N

Dw

Va

b

a

ARRAA

KDCAECWR

1

tMcBPnH2M(D

bo

FHf

0d

ARTICLE IN PRESSG ModelSY-3275; No. of Pages 10

Neuropsychologia xxx (2009) xxx–xxx

Contents lists available at ScienceDirect

Neuropsychologia

journa l homepage: www.e lsev ier .com/ locate /neuropsychologia

ifferent behavioral and eye movement patterns of dyslexic readers with andithout attentional deficits during single word reading�

erena Thalera,∗, Karolina Urtona, Angela Heinea, Stefan Hawelkab, Verena Engla, Arthur M. Jacobsa

Freie Universität Berlin, GermanyUniversität Salzburg, Austria

r t i c l e i n f o

rticle history:eceived 6 August 2008eceived in revised form 7 April 2009ccepted 13 April 2009vailable online xxx

eywords:yslexiaomorbidityttentional problems

a b s t r a c t

Comorbidity of learning disabilities is a very common phenomenon which is intensively studied ingenetics, neuropsychology, prevalence studies and causal deficit research. In studies on the behavioralmanifestation of learning disabilities, however, comorbidity is often neglected. In the present study, wesystematically examined the reading behavior of German-speaking children with dyslexia, of childrenwith attentional problems, of children with comorbid dyslexia and attentional problems and of normallydeveloping children by measuring their reading accuracy, naming latencies and eye movement patternsduring single word reading. We manipulated word difficulty by contrasting (1) short vs. long wordswith (2) either low or high sublexical complexity (indexed by consonant cluster density). Children withdyslexia only (DYS) showed the expected reading fluency impairment of poor readers in regular orthogra-

ye movementsonsonant clusterord length

egular orthography

phies but no accuracy problem. In contrast, comorbid children (DYS + AD) had significantly higher errorrates than all other groups, but less of a problem with reading fluency than DYS. Concurrently recordedeye movement measures revealed that DYS made the highest number of fixations, but exhibited shortermean single fixations than DYS + AD. Word length had the strongest effect on dyslexic children, whereasconsonant cluster density affected all groups equally. Theoretical implications of these behavioral andeye movement patterns are discussed and the necessity for controlling for comorbid attentional deficits

efici

in children with reading d

. Introduction

In the research field of learning disabilities, it is undisputablehat comorbidity is a very common phenomenon (e.g., Gross-Tsur,

anor, & Shalev, 1996; Lewis, Hitch, & Walker, 1994). Typicallyomorbidity of learning disorders is studied in genetics (e.g.,arr et al., 2008; Friedman, Chhabildas, Budhiraja, Willcutt, &ennington, 2003; Willcutt, Pennington, Olson, & DeFries, 2007),europsychology (e.g., Burd, Freeman, Klug, & Kerbeshian, 2005;endriksen et al., 2007), prevalence studies (e.g., Brook & Boaz,

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

005; Capano, Minden, Chen, Schachar, & Ickowicz, 2008; Manor,edad, Zamishlani, & Vurmbrand, 2008) and causal deficit research

e.g., Brookes, Nicolson, & Fawcett, 2007; Crawford & Dewey, 2008;onfrancesco, Mugnaini, & Dell’Uomo, 2005; Haslum & Miles,

� Verena Thaler’s, Angela Heine’s and Verena Engl’s contribution was supportedy a grant (grant number: PLI3029) to A.M. Jacobs from the German Federal Ministryf Education and Research (BMBF).∗ Corresponding author at: Allgemeine und Neurokognitive Psychologie,

achbereich Erziehungswissenschaften und Psychologie, Freie Universität Berlin,abelschwerdter Allee 45, 14195 Berlin, Germany. Tel.: +49 30 838 55626;

ax: +49 30 838 55620.E-mail address: verena.thaler@fu-berlin.de (V. Thaler).

028-3932/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.oi:10.1016/j.neuropsychologia.2009.04.006

ts is highlighted.© 2009 Elsevier Ltd. All rights reserved.

2007). In many studies (e.g., Ackerman & Dykman, 1995; Jongmans,Bouwien, Smits-Engelsman, & Schoemaker, 2003; Landerl, Bevan,& Butterworth, 2004; Purvis & Tannock, 1997; Shalev, Gross-Tsur,& Manor, 1997) it was shown that comorbidity mostly aggravatesthe symptoms and problems of children with a learning disability;Crawford and Dewey (2008), for example, concluded that deficitsin visual memory skills seem to be specific for children withdevelopmental coordination disorder (DCD) and co-occurring read-ing disabilities and/or ADHD. Brook and Boaz (2005) found thathigh school pupils with ADHD and comorbid learning disabilities(dyscalculia, dysgraphia, social science difficulties, fine motor skilldifficulties and spatial adaptation problems) had lower academicachievement. These studies exemplify the importance of control-ling for comorbid deficits; only by ensuring that children withdyscalculia or dyslexia do not suffer from additional comorbid dis-orders, it will be possible to identify the typical patterns or possiblecauses of a specific disorder.

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

1.1. Comorbidity of dyslexia and attentional deficits

Capano et al. (2008) showed that especially for reading disabil-ities the association with ADHD is high. While 18% of their ADHDpopulation had a comorbid mathematical disability, about 26% had

INN

2 ycholo

aaipiVbcwbpf

1

dIGrieHKtCteopttnTtiods

1

iftrtpWdotbtswsDeGlEP

w

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

comorbid reading disability. Yu, Buka, McCormick, Fitzmaurice,nd Indurkhya (2006) further demonstrated that verbal learn-ng disabilities are much more likely associated with behavioralroblems (i.e., externalizing and internalizing behavioral disabil-

ties) than other non-verbal learning disabilities. In their review,ellutino, Fletcher, Snowling, and Scanlon (2004) reported a comor-idity rate of attentional problems between 30% and 70% in dyslexichildren. In light of this figure, it seems essential to control childrenith reading problems for comorbid attentional problems. We will

e able to isolate the characteristic deficits underlying the reading-roblems, only when we ensure that the observed effects resultrom a pure dyslexic group without comorbid attentional problems.

.2. Attentional deficit/hyperactivity disorder (ADHD)

ADHD is a developmental disorder characterized by increasedistractibility, inattention and errors caused by carelessness (DSM-

V-TR German Version: Saß, Wittchen, & Zaudig, 2003; ICD-10erman version, Dilling, Mombour, & Schmidt, 2004). Error and

eaction time patterns of children with attentional deficits are typ-cally measured with continuous performance tasks (e.g., Epsteint al., 2003; Newcorn et al., 2001), flanker tasks (e.g., van Meel,eslenfeld, Oosterlaan, & Sergeant, 2007), go/no go tasks (e.g.,oschak, Kunert, Derichs, Weniger, & Irle, 2003) or visual searchasks (e.g., Wilding, 2003; Wilding, Pankhania, & Williams, 2007).hildren with attentional deficits show similar reaction time pat-erns compared to unimpaired subjects, but drastically increasedrror rates. In a recent ERP study the error-related negativity (ERN)f ADHD children was measured with a modified Eriksen flankeraradigm (van Meel et al., 2007). ADHD children made more errorshan controls especially under time pressure. ERN analyses revealedhat, although the behaviorally measured post-error slowing wasormal in children with ADHD, the ERN amplitude was diminished.he authors suggested that ADHD is correlated with a disruption inhe brain’s error checking system which leads to a failure of employ-ng adequate cognitive control in speeded reaction tasks. This lackf cognitive control further prevents children with ADHD from pre-icting the likelihood of an error and to adapt their performancetrategy accordingly.

.3. Developmental dyslexia

For many decades, developmental dyslexia research was dom-nated from labs in English speaking countries (see Share, 2008or a recent critique). As a consequence, it has long been assumedhat dyslexia is typically characterized by a high error rate duringeading. Studies in more regular orthographies, however, revealedhat the symptoms of dyslexia are highly correlated with the trans-arency of certain orthographies. In one of his pioneering studies,immer (1993) analyzed the reading behavior of German speaking

yslexic children in Grade 2, 3 and 4. He found that, independentf reading material (high-frequency words, pseudowords or con-inuous text), dyslexic children exhibited high reading accuracy,ut extremely slow reading fluency. He, therefore, concluded thathe typical problem of German speaking poor readers is a perva-ive reading speed deficit. There is now an abundance of studieshich support the critical role of transparency for defining dyslexic

ymptoms; whereas dyslexia in deep orthographies (e.g., English,anish) is mainly characterized by high error rates, in transpar-nt orthographies (e.g., Italian, Spanish, Finish, Norwegian, Dutch,erman) dyslexia is defined by significantly prolonged reading

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

atencies, whereas accuracy is relatively preserved (Seymour, Aro,rskine, & COST Action A8 Network, 2003; Wimmer, 1993; Ziegler,erry, Ma-Wyatt, Ladner, & Schulte-Körne, 2003).

Many studies on reading and dyslexia revealed that differentord characteristics differentially affected good and poor read-

PRESSgia xxx (2009) xxx–xxx

ers. Examples are the word frequency effect (e.g., Barca, Burani, DiFilippo, & Zoccolotti, 2006; Katz et al., 2005; Stenneken, Conrad, &Hutzler, 2005), the regularity effect (e.g., Coltheart & Rastle, 1994;Katz et al., 2005; Visser & Besner, 2001), and differences in non-word reading (e.g., Svensson & Jacobson, 2005; Wimmer, 1996; forreview see Rack, Snowling, & Olson, 1992). Other well-establishedeffects of dyslexic reading are the word length effect and the large-unit effect (Di Filippo, De Luca, Judica, Spinelli, & Zoccolotti, 2006;Juphard, Carbonnel, & Valdois, 2004; Landerl, Wimmer, & Frith,1997; Martens & de Jong, 2006; Treiman, Goswami, & Bruck, 1990;Ziegler & Goswami, 2005; Ziegler & Perry, 1998; Ziegler et al., 2003;Zoccolotti et al., 2005).

1.3.1. The word length effectThe word length effect refers to the observation that the reading

time of poor readers increases quasi-linearly with every additionalletter, regardless of whether they read words or pseudowords. Incontrast, proficient readers show hardly any length effects for wordsand only a moderate effect for pseudowords (Ferrand & New, 2003;Valdois et al., 2006; Weekes, 1997). This difference between normaland poor readers indicates that the latter typically rely on a serialdecoding strategy for word recognition, whereas normal readersread whole words or large sublexical units in a parallel fashion (DiFilippo et al., 2006; Juphard et al., 2004; Landerl et al., 1997; Martens& de Jong, 2006; Ziegler et al., 2003; Zoccolotti et al., 2005). In tryingto better understand the reading patterns of poor readers, Zoccolottiand collaborators (De Luca, Borrelli, Judica, Spinelli, & Zoccolotti,2002; De Luca, Di Pace, Judica, Spinelli, & Zoccolotti, 1999) com-plemented accuracy and reaction time measures by examining theeye movements of Italian dyslexic and normal readers. Crucially,they demonstrated that the eye movements of dyslexic childrenonly deviated from normal readers when the task required read-ing, whereas there was no group differences in non-reading controltasks. Hutzler, Kronbichler, Jacobs, and Wimmer (2006) expandedthis finding. They reported a completely normal performance ofpoor readers in a visual search task. For task which requires reading,De Luca et al. (1999) showed that dyslexics (10–17 years) typi-cally exhibit prolonged single fixation durations (290 ms, controls:230 ms) and a doubled number of fixations per word. Furthermore,for the dyslexic readers, number of fixations significantly increasedwith increasing word length which was not the case for unimpairedreaders. Eye movement studies in different regular orthographiesunderlined the hypothesized serial nature of the reading processin reading-disabled children (De Luca et al., 2002, 1999; Hutzler &Wimmer, 2004; MacKeben et al., 2004). In a recent study from ourlab (Thaler, Heine, Engl, & Jacobs, in preparation), 50 disabled read-ers and 45 normal readers were tested during single word reading.All participants attended Grade 3–5 and had a mean age of 8.5 years.Children had to read single words with 3–12 letters on a computerscreen while their eye movements were recorded. For normal read-ers, the average number of fixations increased from three fixationsfor words with three letters to four fixations for words with sixletters and six fixations for words with twelve letters. Their read-ing time increased approximately 65 ms for each additional letter.In marked contrast, the reading time of reading-disabled childrenincreased approximately 350 ms for each additional letter. On aver-age, words with three letters were fixated three times, words withsix letters seven times and words with twelve letters were fix-ated twelve times. These results suggest that disabled readers fixateevery letter of a word at least once while normally developing chil-dren between Grade 3 and 5 start to read words by utilizing larger

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

sublexical units. Thus, the notion of a serial decoding strategy ofpoor readers—at least in transparent orthographies—is generallysupported. However, whether or not reading in regular orthogra-phies is strictly letter-based or whether the grain size (Ziegler &Goswami, 2005) is influenced by orthography and/or linguistic fac-

INN

ycholo

tDZ

1

rrulodtlldZdrrpttaeuaM(SC&W&cwFcociirt‘nrTtfddwmoseri

1

sfd

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

ors is still a matter of debate (Caravolas, 2006; de Jong, 2006;urgunoglu, 2006; Paulesu, 2006; Pugh, 2006; Wimmer, 2006;iegler & Goswami, 2005, 2006).

.3.2. The large-unit effectZiegler et al. (2003) tested the hypothesis that English-speaking

eading-disabled children in comparison to German-speakingeading-disabled children are able to process larger orthographicnits efficiently. Body neighborhood was used as a marker of the

arge-unit effect. Body neighbors are words which share the samerthographic rime such as might, fight and right. English speakingyslexic children showed a facilitatory Body-N effect, but comparedo control children the reduction of the length effect by extractingarger orthographic units was diminished. No such reduction of theength effect was found for German-speaking children (reading-isabled as well as normally reading children). The observation ofiegler et al. thus confirmed the “Grain Size Theory” of readingevelopment which states that “German children (good and pooreaders) rely more or less exclusively on highly efficient sublexicalecoding procedures at small grain sizes” (Ziegler & Goswami, 2005,. 13). As mentioned above, it is still a matter of debate whetherhis strong version of the grain size assumption—claiming that inransparent orthographies mostly phoneme-grapheme decoding ispplied for word recognition—really holds true. There is empiricalvidence that during reading in regular orthographies sublexicalnits other (i.e., larger) than grapheme-phoneme correspondencesre applied (e.g., Arduino & Burani, 2004; Elbro, 2005; Martensen,aris, & Dijkstra, 2000). Different sublexical units such as syllables

Conrad, Grainger, & Jacobs, 2007; Ferrand & New, 2003; Hofmann,tenneken, Conrad, & Jacobs, 2007; Hutzler et al., 2004; Hutzler,onrad, & Jacobs, 2005; Rouibah & Taft, 2001), bigrams (Doignon

Zagar, 2005; Grainger & Jacobs, 1993; Hofmann et al., 2007;hiteley & Walker, 1997; Whitney, 2001), or morphemes (CarlisleKatz, 2006; Hyönä & Pollatsek, 1998; Taft & Forster, 1975) are dis-

ussed as likely candidates. Recently, effects of consonant clustersere also scrutinized (Martens & de Jong, 2006; Ziegler et al., 2003).

or example, Martens and de Jong (2006, p. 147) state in their dis-ussion: “Interestingly, most of the difference in the effect of lengthn reading speed between dyslexics and their normal reading peersould be ascribed to a larger increase in response times for the dyslex-cs from 3 to 4 letters. Although items of different length were matchedn consonant–vowel structure as much as possible, the increase inesponse time from 3 to 4 letters may have been caused (partly) byhe much higher frequency of occurrence of consonant clusters (e.g.,gr..’ and ‘..ms’) in items consisting of 4 letters (and more).” Unfortu-ately, empirical research on the effect of consonant clusters is stillare (Brand, Groux, Puijalon, & Rey, 2007; Bruck & Treiman, 1990;reiman, 1991). Thaler et al. (in preparation) found that words ofhe same length, syllable frequency, word frequency, and bigramrequency but high consonant cluster density elicited longer gazeurations, higher number of fixations and longer mean fixationurations compared to words with a low cluster density. This effectas particularly strong for poor readers, but also appeared for nor-al readers. Thus, although dyslexic readers (at least in regular

rthographies) seem to rely on a serial decoding strategy, largerublexical units—in this case consonant clusters—exert an influ-nce and such differential effects of sublexical units on single wordeading between normal and dyslexic readers will provide furthernsights into the specific problems of dyslexic readers.

.4. Controlling for comorbidity

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

As mentioned above, in order to ensure that dyslexia-specificymptoms and distal causes are discovered, it is important to care-ully diagnose whether participants of a study are truly (i.e., solely)yslexic (according to internationally comparable criteria) and do

PRESSgia xxx (2009) xxx–xxx 3

not suffer from a additional comorbid disorders which potentiallyaffects (i.e., alters or/and aggravates) their reading deficit. We willnow briefly describe why controlling for comorbidity is impor-tant. Recently, the comorbidity of reading disability and ADHDwas shown to be especially critical in revealing the defensibilityof a causal theory—the cerebellar deficit hypotheses. The Sheffieldgroup (e.g. Brookes et al., 2007; Needle, Fawcett, & Nicolson, 2006;Nicolson & Fawcett, 1990, 1995, 2007) proposed that difficultiesin learning to read are caused by abnormalities in the (func-tioning of the) cerebellum. Cerebellar problems lead to a generaldeficit in automatizing repeatedly performed behavioral or cogni-tive sequences. In their studies, the automatization deficit was oftentested with a dual-task paradigm—that is, concurrently balancingand counting. Wimmer and colleagues (Raberger & Wimmer, 1999,2003; Wimmer, Mayringer, & Raberger, 1999) addressed the ques-tion whether balancing problems of the reading-deficit group (RD)mainly existed because of comorbidity with ADHD. Their studiesindeed confirmed that poor balancing was associated with ADHD,but not with reading deficits. Other studies confirmed this con-found of ADHD in the observed balancing problems in RD (Ramus,Pidgeon, & Frith, 2003; Van Daal & van der Leij, 1999), whereasrecent studies of the Sheffield group again found (at least mild)cerebellar deficits of dyslexic children even when they checked forcomorbid ADHD with balancing tasks (Needle et al., 2006) and othermeasures of cerebellar deficits (e.g., prism adaptation; Brookes etal., 2007). Apparently, whether cerebellar deficits play a role indyslexia seems far from closed.

The ongoing debate, whether cerebellar deficits are the cause of(or contribute to) the dyslexic reading deficit, clearly demonstratesthe importance of controlling for comorbidity. For studying clini-cal populations one needs a firm and clear-cut assessment (1) ofthe specific deficit one is interested in (e.g., dyslexia, dyscalculia orADHD) and (2) of possible comorbid problems of the participants.This is not only important for the validation of causal hypothesesunder investigation, but for every aspect of a specific deficit. Aretypically reported effects influenced by comorbidity as well? Weargue, that every study which, for example, investigates aspects ofdyslexia, has to control for comorbidity to ensure that the observedeffects (e.g., the word length effect) are truly a dyslexic phe-nomenon and not (to some extant) caused by a comorbid disorder.

1.5. The current study

As aforementioned, developmental dyslexia in regular orthogra-phies is primarily characterized by extremely slow reading speed,whereas accuracy is relatively preserved (de Jong, 2003; Landerl,2003; Wimmer, 1993). Nonetheless, there are some studies that didreveal accuracy problems of dyslexic children in regular orthogra-phies (Klicpera & Schabmann, 1993; Landerl et al., 1997; Nergard-Nilsen, 2005). One tentative explanation for the appearance ofdeficient reading accuracy in some studies in transparent orthogra-phies is that these studies included dyslexic children with comorbid(but unidentified) attentional problems. Klicpera and Schabmann(1993), for example, showed that hyperactivity was a stable predic-tor of reading difficulties throughout school (first to eighth grade).The common pattern of children with attentional problems—thatis, high error rates caused by carelessness—could account for theimpaired reading accuracy of children in these studies.

In the present study, we aim at investigating systematicallythe reading behavior of children with reading-problems-only(DYS), children with attentional-problems-only (AD), children

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

with comorbid-reading-and-attentional-problems (DYS + AD) andunimpaired control children (CG). All children were tested withstandardized reading and attentional tests. Subsequently, all groupshad to read single words on a computer screen. Eye movementswere simultaneously recorded. The words were manipulated in

IN PRESSN

4 ychologia xxx (2009) xxx–xxx

tc(sdsibeaa

2

2

p1asMTa

pdhwm9rp

sd[Gga“sfcttpciT(p

2

bdwco

TMmd(

AIRE

Table 2Mean number of letters, mean proportion of consonants in a cluster, number ofsyllables and mean word frequency (1 per million) for the four word categories(short-CVC, short-CCC, long-CVC, long-CCC).

Short-CVC Short-CCC Long-CVC Long-CCC

M SD M SD M SD M SD

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

heir length and consonant cluster density. We hypothesized thathildren with attentional problems are affected by word lengthbecause of their inattention), but not by cluster density. The effecthould be particularly expressed in higher error rates. Children withyslexia should be affected by both word length and cluster den-ity. In this group reading time should abnormally increase withtem length and consonant cluster density, but error rate shoulde about the same for all word categories. Comorbid children arexpected to exhibit both higher reading times and higher error ratesnd these effects are expected to be modulated by both word lengthnd cluster density.

. Methods

.1. Participants

Participating children were selected from a large sample of 2455 pupils from 13rimary schools in Berlin. They all attended Grade 2–5, their age ranged from 85 to40 months (M = 112; SD = 11). All children had normal or corrected to normal visualcuity. Informed consent of the parents was obligatory for the participation in ourtudy. Only children with an IQ higher than 85 assessed by the Coloured Progressiveatrices (Bulheller & Häcker, 2002) were included in the final sample (N = 74; see

able 1 for further details). The four groups (DYS, AD, DYS + AD, CG) did not differ inge and IQ—F(3, 70) = 1.28, p = .287 and F(3, 70) = 1.45, p = .235.

Dyslexics (DYS; N = 20) and children with combined dyslexic and attentionalroblems (DYS + AD; N = 20) had to have a reading quotient (RQ) below 85 on a stan-ardized German reading test (Mayringer & Wimmer, 2003). In this test, childrenave to read and rate the correctness of the content of as many sentences as possibleithin 3 min (e.g., Strawberries are blue). Children belonging to one of the two nor-al reading groups—AD and CG, n = 14 and 20, respectively—showed an RQ of at least

0 (see Table 1). The dyslexic group (DYS and DYS + AD) showed significantly lowereading quotients than the normally reading group (AD and CG), F(1, 72) = 326.76,< .001.

Attentional problems were defined by a T-value below 37 on three out of fiveubtests (Distractability, Alertness, Flexibility; Go/Nogo, Visual scanning) of a stan-ardized test battery (Testbatterie zur Aufmerksamkeitsprüfung für Kinder (KITAP)test battery for the assessment of attentional deficits in children]; Zimmermann,ordan, & Fimm, 2002) was obligatory. In the subtest “Distractability” children see ahost with a laughing or a sad mouth. Only if the ghost is sad, children have to pressbutton. In half of the trials a distractor appears right before the ghost. In the subtest

Alertness” a witch has to be removed as quickly as possible (with a button press) ifhe appears in a window. “Flexibility” is tested with a task where children have toeed two families of dragons (blue and green family) in an alternating manner. Thehild has to press the left or right button, depending on the side of the door wherehe dragon with the correct colour appears. During the “Go/Nogo” task children learnhat a bat which chases a cat has to be removed (with a button press) as quickly asossible. If the cat appears the child should do nothing. During “Visual scanning”hildren see witches riding on their brooms (10 × 10 per screen). Children have tondicate by button press, if all witches ride in the same direction or not. Mean error--values significantly differed between the two groups with attentional difficultiesAD and DYS + AD) and the two non-impaired groups (CG and DYS), F(1,72) = 39.83,< .001.

.2. Stimulus material

Four groups of words (n = 30) were manipulated using a 2 × 2 factorial com-

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

ination of word length and cluster density: short words, low consonant clusterensity (short-CVC), short words, high consonant cluster density (short-CCC), longords, low consonant cluster density (long-CVC) and long words, high consonant

luster density (long-CCC). Consonant cluster density was defined as the proportionf consonants in a cluster in relation to all letters of a word. For example the word

able 1ean age in month, mean intelligence quotient (IQ), mean reading quotient (RQ) andean attentional Error-T-value (error-T-value) of the four groups (dyslexics (DYS),

yslexics + ADHD (DYS + AD), children with attentional deficits (AD), control groupCG)).

DYS DYS + AD AD CG

♂13, ♀7 ♂16, ♀4 ♂10, ♀4 ♂12, ♀8

M SD M SD M SD M SD

ge in month 113.4 11.5 113.5 12.0 106.9 7.5 113.4 12.1Q 100.3 10.5 100.4 10.1 101.2 11.1 106.4 11.5Q 75.0 6.7 74.4 3.6 95.14 4.7 99.7 5.6rror-T-value 45 5.3 36 7.1 38 5.0 50 5.6

Letters 6.40 1.19 6.40 1.19 10.13 1.55 10.16 1.57Prop. consonants 0.00 0.00 0.61 0.13 0.23 0.20 0.49 0.14Num. of syllables 2.77 0.89 1.50 0.51 3.50 0.94 2.97 0.89Frequency/1Mio 0.05 0.09 0.04 0.06 0.01 0.01 0.01 0.02

Esel [donkey] has no consonant in a cluster and therefore a proportion of .00, whileKampf [fight] has three consonants in a cluster and therefore reaches a proportion of.60. For further details about number of letters, proportion of consonants in a cluster,number of syllables and word frequency see Table 2. The number of letters signif-icantly differed between short and long words, t(118) = 14.91, p < .001. Words withlow cluster density differed significantly from words with high cluster density inthe proportion of consonants in a cluster, t(112.76) = 14.68, p < .001 and in the num-ber of syllables, t(118) = 5.16, p < .001. The median split of frequency of short wordsin comparison to long words in our study showed that short words had a higherfrequency than long words (mean frequency of .012 vs. .005). Stimulus material islisted in Appendix.

2.3. Apparatus and procedure

The experiment was conducted in a mobile eye-tracking lab (“Guckomobil”—aDaimler-Benz van; www.guckomobil.de). All children were tested on the campusof their schools. The interior of the lab was shaded. Eye movements were recordedwith an EyeLink®-1000 (PC, Pentium-IV Processor) eye tracking system (SR Research,Canada) with a sampling rate of 1000 Hz. The viewing distance was 60 cm. Wordswere presented one by one in yellow letters (24pt Times New Roman) on a darkgray background in the center of a 20 in. computer screen with a refresh rate of120 Hz and a resolution of 1024 × 768. Uppercase letters of the reading material were6 mm in height which corresponded to a visual angle of 5.65◦ . We used a horizontal3-point-calibration at the beginning of the experiment which was followed by are-calibration after every third trial.

The presentation routine for each word was as follows: a fixation cross appearedon the left side of the screen. If the eye-tracker detected that the child was fixating thecross, a pre-stimulus mask (XXXXX) appeared 75 pixels to the right of the fixationcross. The number of X’s corresponded to the word length of the upcoming item.Once the child’s gaze crossed an invisible boundary (aiming to look at the X’s) a wordappeared and the child had to name the word aloud. The child was instructed to namethe word only if she already knew the pronunciation of the whole word. Reading timewas defined as the time between the appearance of the word and the voice onset.Immediately after the child finished reading the word aloud, she had to look at across on the right side of the screen. The correctness of the response was recorded.The participants were familiarized to the experiment with five training words.

3. Results

ANOVAS with group (dyslexia, dyslexia + attentional problems,attentional problems, controls) as unrepeated factor and wordlength (short vs. long) and word density (CVC vs. CCC) as repeatedfactors were computed.

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

3.1. Reading accuracy

Reading accuracy for all groups except for DYS + AD was high (seeTable 3). This fact was also revealed in an ANOVA which showed asignificant effect of group, F(3,70) = 10.99, p < .001 (�2 = .32). Bon-

Table 3Mean and standard deviation of percentages of correctly read words for the differentword categories (short-CVC, short-CCC, long-CVC, long-CCC) for the different groups(dyslexics (DYS), children with attentional problems (AD), children with comorbidproblems (DYS + AD), control children (CG)).

DYS DYS + AD AD CG

M SD M SD M SD M SD

Short-CVC 84 12 67 27 90 7 94 5Short-CCC 80 12 68 24 82 8 88 8Long-CVC 81 14 65 27 89 8 92 8Long-CCC 80 16 63 28 86 10 92 8

IN PRESSN

ychologia xxx (2009) xxx–xxx 5

fsotitrdaht

3

ral

rr(Daaddl(issocobws

3

DASc(aaip

F(

Fig. 2. Means and standard errors of number of fixations for different word cate-gories (short-CVC, short-CCC, long-CVC, long-CCC) and groups (DYS, DYS + AD, AD,CG).

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

erroni adjusted pairwise comparisons confirmed that DYS + ADhowed significantly lower rates of correctly read words than allther groups (p < .016). Words with high density led to more errorshan words with low density, F(1,70) = 7.78, p < .001 (�2 = .20). Thenteraction length × density, F(3,70) = 4.88, p < .05 (�2 = .07) showedhat only for short words density had an influence on reading accu-acy, t(73) = 3.85, p < .001. Interestingly, the interaction betweenensity and group, F(3,70) = 2.72, p = .05 (�2 = .11), revealed that DYSnd DYS + AD were not influenced by density (t < 1.5), whereas itad an effect on the error rate of the group with AD and the CG,(13) = 5.32, p < .001 and t(19) = 2.48, p < .05.

.2. Reading time

For the analyses of reading time we considered only correctlyead words. Furthermore, words with a reading onset latency belown individual z-value of 2.5 were excluded. These selection criteriaeft us with M = 78% (SD = 17.34) of the full set of items for analyses.

The main effect of group, F(3,70) = 16.63, p < .001 (�2 = .42),evealed that poor readers (DYS; DYS + AD) had significantly highereading times in comparison to children without reading problemsAD and CG). Mean onset times of the poor reading groups (DYS;YS + AD) were more than twice as long as mean onset times of thege-adequate readers [poor readers: M = 3697 ms (S.E. = 280), age-dequate readers: M = 1594 ms (S.E. = 96)]. Words with high clusterensity in comparison to words with low cluster density led toelayed onsets, F(1,70) = 12.74, p = .001 (�2 = .15). Onsets occurred

ater for long words than for short words F(1,70) = 89.71, p < .001�2 = .56). Bonferroni adjusted pairwise comparisons for the signif-cant interaction of length by group, F(3,70) = 8.58, p < .001 (�2 = .27),howed that DYS and DYS + AD had an almost identical onset forhort words, p = 1.00, which deviated significantly from the meannset of the non-impaired readers, p’s < .003. For long words theomorbid group (DYS + AD) had shorter (although non-significant)nset times than the DYS group, p = .16 and showed more similar,ut still significantly different onset patterns compared to the groupith attentional problems, p = .044. Any other interaction did reach

ignificance, all F’s < 3.43. (For further details see Fig. 1.)

.3. Number of fixations

DYS showed a mean number of 7.35 fixations (SD = 2.67),YS + AD a mean number of 5.03 fixations (SD = 2.22), whereasD and CG had about equal number of fixations (AD: M = 3.60,D = 1.37; CG: M = 3.27, SD = 1.04). Bonferroni corrected pairwiseomparisons of the main effect of group, F(3,70) = 16.81, p < .001

2

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

� = .42), showed that DYS made significantly more fixations thanll other groups (DYS + AD: p < .01; AD and CG, p’s < .001). Lengths well as the interaction of length × group reached highly signif-cant p-values, F(1,70) = 215.14, p < .001 (�2 = .76) and F(1,70) = 8.18,< .001 (�2 = .42). For short and long words DYS made consistently

ig. 1. Means and standard errors of reading times for different word categoriesshort-CVC, short-CCC, long-CVC, long-CCC) and groups (DYS, DYS + AD, AD, CG).

Fig. 3. Means and standard errors of single fixation durations for different wordcategories (short-CVC, short-CCC, long-CVC, long-CCC) and groups (DYS, DYS + AD,AD, CG).

more fixations than all other groups (DYS + AD: p < 01; AD and CG:p’s < .001). On the other hand, the profile of DYS + AD in compari-son to the other groups changed somewhat between short and longwords. Reading short words, DYS + AD made clearly more fixationsthan AD (p = .058) and CG (p = .01), but when reading long words thisdifference was no longer observable (in comparison to AD: p = .592;in comparison to CG: p = .121). For all groups, words with high con-sonant density caused more fixations than words with low density,F(1,70) = 7.19, p < .01 (�2 = .09). For further details see Fig. 2.

3.4. Mean single fixation duration

Fig. 3 shows mean single fixation durations for the four groupsand the four word categories. In contrast to mean number of fix-ations the main effect of group, F(3,70) = 9.47, p < .001 (�2 = .29)for mean single fixation durations revealed the following pattern;whereas DYS had longer mean fixations than CG (p < .01), DYS + ADshowed longer mean fixations than CG and than AD, p < .001 andp = .011. For all groups the average fixation duration was longer forshort words than for long words, F(1,70) = 33.14, p < .001 (�2 = .32).The interaction of length x density, F(1,70) = 7.51, p < .01 (�2 = .10),revealed that for short words density had no additional effect,whereas long words with high density provoked longer mean fixa-tion durations than long words with low density, t(73) = 2.98, p < .01.

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

3.5. Orthographic characteristics of consonant clusters1

In our study, consonant clusters were defined as letter sequencesof non-vocalic letters. However, this definition is rather untypi-

1 The issue of possible confounds in our definition of orthographic characteristicswas raised by one of the reviewers.

ARTICLE ING ModelNSY-3275; No. of Pages 10

6 V. Thaler et al. / Neuropsycholo

Table 4Ratio of the number of fixations: mean and standard deviations for different lin-guistic categories (di(tri)graph-cluster vs. phonologically based consonant cluster;cluster within one syllable vs. clusters spanning two syllables; consonant clusterwithin one morpheme vs. cluster spanning two morphemes) for the four groups(DYS, DYS + AD, AD, CG).

DYS DYS + AD AD CG

M SD M SD M SD M SD

Di(tri)graph cl. .83 .28 .59 .24 .44 .19 .41 .11Phon. cl. .83 .58 .69 .43 .30 .20 .33 .22

1-syllable cl. .80 .36 .61 .37 .43 .14 .41 .142

12

cdFtt“wK

orwtwacp[ci

ttnmiaDrttwbaate

4

iLatHL

ao

-syllables cl. .96 .50 .73 .43 .41 .28 .35 .16

-morpheme cl. .87 .39 .66 .30 .40 .21 .38 .14-morphemes cl. 1.05 .42 .79 .55 .50 .23 .52 .23

al because consonant groups like ch or sch are usually defined asigraphs or trigraphs, because they represent a single phoneme.urthermore, we did not differentiate between consonant clus-ers within one syllable and consonant clusters which span acrosswo syllables (e.g., Kampf [fight] vs. Kir-sche [cherry]). Additionally,long words – CCC” mainly consisted of consonant clusters thatere generated through the composition of two morphemes (e.g.,eller-treppe [basement stairs]).

To analyze the effects of these different orthographic factorsn the reading behavior of the different groups, we analyzed theatio of the number of fixations per letters of a cluster (1) for wordsith phonologically based consonant clusters vs. consonant clus-

ers consisting of digraphs and trigraphs, (2) consonant clustersithin one syllable vs. consonant clusters spanning two syllables,

nd (3) consonant clusters within one morpheme vs. consonantlusters spanning across two morphemes. The ratio of fixationser letters of a consonant cluster was calculated as follows: Kampffight] consists of three letters within the consonant cluster. If ahild fixated a 3-consonant cluster twice than the ratio of fixationss 2/3 = .67.

Means and standard deviations for the ratio of fixations are illus-rated in Table 4. T-tests showed that for DYS, DYS + AD, and CGhe difference between the two groups of consonant clusters wasot significant t’s(19) < 1.88. Only AD children made significantlyore fixations, if they had to read consonant clusters consist-

ng of digraphs or trigraphs, t(13) = 3.319, p < .01. Furthermore, thenalyses revealed that for children with reading problems (DYS,YS + AD) clusters spanning two syllables were more difficult to

ead than clusters within one syllable, t(19) = −2.09, p = .05 and(19) = −2.59, p < .05. For AD there was no significant difference,(13) = −1.7, p = .11. A particularly important finding was that CGas the only group that showed an advantage for a syllable splitetween consonant clusters, t(19) = 2.17, p < .05. Finally, DYS, CGnd AD needed more fixations when a consonant clusters spannedcross the boundary of morphemes, t’s(19) > −2.93, p < .01 and(13) = −3.45, p < .01. For DYS + AD, there was no significant differ-nce, t(13) = −1.55, p = .14.

. Discussion

Comorbidity is a very common phenomenon in various learn-ng disabilities (e.g., Capano et al., 2008; Gross-Tsur et al., 1996;ewis et al., 1994; Manor et al., 2008). Nonetheless, studies whichssessed behavioral symptoms of learning disabilities often neglecthis important issue (e.g., Barca et al., 2006; Di Filippo et al., 2006;

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

utzler & Wimmer, 2004; Juphard et al., 2004; Katz et al., 2005;anderl et al., 1997; Visser & Besner, 2001; Ziegler et al., 2003).

The effect of attentional problems was repeatedly shown to bepotential confound in studies examining the causal hypotheses

f dyslexia (Bishop, 2002; Pennington, 2006; Ramus et al., 2003).

PRESSgia xxx (2009) xxx–xxx

Wimmer and Raberger (Raberger & Wimmer, 1999; Wimmer etal., 1999), for example, showed that once dyslexic children werechecked for additional attentional deficits, anomalies in tasks sen-sitive to a cerebellar deficit were not observed. Nonetheless, untilnow, the importance of controlling for comorbid attentional prob-lems in studies of reading disability has been widely neglected.

The goal of our study was to investigate the contribution of atten-tional problems to different reading patterns of reading impairedchildren and children with no specific reading impairment butattentional problems. Therefore, four groups of children—dyslexicreaders, dyslexic readers with attentional problems, children withattentional problems and unaffected control children—read sin-gle words of different lengths and orthographic complexity. Todemonstrate possible effects of attentional problems on reading,two well-established phenomena of dyslexic reading—the wordlength effect and the large-unit effect—were analyzed in the fourgroups. For a fine-grained analyses of reading characteristics, eyemovements were recorded.

4.1. Group effects

4.1.1. Reading accuracyReading accuracy of unimpaired readers (AD and CG) was

high (between 86% and 94% correctly read words). Childrenwith reading-problems-only (DYS) had a reading accuracy ratecomparable to normal readers (between 81% and 84%). In con-trast, children with dyslexia and an attentional deficit (DYS + AD)revealed a particularly low reading accuracy (between 63% and64%). While most studies in regular orthographies have shownthat dyslexic readers have no problems with reading accuracy(e.g., de Jong, 2003; Landerl, 2003; Wimmer, 1993), other studieshave revealed accuracy problems (Klicpera & Schabmann, 1993;Landerl et al., 1997; Nergard-Nilsen, 2005). Our study confirmedthe assumption that dyslexic children in regular orthographieshave rather little problem with respect to accuracy. Additionally,we added an important aspect to the puzzle of understanding whysome studies in regular orthographies possibly deviate from thisobservation. We could show that children with dyslexia, who haveadditional problems with attention, do make a higher numberof errors during reading. The typical symptoms of children withattentional problems are increased distractibility, inattention andcareless errors (e.g., Epstein et al., 2003; Johnson et al., 2007;O’Connell, Bellgrove, Dockree, & Robertson, 2004; van Meel et al.,2007; Wilding et al., 2007). Consequently, one would assume thatAD children—independent of comorbidity—should also make moreerrors during reading. However, since children frequently have toread sentences and texts in class, they should be highly practiced inreading long texts. It thus seems reasonable to assume that readingsingle words does no longer require an especially high level of con-centration and sustained attention over a longer period in childrenwith no reading problems. Therefore, it is plausible to assume thatchildren with attentional-deficits-only should show comparablerates of careless errors compared with normally developing controlchildren. For DYS + AD reading single words requires a high level ofconcentration and sustained attention which they cannot afford—hence it was predictable that children with reading and attentionaldeficits should show considerably more errors than all other groups.

4.1.2. Reading speedAnalysis of reading speed (as well as reading accuracy) con-

firmed the typical result of studies with dyslexic children in regular

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

orthographies. Their main problem is a massively slow readingspeed. DYS needed almost three times longer to read a word thanCG. The differentiation between DYS and DYS + AD revealed thatDYS + AD differs less from normal readers than DYS (reading onset:M = 1900 ms vs. 2700 ms). This is at odds with the interpretation

INN

ycholo

tttJSmpdbmpsltofi(osdbnnA

4

dbofisttww

4

ascreffscElwitpdftDcEeswae

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

hat children with comorbid disorders—irrespective of the kind ofhe disorders (e.g., dyscalculia, dyslexia, dyspraxia, ADHD)—showhe most aggravated deficits (e.g., Ackerman & Dykman, 1995;ongmans et al., 2003; Landerl et al., 2004; Purvis & Tannock, 1997;halev et al., 1997). In our case, the comorbid group did not showore severe problems than the dyslexic group, but their reading

roblems were qualitatively different. DYS mainly suffered fromeficient reading speed; DYS + AD suffered stronger from a accuracyut a less from a reading speed impairment. Additionally, eye move-ents confirmed that DYS and DYS + AD have very different reading

rofiles. Eye movements of DYS closely reflect their serial decodingtrategy. DYS children fixated almost every letter (mean number ofetters: 8.27; mean number of fixations: 7.35). The duration fixa-ions were longer than that of CG and AD, but not as long as thatf DYS + AD. DYS + AD particularly exhibited anomalies concerningxation durations but not concerning mean number of fixationswhich was significantly lower than in DYS and comparable to thatf AD). The comorbid group did not rely on a letter-by-letter readingtrategy, but analyzed larger chunks—resulting in higher fixationurations. The reading strategy of DYS + AD—that is the decoding ofigger units rather than serial, letter-by-letter word decoding—doesot seem to be adequate for their reading abilities and resulted in aoticeable higher error rate compared to all other groups (DYS, CG,D).

.1.3. ConclusionThe comparison of the four tested groups showed considerable

ifferences in their reading strategies. While DYS read accurately,ut exhibited the longest reading times and the highest numberf fixations, DYS + AD made fewer fixations, but had the highestxation durations and the highest error rate. AD and CG showedimilar reading patterns in every aspect (reading accuracy, readingimes, number of fixations, mean single fixation duration). Addi-ionally, we assessed the effect of comorbid attentional problemsith respect to two well-established effects in dyslexic readers—theord length effect and the large-unit effect.

.2. The word length effect

All four groups showed a word length effect in that onset timesnd number of fixations were higher for longer words. Analy-is of reading times confirmed that word length affected dyslexichildren (DYS and DYS + AD) to a greater extent than the normaleaders (CG and AD). DYS were the most affected children. Inter-stingly, mean reading time of DYS differed more from DYS + ADor long words than for short words. At the same time, the dif-erence in reading time between DYS + AD and AD was larger forhort words. It is well established that dyslexic children have diffi-ulties in building up an orthographic lexicon (e.g., Dejerine, 1891;hri, 1998; McCandliss, Cohen, & Dehaene, 2003). The orthographicexicon is assumed to be a store of frequently encountered words

hich serves fast and effortless recognition of already anchoredtems. Rarely encountered, low frequency words are not supposedo be stored in the orthographic lexicon and therefore have to berocessed by serial decoding. One consequence of a lacking orefect orthographic lexicon is that almost all words—even high-requency words—cannot be recognized “automatically”, but haveo be decoded in a serial manner. Both dyslexic groups (DYS andYS + AD) did not seem to have built up an orthographic lexi-on yet and, therefore, even short high-frequency words such assel [donkey] or Fisch [fish] had to be decoded serially leading to

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

xorbitant reading times. As our participants were children fromecond to fifth grade it is not surprising that long, low frequencyords are not yet part of their orthographic lexicon. Therefore,

ll children showed higher reading times for long words. How-ver, the reading onset of dyslexic children was significantly higher

PRESSgia xxx (2009) xxx–xxx 7

than that of normally developing children confirming the persis-tence of the slow serial decoding strategy of dyslexic readers. DYSshowed clearly more severe decoding difficulties than DYS + ADwhich resulted in higher onset times in reading words for the for-mer group. The analyses of eye movements proved to be a feasiblemethod for showing the differential problems of DYS and DYS + AD.During reading long words DYS + AD made significantly fewer fix-ations than DYS and a similar number as AD and CG. In contrast,for short words DYS + AD made significantly fewer fixations thanDYS, but needed significantly more fixations than AD and CG. Forboth dyslexic groups it can be assumed that they have deficits inbuilding up an orthographic lexicon, but DYS + AD decoded sup-posedly non-stored words with significantly fewer fixations thanDYS. One could interpret this pattern as indicating a better decodingstrategy of DYS + AD, but has to put into perspective that DYS + ADshowed the highest error rates for all types of words. Reasonably,their impulsivity and their problem with sustained attention dis-rupt their reading process resulting in a significantly higher errorrate compared to each of the other groups. In other words, eventhough DYS + AD can read words correctly by analyzing bigger sub-lexical units, their attentional and impulsivity problems lead tovery high error rates. Additionally, the building up of a store forfrequently encountered words seems to be a major challenge forcomorbid children. Recent studies on children with ADHD indicatedsubstantial deficits in their memory system(s) (e.g., Dige, Maahr,& Backenroth-Ohsako, 2008; Mariani & Barkley, 1997; Seidman,2006). As aforementioned, memory problems—especially in build-ing up an orthographic lexicon—are generally assumed for dyslexicsas well (e.g., Ehri, 1998; Pugh et al., 2000). Taking into regard thedouble vulnerability of memory deficits of the comorbid group, itis not surprising that DYS + AD showed the greatest difficulties instoring short, frequently encountered words.

4.2.1. ConclusionAll groups were affected by word length, but DYS showed the

greatest effects. For long words, DYS made significantly more fixa-tions than all other groups indicative of their serial, letter-by-letterdecoding strategy. DYS + AD more strongly deviated from normalreaders concerning short words. We suggest that their comorbiddeficits—both associated with poor memory capabilities—led toadditional problems in building up an orthographic lexicon evenfor short, frequently encountered words.

4.3. The large-unit effect

4.3.1. Orthographic characteristics of consonant clusters. Before discussing the large-unit effect, we will go into further

detail about the orthographic characteristics of our consonant clus-ters. By means of post hoc analysis we looked into the effect of (1)di(tri)graph clusters vs. phonologically based clusters, (2) clusterswithin one syllable vs. clusters spanning across two syllables and(3) clusters within one morpheme vs. clusters spanning across twomorphemes. With regards the first factor, we could show that ourparticipants (children from Grade 2 to 5) needed the same num-ber of fixations for the two classes of consonant clusters. It can beassumed that children in this age group either do or do not clusterconsonantal letters independent of their phonological characteris-tics. In contrast, CG already seemed to use a reading strategy withsyllable structure, whereas DYS and DYS + AD had even more prob-lems reading clusters spanning across two syllables. Clusters, thatwere divisible between two morphemes, needed more fixations

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

than clusters within one morpheme which indicates that the skillfor recognizing such boundaries was not yet fully developed in ourparticipants.

Our word material surely was not adequate to deduce firmconclusions on different orthographic effects, but it appears as

INN

8 ycholo

iptftTmi

4

GtuoostssewcdnDouodgwlfwdgshwonroimgdteosscWeotcra

4

do

We thank Manon Jones for proofreading the manuscript. We aregrateful for the helpful comments of the two anonymous reviewers.

Appendix A.

Stimulus list: Short words with low and high cluster density and long words withlow and high cluster density.

Short-CVC Short-CCC Long-CVC Long-CCC

Esel Obst Tierarzt FeldmausPedal Dachs Segelohr GipsbeinMonat Durst Sägemehl HalsbandBoden Fisch Maikäfer HandtuchHafen Kampf Teekanne SalzdoseRaupe Ochse Staudamm WitzbuchMagen Sumpf Weidezaun BesenstielMotor Wurst Blaulicht GeldbeutelRakete Frucht Seeräuber SandhaufenMinute Kirche Probefahrt SchuhgrößeTapete Delphin Schneeball TürpfostenDaumen Fenster Steinadler HaferflockeAnorak Förster Pferdekopf KellertreppeBeutel Hamster Feueralarm PuppenstubeKanone Mädchen Autofahrer KaspertheaterZuhause Herbst Schneehase Randstein

ARTICLEG ModelSY-3275; No. of Pages 10

V. Thaler et al. / Neurops

f children with reading disabilities in a transparent orthogra-hy seem to rely on a serial letter-based decoding strategy, notaking phonological characteristics into account. Normal readersrom Grade 2 to 5, in contrast, seem to profit from split-ing complex orthographic units into syllables or morphemes.his vague hypothesis—based on post hoc analysis of our wordaterial—shows the importance of investigating these questions

n future research.

.3.2. Large-unit effectWithin the framework of their Grain Size Theory, Ziegler and

oswami (2005) propose that the orthography of a language defineshe typical grain size of reading units (e.g., grapheme-phonemenits; rime units; body neighbors). For the transparent Germanrthography they propose that children rely more or less exclusivelyn highly efficient sublexical decoding procedures at small grainizes (i.e., phoneme-grapheme correspondences). According toheir theory, regular orthographies induce the usage of highly con-istent grapheme-phoneme correspondences (rather than largerublexical units). This hypothesis fits well with results of Zieglert al. (2003), showing that for German children the length effectas not reduced by high body neighborhood (in contrast to English

hildren). This was true for normal readers as well as for reading-isabled readers. Nonetheless, there exist several studies that doot fit to this observation (e.g., Caravolas, 2006; de Jong, 2006;urgunoglu, 2006; Paulesu, 2006; Wimmer, 2006). Our analysesf cluster density showed that children in a regular orthographyse larger sublexical units than proposed by the Grain Size The-ry. In general, reading accuracy of CG and AD was affected byensity while this was not the case for the two reading disabledroups (DYS and DYS + AD). Independently of group, short wordsith high density lead to a higher error rate than short words with

ow density. Reading times and number of fixations was higheror words with a high consonant cluster density than for wordsith a low one. This pattern held for all groups. The effect ofensity on mean single fixation duration was also similar for allroups. For short words, mean single fixation durations were theame for the low and high density conditions. Long words withigh density caused higher single fixation durations than wordsith low density. Our study thus revealed that children in regular

rthographies are influenced by consonant cluster density and thatormal and poor readers are influenced to the same degree. Theseesults have two implications: first it suggests that the statementf the grain size proponents—that reading in a regular orthographys mostly letter-based—does not seem to be adequate for Ger-

an readers, because they are using sublexical chunks larger thanrapheme-phoneme correspondences. Second, consonant clusterensity—almost neglected by reading researchers (with the excep-ions of Bruck & Treiman, 1990; Treiman, 1991)—seems to be anffect-carrying unit with a reliable impact on the reading efficacyf German-speaking children. This suggests that it is worthwhiletudying this effect more intensively. That the processing of con-onant clusters is relevant for increasing reading speed of dyslexichildren was previously shown in a training study (Thaler, Ebner,

immer, & Landerl, 2004). In this study, 20 dyslexic readers repeat-dly read 32 training words with different consonant clustersver a period of 25 days. Post-tests—one and five weeks afterraining—showed that reading fluency of the poor readers signifi-antly increased over the remediation period although they did noteach an age-adequate level. This indicates that consonant clustersre units relevant for reading and reading remediation.

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

.3.3. ConclusionAll groups were influenced by cluster density to the same

egree. This shows that all children in the context of the Germanrthography—independent of reading level—use sublexical units

PRESSgia xxx (2009) xxx–xxx

above the grapheme-phoneme level and that consonant clustersare a sublexical unit that deserves being studied in more detail infurther studies.

4.4. General conclusion

Our study revealed that dyslexic children (DYS) and childrenwith dyslexic and attentional deficits (DYS + AD) show distinguish-able reading profiles in a transparent orthography. Dyslexic childrenshowed the typical pattern of high reading accuracy, but severedeficits in reading fluency. In contrast, comorbid children showedespecially low reading accuracy and milder deficits in readingfluency. Eye movement analysis offered a more detailed pictureabout reading deficits of the two groups (DYS and DYS + AD). Whiledyslexic children fixated every letter of a word independently ofword length, comorbid children had particular difficulties withshort, high-frequency words. Comorbid children’s number of fix-ations for longer words was comparable to that of normal readers.Their marked problem with short words highlights their difficultiesin building up a store of often encountered words. The comor-bid attentional problems can be assumed to hamper DYS + AD inadapting their reading strategy to their reading abilities resultingin a dramatically higher error rate. However, their reading deficitseems to be qualitatively different to that of DYS, as their readingstrategy deviates less from normal readers with and without atten-tional problems. DYS + AD therefore (try to) read words by utilizingbigger sublexical chunks which sometimes results in a success andfrequently results in an erroneous word production.

In conclusion, we demonstrated the importance of controllingfor additional problems of attention in dyslexia studies for theadequate description and the examination of causal theories ofdevelopmental dyslexia. Especially when dyslexia is diagnosed viaa high number of reading errors—which is the common criterion inEnglish-based studies—one has to take care not to include dyslexicreaders with comorbid attentional problems.

Acknowledgements

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

Kapitän Hirsch Taxifahrer SturmflutAmeisen Schlaf Hundeleine SturzflugFamilie Schlag Speisewagen BahnbrückeMedizin Schlot Vogeleltern DampfnudelPapagei Ärmchen Montagabend FuchshöhleKönigin Kirsche Hausaufgabe Apfelschale

ARTICLE ING ModelNSY-3275; No. of Pages 10

V. Thaler et al. / Neuropsycholo

Appendix A (Continued )

Short-CVC Short-CCC Long-CVC Long-CCC

Bananen Rutsche Fotoapparat AprilscherzTelefon Schmerz Schlafanzug RegenschirmPolizei Schmuck Winterabend SprungbrettParadies Lutscher Sommeranzug StrumpfhoseZauberer Peitsche Mittagessen GlassplitterDienerin Schlange Siegerurkunde DampfmaschineLS

R

A

A

B

B

B

B

B

B

B

B

B

C

C

C

C

C

C

D

d

d

D

D

D

D

D

MacKeben, M., Trauzettel-Klosinski, S., Reinhard, J., Dürrwächter, U., Adler, M., &Klosinski, G. (2004). Eye movement control during single-word reading in

imonade Schlinge Autoreparatur Tomatenpflanzeanitäter Abschluss Kaffeekännchen Marktplatz

eferences

ckerman, P. T., & Dykman, R. A. (1995). Reading-disabled students with andwithout comorbid arithmetic disability. Developmental Neuropsychology, 11(3),351–371.

rduino, L. S., & Burani, C. (2004). Neighborhood effects on nonword visual process-ing in a language with shallow orthography. Journal of Psycholinguistic Research,33(1), 75–95.

arca, L., Burani, C., Di Filippo, G., & Zoccolotti, P. (2006). Italian developmentaldyslexic and proficient readers: Where are the differences? Brain and Language,98, 347–351.

arr, C. L., Wigg, K., Laurin, N., Misener, V., Cate-Carter, T., Maureen, L. W., Kerr, E., &Couto, J. (2008). Attention deficit/hyperactivity disorder and dyslexia: Evidencefor shared genetic susceptibility. European Psychiatry, 23, 25.

ishop, D. V. M. (2002). Cerebellar abnormalities in developmental dyslexia: Cause,correlate or consequence? Cortex, 38, 491–498.

rand, M., Groux, I., Puijalon, C., & Rey, A. (2007). Syllable onsets are perceptualreading units. Memory & Cognition, 35(5), 966–973.

rookes, R. L., Nicolson, R. I., & Fawcett, A. J. (2007). Prisms throw light on develop-mental disorders. Neuropsychologia, 45, 1921–1930.

rook, U., & Boaz, M. (2005). Attention deficit and learning disabilities (ADHD/LD)among high school pupils in Holon (Israel). Patient Education and Counseling, 58,164–167.

ruck, M., & Treiman, R. (1990). Phonological awareness and spelling in normal chil-dren and dyslexics: The case of initial consonant clusters. Journal of ExperimentalChild Psychology, 50, 156–178.

ulheller, S. & Häcker, H. (2002). Coloured progressive matrices—German version.Manual for Raven’s Progressive Matrices and Vocabulary Scales of J Raven, JCRaven and JH Court. Frankfurt: Swets Test Services.

urd, L., Freeman, R. D., Klug, M. G., & Kerbeshian, J. (2005). Tourette syndrome andlearning disabilities. BMC Pediatrics, 5(34), 1471–2431.

apano, L., Minden, D., Chen, S., Schachar, R. J., & Ickowicz, A. (2008). Mathemati-cal learning disorder in school-age children with attention-deficit hyperactivitydisorder. La Revue canadienne de psychiatrie, 53(6), 392–399.

aravolas, M. (2006). Refining the psycholinguistic grain size theory: Effects ofphonotactics and word formation on the availability of phonemes to preliteratechildren. Developmental Science, 9(5), 445–447.

arlisle, J. F., & Katz, L. A. (2006). Effects of word and morpheme familiarity on readingof derived words. Reading and Writing, 19, 669–693.

oltheart, M., & Rastle, K. (1994). Serial processing in reading aloud—evidence fordual-route models of reading. Journal of Experimental Psychology: Human Percep-tion and Performance, 20(6), 1197–1211.

onrad, M., Grainger, J., & Jacobs, A. M. (2007). Phonology as the source of syllablefrequency effects in visual word recognition: Evidence from French. Memory &Cognition, 35(5), 974–983.

rawford, S. G., & Dewey, D. (2008). Co-occurring disorders: A possible key tovisual perceptual deficits in children with developmental coordination disorder?Human Movement Science, 27, 154–169.

ejerine, J. (1891). Sur un cas de cecite verbale avec agraphic, suivi d’autopsie. SocieteBiologique, 3, 197–201.

e Jong, P. F. (2003). Problems in the acquisition of fluent word decoding in Dutchchildren. In N. Goulandris (Ed.), Dyslexia: A cross-linguistic comparison (pp.33–52). London: Whurr.

e Jong, P. F. (2006). Units and routes of reading in Dutch. Developmental Science,9(5), 441–442.

e Luca, M., Borrelli, M., Judica, A., Spinelli, D., & Zoccolotti, P. (2002). Reading wordsand pseudowords: An eye movement study of developmental dyslexia. Brain andLanguage, 80, 617–626.

e Luca, M., Di Pace, E., Judica, A., Spinelli, D., & Zoccolotti, P. (1999). Eye movementpatterns in linguistic and non-linguistic tasks in developmental surface dyslexia.Neuropsychologia, 37, 1407–1420.

i Filippo, G., De Luca, M., Judica, A., Spinelli, D., & Zoccolotti, P. (2006). Lexicalityand stimulus length effects in Italian dyslexics: Role of the over-additivity effect.Child Neuropsychology, 12(2), 141–149.

ige, N., Maahr, E., & Backenroth-Ohsako, G. (2008). Memory tests in subgroupsof adult attention deficit hyperactivity disorder reveals simultaneous capacity

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

deficit. International Journal of Neuroscience, 118(4), 569–591.illing, H., Mombour, W., & Schmidt, M. H. (2004). ICD-10 Kapitel V (F): Klinisch

diagnostische Leitlinien [ICD-10 chapter V (F): Clinical diagnostic guidelines]. Bern:Verlag Hans Huber.

PRESSgia xxx (2009) xxx–xxx 9

Doignon, N., & Zagar, D. (2005). Illusory conjunctions in French: The nature of sub-lexical units in visual word recognition. Language and Cognitive Processes, 20(3),364–443.

Donfrancesco, R., Mugnaini, D., & Dell’Uomo, A. (2005). Cognitive impulsivity inspecific learning disabilities. European Child—Adolescent Psychiatry, 14, 270–275.

Durgunoglu, A. Y. (2006). Learning to read in Turkish. Developmental Science, 9(5),437–439.

Ehri, L. C. (1998). Grapheme-phoneme knowledge is essential for learning to readwords in English. In J. L. Metsala, & L. C. Ehri (Eds.), Word recognition in beginningliteracy (pp. 3–40). London: Lawrence Erlbaum Associates Publishers.

Elbro, C. (2005). Literacy acquisition in Danish: A deep orthography in cross-linguistic light. In R. M. Joshi, & P. G. Aaaron (Eds.), Handbook of orthographyand literacy (pp. 300–350). Mahwa, NJ: Lawrence Erlbaum.

Epstein, J. N., Erkanli, A., Conners, C. K., Klaric, J., Costello, J. E., & Arngold, A. (2003).Relations between continuous performance test performance measures andADHD behaviors. Journal of Abnormal Child Psychology, 31(5), 543–554.

Ferrand, L., & New, B. (2003). Syllabic length effects in visual word recognition andnaming. Acta Psychologica, 113, 167–183.

Friedman, M. C., Chhabildas, N., Budhiraja, N., Willcutt, E. G., & Pennington, B. F.(2003). Etiology of the comorbidity between RD and ADHD: Exploration of thenon-random mating hypothesis. American Journal of Medical Genetics Part B (Neu-ropsychiatric Genetics), 120B, 109–115.

Grainger, J., & Jacobs, A. M. (1993). Masked partial-word priming in visual wordrecognition: Effects of positional letter frequency. Journal of Experimental Psy-chology: Human Perception and Performance, 19, 951–964.

Gross-Tsur, V., Manor, O., & Shalev, R. S. (1996). Developmental dyscalculia: Preva-lence and demographic features. Child Neurology, 38(1), 25–33.

Haslum, M. N., & Miles, T. R. (2007). Motor performance and dyslexia in a nationalcohort of 10-year-old children. Dyslexia, 13, 257–275.

Hendriksen, J. G. M., Keulers, E. H. H., Feron, F. J. M., Wassenberg, R., Jolles, J., & Vles, J. S.H. (2007). Subtypes of learning disabilities. Neuropsychological and behaviouralfunctioning of 495 children referred for multidisciplinary assessment. EuropeanChild—Adolescent Psychiatry, 16, 517–524.

Hofmann, M., Stenneken, P., Conrad, M., & Jacobs, A. M. (2007). Sublexical frequencymeasures for orthographic and phonological units in German. Behavior ResearchMethods, 39(3), 620–629.

Hutzler, F., Bergmann, J., Conrad, M., Kronbichler, M., Stenneken, P., & Jacobs, A. M.(2004). Inhibitory effects of first syllable-frequency in lexical decision: An event-related potential study. Neuroscience Letters, 373(3), 179–184.

Hutzler, F., Conrad, M., & Jacobs, A. M. (2005). Effects of syllable-frequency in lex-ical decision and naming: An eye-movement study. Brain and Language, 92(2),138–152.

Hutzler, F., Kronbichler, M., Jacobs, A. M., & Wimmer, H. (2006). Perhaps correlationalbut not causal: No effect of dyslexic readers’ magnocellular system on their eyemovements during reading. Neuropsychologia, 44, 637–648.

Hutzler, F., & Wimmer, H. (2004). Eye movements of dyslexic children when readingin a regular orthography. Brain and Language, 89, 235–242.

Hyönä, J., & Pollatsek, A. (1998). Reading Finnish compound words: Eye fixations areaffected by component morphemes. Journal of Experimental Psychology: HumanPerception and Performance, 24(6), 1612–1627.

Johnson, K. A., Kelly, S. R., Bellgrove, M. A., Barry, E., Cox, M., Gill, M., et al. (2007).Response variability in attention deficit hyperactivity disorder: Evidence forneuropsychological heterogeneity. Neuropsychologia, 45, 630–638.

Jongmans, M. J., Bouwien, C., Smits-Engelsman, M., & Schoemaker, M. M. (2003). Con-sequences of comorbidity of developmental coordination disorders and learningdisabilities for severity and pattern of perceptual-motor dysfunction. Journal ofLearning Disabilities, 36(6), 528–537.

Juphard, A., Carbonnel, S., & Valdois, S. (2004). Length effect in reading and lexicaldecision: Evidence from skilled readers and a developmental dyslexic partici-pant. Brain and Cognition, 55, 332–340.

Katz, L., Lee, C. H., Tabor, W., Frost, S. J., Mencl, W. E., Sandak, R., et al. (2005). Behav-ioral and neurobiological effects of printed word repetition in lexical decisionand naming. Neuropsychologia, 43(14), 2068–2083.

Klicpera, C., & Schabmann, A. (1993). Do German-speaking children have a chanceto overcome reading and spelling difficulties? A longitudinal survey from thesecond until the eighth grade. European Journal of Psychology of Education, 8(3),307–323.

Koschak, J., Kunert, H. J., Derichs, G., Weniger, G., & Irle, E. (2003). Impaired andenhanced attentional function in children with attention deficit/hyperactivitydisorder. Psychological Medicine, 33(3), 481–489.

Landerl, K. (2003). Dyslexia in German-speaking children. In N. Goulandris (Ed.),Dyslexia: A cross-linguistic comparison (pp. 1–32). London: Whurr.

Landerl, K., Bevan, A., & Butterworth, B. (2004). Developmental dyscalculia and basicnumerical capacities: A study of 8-9-year-old students. Cognition, 93(2), 99–125.

Landerl, K., Wimmer, H., & Frith, U. (1997). The impact of orthographic consistencyon dyslexia: A German-English comparison. Cognition, 63, 315–334.

Lewis, C., Hitch, G. J., & Walker, P. (1994). The prevalence of specific arithmetic diffi-culties and specific reading difficulties in 9-year-old to 10-year-old boys and girls.Journal of Child Psychology and Psychiatry and allied Disciplines, 35(2), 283–292.

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

dyslexics. Journal of Vision, 4, 388–402.Manor, I., Medad, S., Zamishlani, Z., & Vurmbrand, N. (2008). ADHD frequency and

characteristics in students suffering from learning disabilities. European Psychi-atry, 23, 26–27.

INN

1 ycholo

M

M

M

M

M

N

N

N

N

N

N

O

P

P

P

P

P

R

R

R

R

R

S

S

S

S

S

S

S

ARTICLEG ModelSY-3275; No. of Pages 10

0 V. Thaler et al. / Neurops

ariani, M. A., & Barkley, R. A. (1997). Neuropsychological and academic functioningin preschool boys with attention deficit hyperactivity disorder. DevelopmentalNeuropsychology, 13(1), 111–129.

artens, V. E. G., & de Jong, P. F. (2006). The effect of word length on lexical decisionin dyslexic and normal reading children. Brain and Language, 98, 140–149.

artensen, H., Maris, E., & Dijkstra, T. (2000). When does inconsistency hurt? On therelation between phonological consistency effects and the reliability of sublex-ical units. Memory & Cognition, 28(4), 648–656.

ayringer, H., & Wimmer, H. (2003). Salzburger Lese-Screening für die Klassenstufen1-4 (SLS 1-4) [Salzburg Reading-Screening from Grade 1–4]. Göttingen: HogrefeVerlag.

cCandliss, B. D., Cohen, L., & Dehaene, S. (2003). The visual word form area: Exper-tise for reading in the fusiform gyrus. Trends in Cognitive Science, 7(7), 293–299.

eedle, J. L., Fawcett, J., & Nicolson, R. I. (2006). Balance and dyslexia: An investigationof adults’ abilities. European Journal of Cognitive Psychology, 18(6), 909–936.

ergard-Nilsen, T. (2005). Word-decoding deficits in Norwegian: The impact of psy-cholinguistic marker effects. Reading and Writing, 19, 265–290.

ewcorn, J. H., Halperin, J. M., Jensen, P. S., Abikoff, H. B., Arnold, L. E., Cantwell, D. P.,et al. (2001). Symptoms profiles in children with ADHD: Effects of comorbidityand gender. Journal of the American Academy of Child and Adolescent Psychiatry,40(2), 137–146.

icolson, R. I., & Fawcett, A. J. (1990). Automaticity: A new framework for dyslexiaresearch?? Cognition, 35, 159–182.

icolson, R. I., & Fawcett, A. J. (1995). Dyslexia is more than a phonological disability.Dyslexia, 1, 19–36.

icolson, R. I., & Fawcett, A. J. (2007). Procedural learning difficulties: Reuniting thedevelopmental disorders? Trends in Neurosciences, 30(4), 135–141.

’Connell, R. G., Bellgrove, M. A., Dockree, P. M., & Robertson, I. H. (2004). Reducedelectrodermal response to errors predicts poor sustained attention performancein attention deficit hyperactivity disorder. Neuroreport, 15(16), 2535–2538.

aulesu, E. (2006). On the advantage of “shallow” orthographies: Number and grainsize of the orthographic units or consistency per se? Developmental Science, 9(5),443–444.

ennington, B. F. (2006). From single to multiple deficit models of developmentaldisorders. Cognition, 101, 385–413.

ugh, K. (2006). A neurocognitive overview of reading acquisition and dyslexia acrosslanguages. Developmental Science, 9(5), 448–450.

ugh, K. R., Mencl, W. E., Jenner, A. R., Katz, L., Frost, S. J., Lee, J. R., et al. (2000). Func-tional neuroimaging studies of reading and reading disability (developmentaldyslexia). Mental Retardation and Developmental Disabilities Research Reviews, 6,207–213.

urvis, K. L., & Tannock, R. (1997). Language abilities in children with attentiondeficit hyperactivity disorder, reading disabilities, and normal controls. Journalof Abnormal Child Psychology, 25(2), 133–144.

aberger, T., & Wimmer, H. (1999). Ist Leseschwäche durch ein Automatisierungs-defizit verursacht? [Is poor reading caused by an automatisation deficit?].Zeitschrift für Pädagogische Psychologie, 13(1/2), 74–83.

aberger, T., & Wimmer, H. (2003). On the automaticity/cerebellar deficit hypothe-sis of dyslexia: Balancing and continuous rapid naming in dyslexic and ADHDchildren. Neuropsychologia, 41, 1493–1497.

ack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword reading deficit indevelopmental dyslexia: A review. Reading Research Quarterly, 27(1), 28–53.

amus, F., Pidgeon, E., & Frith, U. (2003). The relationship between motor controland phonology in dyslexic children. Journal of Child Psychology and Psychiatryand Allied Disciplines, 44(5), 712–722.

ouibah, A., & Taft, M. (2001). The role of syllabic structure in French visual wordrecognition. Memory & Cognition, 29(2), 373–381.

aß, H., Wittchen, H.-U., & Zaudig, M. (2003). Diagnostisches und statistisches Manualpsychischer Störungen (DSM-IV-TR) [Diagnostic statistic manual of mental disor-ders]. Göttingen: Hogrefe.

eidman, L. J. (2006). Neuropsychological functioning in people with ADHD acrossthe lifespan. Clinical Psychology Review, 26, 466–485.

eymour, P. H. K., Aro, M., Erskine, J. M., & COST Action A8 Network. (2003). Founda-tion literacy acquisition in European orthographies. British Journal of Psychology,94, 143–174.

halev, R. S., Gross-Tsur, V., & Manor, O. (1997). Neuropsychological aspects of devel-opmental dyscalculia. Mathematical Cognition, 3, 105–170.

hare, D. L. (2008). On the anglocentricities of current reading research and prac-

Please cite this article in press as: Thaler, V., et al. Different behavioralattentional deficits during single word reading. Neuropsychologia (200

tice: The perils of overreliance on an “outlier” orthography. Psychological Bulletin,134(4), 584–615.

tenneken, P., Conrad, M., & Hutzler, F. (2005). Frequency effects with visual wordsand syllables in a dyslexic reader. Behavioural Neurology, 16(2–3), 103–117.

vensson, I., & Jacobson, C. (2005). How persistent are phonological difficulties? Alongitudinal study of reading retarded children. Dyslexia, 12, 3–20.

PRESSgia xxx (2009) xxx–xxx

Taft, M., & Forster, K. I. (1975). Lexical storage and retrieval of prefixed words. Journalof Verbal Learning and Verbal Behavior, 14, 638–647.

Thaler, V., Ebner, E., Wimmer, H., & Landerl, K. (2004). Training fluency in dysfluentreaders with high reading accuracy: Word specific effects but low transfer tountrained words. Annals of Dyslexia, 54, 89–113.

Thaler, V., Heine, A., Engl, V., & Jacobs, A. M. (in preparation). Effects of consonantcluster density during reading single words and non-words of different length:A differentiation of dyslexic and normal reading children.

Treiman, R. (1991). Children’s spelling errors on syllable-initial consonant clusters.Journal of Educational Psychology, 83, 346–360.

Treiman, R., Goswami, U., & Bruck, M. (1990). Not all nonwords are alike: Implicationsfor reading development and theory. Memory & Cognition, 18(6), 559–567.

Valdois, S., Carbonnel, S., Juphard, A., Baciu, M., Ans, B., Peyring, C., et al. (2006).Polysyllabic pseudo-word processing in reading and lexical decision: Convergingevidence from behavioural data, connectionist simulations and functional MRI.Brain Research, 1085, 149–162.

Van Daal, V., & van der Leij, A. (1999). Developmental dyslexia: Related to specific orgeneral deficits? Annals of Dyslexia, 49, 71–104.

van Meel, C. S., Heslenfeld, D. J., Oosterlaan, J., & Sergeant, J. A. (2007). Adaptivecontrol deficits in attention-deficit/hyperactivity disorder (ADHD): The role oferror processing. Psychiatry Research, 151, 211–220.

Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Specific readingdisability (dyslexia): What have we learned in the past four decades? Journal ofChild Psychology and Psychiatry, 45(1), 2–40.

Visser, T. A. W., & Besner, D. (2001). On the dominance of whole-word knowledge inreading aloud. Psychonomic Bulletin & Review, 8(3), 560–567.

Weekes, B. S. (1997). Differential effects of number of letters on word and nonwordnaming latency. The Quarterly Journal of Experimental Psychology Section A, 50(2),439–456.

Whiteley, H. E., & Walker, P. (1997). Multiletter units in visual word recognition:Direct activation by supraletter visual features. Visual Cognition, 4(1), 69–110.

Whitney, C. (2001). How the brain encodes the order of letters in a printed word:The SERIOL model and selective literature review. Psychonomic Bulletin & Review,8(2), 221–243.

Willcutt, E. G., Pennington, B. F., Olson, R. K., & DeFries, J. (2007). Under-standing Comorbidity: A Twin Study of Reading Disability and Attention-Deficit/Hyperactivity Disorder. American Journal of Medical Genetics Part B(Neuropsychiatric Genetics), 144B, 709–714.

Wilding, J. M. (2003). Attentional difficulties in children: Weakness in executivefunction or problems in coping with difficult tasks? British Journal of Psychology,94, 427–436.

Wilding, J., Pankhania, P., & Williams, A. (2007). Effects of speed and accuracy instruc-tions on performance in a visual search task by children with good or poorattention. British Journal of Psychology, 98, 127–139.

Wimmer, H. (1993). Characteristics of developmental dyslexia in a regular writingsystem. Applied Psycholinguistics, 14, 1–33.

Wimmer, H. (1996). The nonword reading deficit in developmental dyslexia: Evi-dence from children learning to read German. Journal of Experimental ChildPsychology, 61, 80–90.

Wimmer, H. (2006). Don’t neglect reading fluency!. Developmental Science, 9(5),447–448.

Wimmer, H., Mayringer, H., & Raberger, T. (1999). Reading and dual-task balanc-ing: Evidence against the automatization deficit explanation of developmentaldyslexia. Journal of Learning Disabilities, 32(5), 378–473.

Yu, J. W., Buka, S. L., McCormick, M. C., Fitzmaurice, G. M., & Indurkhya, A. (2006).Behavioral problems and the effects of early intervention on eight-year-old chil-dren with learning disabilities. Maternal and Child Health Journal, 10(4), 329–338.

Ziegler, J. C., & Goswami, U. (2005). Reading acquisition, developmental dyslexia, andskilled reading across languages: A psycholinguistic grain size theory. Psycholog-ical Bulletin, 131, 3–29.

Ziegler, J. C., & Goswami, U. (2006). Becoming literate in different languages: Similarproblems, different solutions. Developmental Science, 9(5), 429–436.

Ziegler, J. C., & Perry, C. (1998). No more problems in Coltheart’s neighborhood:Resolving neighborhood conflicts in the lexical decision task. Cognition, 68,B53–B62.

Ziegler, J. C., Perry, C., Ma-Wyatt, A., Ladner, D., & Schulte-Körne, G. (2003). Develop-mental dyslexia in different languages: Language-specific or universal? Journalof Experimental Child Psychology, 86, 169–193.

and eye movement patterns of dyslexic readers with and without9), doi:10.1016/j.neuropsychologia.2009.04.006

Zimmermann, P., Gordan, M., & Fimm, B. (2002). KITAP-Testbatterie zur Aufmerk-samkeitsprüfung für Kinder [KITAP-test battery for the assessment of attentionaldeficits in children]. Herzogenrath: Psytest.

Zoccolotti, P., De Luca, M., Di Pace, E., Gasperini, F., Judica, A., & Spinelli, D. (2005).The word length effect in early reading and in developmental dyslexia. Brain andLanguage, 93(3), 369–373.