolfram, G., Boltze, J., & Ahnert, P. (2009). The

role of gene DCDC2 in German dyslexics. Ann Dyslexia, 59 (1), 1–11.

doi:10.1007/s11881-008-0020-7

Willems, R. M., Frank, S. L., Nijhof, A. D., Hagoort, P., & Van den Bosch, A. (2016).

Prediction during natural language comprehension. Cerebral Cortex, 26 (6),

2506–2516.

Wilson, S. M., DeMarco, A. T., Henry, M. L., Gesierich, B., Babiak, M., Mandelli, M. L.,

. . . Gorno-Tempini, M. L. (2014). What role does the anterior temporal lobe play in

sentence-level processing? Neural correlates of syntactic processing in semantic

variant primary progressive aphasia. Journal of Cognitive Neuroscience, 26 (5),

970–985.

https://dx.doi.org/10.1111/j.1467-7687.2006.00483.x

https://dx.doi.org/10.1007/s11881-008-0020-7


Wolf, M. & Bowers, P. G. (1999). The double-deficit hypothesis for developmental

dyslexias. Journal of Educational Psychology, 91 (3), 415–438.

doi:10.1037/0022-0663.91.3.415

World Health Organization. (1992). The ICD-10 classification of mental and behavioural

disorders: Clinical descriptions and diagnostic guidelines. World Health Organization.

World Health Organization. (1993). The ICD-10 classification of mental and behavioural

disorders: Diagnostic criteria for research. World Health Organization.

Yarkoni, T., Poldrack, R. A., Nichols, T. E., Van Essen, D. C., & Wager, T. D. (2011).

Large-scale automated synthesis of human functional neuroimaging data. Nat

Methods, 8 (8), 665–70. doi:10.1038/nmeth.1635

Yushkevich, P. A., Piven, J., Hazlett, H. C., Smith, R. G., Ho, S., Gee, J. C., & Gerig, G.

(2006). User-guided 3d active contour segmentation of anatomical structures:

Significantly improved efficiency and reliability. Neuroimage, 31 (3), 1116–1128.

Zoubrinetzky, R., Bielle, F., & Valdois, S. (2014). New insights on developmental dyslexia

subtypes: Heterogeneity of mixed reading profiles. PLoS One, 9 (6).

https://dx.doi.org/10.1037/0022-0663.91.3.415

https://dx.doi.org/10.1038/nmeth.1635

Dyslexia Beyond the Word: An Ecological Study of Specific ...

Documents

Transcript of Dyslexia Beyond the Word: An Ecological Study of Specific ...