Efficiency of visual information processing in children at-risk for dyslexia: Habituation of single-trial ERPs

www.elsevier.com/locate/b&l

Brain and Language 98 (2006) 319–331

Efficiency of visual information processing in children at-riskfor dyslexia: Habituation of single-trial ERPs

Anne G.F.M. Regtvoort *, Theo H. van Leeuwen, Reinoud D. Stoel, Aryan van der Leij

Faculty of Behavioural and Social Sciences, Department of Educational Sciences, University of Amsterdam, P.O. Box 94208,

1090 GE Amsterdam, The Netherlands

Accepted 1 June 2006Available online 25 July 2006

Abstract

To investigate underlying learning mechanisms in relation to the development of dyslexia, event-related potentials to visual standardswere recorded in five-year-old pre-reading children at-risk for familial dyslexia (n = 24) and their controls (n = 14). At the end of secondgrade the children aged 8 years were regrouped into three groups according to literacy level and risk factor. Single-trial analyses revealedN1 habituation in the normal-reading controls, but not in the normal-reading at-risks, and a N1 amplitude increase in the group of poor-reading at-risks and poor-reading controls. No P3 habituation was found in either group. The normal-reading at-risk group exhibited thelongest N1 and P3 latencies, possibly compensating for their reduced neuronal activity during initial information extraction. In contrast,the poor-reading group only showed prolonged P3, and their increase in (initial small) N1 amplitude together with normal N1 latencies,suggests inefficient processing in an early time window, which might explain automatisation difficulties in dyslexic readers.� 2006 Elsevier Inc. All rights reserved.

Keywords: Dyslexia; Habituation; Event-related potential; Single trial; Visual; N1; P3; Children

1. Introduction

It is internationally recognised that dyslexia is neurobio-logical in origin and characterised by slow and inaccuratereading as well as poor spelling. These difficulties are oftenunexpected in relation to other cognitive abilities and theprovision of effective classroom instruction (Goswami,2003; Lyon, Shaywitz, & Shaywitz, 2003). Up to 10% ofthe population shows mild to severe symptoms of dyslexia.In the case of a family-related history of dyslexia preva-lence is estimated between 33% and 50% (Pennington &Lefly, 2001). Having at least one dyslexic parent forms aconsiderable risk for a young child to develop dyslexia him-self (Gilger, Hanebuth, Smith, & Pennington, 1996). Overthe years a strong consensus has arisen among investigatorsthat reading is primarily a linguistic skill and that the cen-

0093-934X/$ - see front matter � 2006 Elsevier Inc. All rights reserved.

doi:10.1016/j.bandl.2006.06.006

* Corresponding author. Fax: +31 20 5251200.E-mail address: [email protected] (A.G.F.M. Regtvoort).

tral difficulty in dyslexia reflects a deficit within the lan-guage system, in particular a phonological deficit (e.g.,Lyon et al., 2003; Pennington, 1990; Ramus et al., 2003;Vellutino, Fletcher, Snowling, & Scanlon, 2004). Neuro-biological evidence supporting this hypothesis comes froma large amount of literature on reading disability in adultsand children, describing dysfunction in particular left hemi-sphere posterior brain systems (e.g., Helenius, Tarkiainen,Cornelissen, Hansen, & Salmelin, 1999; Paulesu et al.,2001; Rumsey, Andreason, Zametkin, & Aquino, 1992;Rumsey et al., 1997; Shaywitz et al., 2002; Silani et al.,2005; Simos, Breier, Fletcher, Bergman, & Papanicolaou,2000), considered to be critical in the development ofskilled, automatised reading (Lyon et al., 2003).

In comparison to the abilities of normal-reading chil-dren, it is evident that the reading and spelling of dyslexicchildren lack automaticity, in particular indicated by rela-tively slow response rates on measurements of (repeated)reading. For example, it has become clear that althoughfamiliar words may be processed accurately, the processing

mailto:[email protected]

320 A.G.F.M. Regtvoort et al. / Brain and Language 98 (2006) 319–331

rate stays relatively slow (Seymour, 1986; van der Leij &van Daal, 1999). Furthermore, the process of learninghow to read and spell lacks the normal characteristics ofskill automatisation (see for a review Adams, 1990). Skillsthat become automatised are performed more quickly, withless effort, autonomously or unintentionally, and with lessconscious awareness (Logan, 1997). Although it is com-monly assumed that dyslexic children have a specific learn-ing disorder concerning the domain of reading andspelling, the problem in mastering automaticity mightimply a learning problem of a more general nature than dif-ficulties in the learning of correspondences between lettersand constituent sounds per se.

Evidence for this line of hypothesizing comes from twostudies by Nicolson and Fawcett (2000) concerning long-term training in dyslexic children on a keyboard spatialtask and a choice reaction task. They showed that thedyslexic subjects acquired a normal ‘strength’ of automat-isation as measured by resistance to unlearning, andlong-term retention. In terms of speed and accuracy, how-ever, their automatised performance was of a lower quality.The learning curves of the dyslexic subjects in the abovetasks displayed slower and more error-prone initial perfor-mance leading to lower final skill levels as compared tosubjects without dyslexia. These ‘dyslexic’ rates of learningresemble the rates derived from repeated practice in thedomain of reading (Thaler, Ebner, Wimmer, & Landerl,2004; van der Leij & van Daal, 1990). Assuming that auto-maticity develops with learning (Logan, 1997) and thatautomaticity in children with dyslexia is of a lower qualityin terms of speed but not in strength (Nicolson, Fawcett, &Dean, 2001), how are we to explain this lower quality? Therelevance of this question is not restricted to childrenalready facing severe reading and spelling problems, butmay also hold for children genetically at-risk for develop-mental dyslexia, who are still in the pre-reading age andhave not developed any reading skill.

An approach that may be crucial in explaining the auto-matisation difficulties in dyslexic children is examining itspossible relation with underlying mechanisms of informa-tion processing. As all learning consists of linking new toold information and new to old abilities, the degree inwhich the outcome of that process can be called automa-tised is determined by the ability to process new and chang-ing information and to differentiate between relevant andirrelevant information. Of particular interest in this lightare the orienting response (OR) and habituation of theOR. As the OR is a mechanism controlling involuntarychanges in information-processing ability, leading to anincrease in alertness and attention, the functional signifi-cance of the OR is to help investigate the eliciting stimuliin greater detail (Spinks & Siddle, 1983). However, afterrepeated presentation of stimuli, decline or habituation ofthe OR is thought to take place because the updating ofthe neuronal model of the stimulus becomes automated(Ravden & Polich, 1998). All information contained withinthe stimulus is ‘learned’ and the brain no longer responds

to it. At a functional level habituation can be seen as an ele-mentary form of learning (Sambeth, Maes, Quian Quiroga,Van Rijn, & Coenen, 2004; Stephenson & Siddle, 1983).Assuming that rate of habituation reflects the level of effi-ciency of the brain to process information in a progressive-ly automated manner, habituation may reflect a rate oflearning. As a result of the automatised processing, thereis a reduced need for attentional resources, and the limitedprocessing capacity of short-term store becomes availablefor evaluation of novel or significant stimuli.

Both orienting and habituation can be defined by behav-ioural (overt) and physiological (covert) responses. Gaininginsight into the latter responses calls for neurophysiologicalinvestigations, for example of the brain mechanismsinvolved, by registration of Event-Related Potentials(ERPs). ERPs are relatively small changes in the electroen-cephalogram (EEG) in time locked to a given event or stim-ulus. ERP components of relevance to the dynamics of theOR can provide evidence for the existence, order and tim-ing of independent processes in information processing (seeKenemans, 1990; Polich & Kok, 1995; Sokolov, Spinks,Naatanen, & Lyytinen, 2002, for extensive reviews). Giventhat habituation usually takes place within 10 to 20 trials(Kenemans, 1990; Verbaten, Roelofs, Sjouw, & Slangen,1986) and traditional ERPs require many artifact free tri-als, a technique is needed that allows deriving single-trialERPs. There are only few habituation studies using sin-gle-trial techniques that compute single ERPs within ahabituation series of subsequent trials, such as orthogonalpolynomial trend analysis (OPTA, see Woestenburg,Verbaten, Van Hees, & Slangen, 1983) or a dynamic factormodel (Molenaar, 1994). In a sample of young adults,Kenemans, Verbaten, Roelofs, and Slangen (1989)observed an amplitude decrease of single-trial ERP compo-nents to a sequence of identical neutral as well as task-rel-evant visual stimuli for a non-specific N1 and for aposteriorly distributed P3. The N1, thought to representthe initial extraction of information from the sensory anal-ysis of the stimulus, is in particular affected by stimulusnovelty, and shows its sensitivity to subsequent repetitionof a stimulus with rapid decrement (e.g., Kenemans,1990; Naatanen & Picton, 1987). Although the P3 is asso-ciated with higher level processing, upon for instance, stim-ulus relevance, with short term repetition the P3 appears todecrease as rapidly as well (e.g., Kenemans, 1990; Polich &Kok, 1995). Habituation research including various single-trial procedures shows that occurrence and rate of habitu-ation are heavily dependent on stimulus type, complexityand relevance, task and inter-stimulus interval (ISI) condi-tions, and subjects (e.g., Cohen & Polich, 1997; Geisler &Polich, 1994; Lew & Polich, 1993; Ravden & Polich,1998; Riggins & Polich, 2002; Verbaten et al., 1986).Although the evidence about atypical brain- and skill-relat-ed development of young children with a family history ofdyslexia is still growing (e.g., Elbro & Petersen, 2004; Kos-ter et al., 2005; Lyytinen et al., 2005; Maurer, Bucher,Brem, & Brandeis, 2003; Snowling, Gallagher, & Frith,

A.G.F.M. Regtvoort et al. / Brain and Language 98 (2006) 319–331 321

2003; van Leeuwen et al., 2006), to our knowledge ERPhabituation has not been studied in pre-reading at-risk chil-dren. The scarce evidence regarding differences betweengroups of older children appears to be inconsistent. Forexample, no differences in habituation to simple visualand auditory stimuli were found between normally learningand learning disabled groups (Hirano, Russell, Ornitz, &Liu, 1996; John et al., 1989). Conversely, absence of habit-uation of the N1 to auditory stimuli was reported in a smallgroup of children with fragile X syndrome in contrast tohabituation shown by the control children (Castren,Paakkonen, Tarkka, Ryynanen, & Partanen, 2003).

Rate of learning depends heavily on the speed at whichnew and changing information is processed. Assuming thatthe latencies of ERP components reflect the processing timerequired for information extraction, stimulus evaluationand categorization, the lower quality of automatised per-formance in dyslexic children may be reflected in prolongedlatencies of N1 and P3. Of interest to the present study areany delays in non-linguistic information processing, andERP literature of relevance provides some evidence forspeed differences between dyslexic and normal-reading sub-jects. Prolonged lower-level (visual) processing in dyslexicreaders is, however, mostly reported to targets or low-probability stimuli (Breznitz & Meyler, 2003; Fawcettet al., 1993; Silva Pereyra et al., 2003; Taylor & Keenan,1990). Conversely, studies using non-targets in passivetasks report no (Bernal et al., 2000; Stelmack, Rourke, &van der Vlugt, 1995) or inconsistent evidence for prolongedprocessing (see for a review, Dool, Stelmack, & Rourke,1993). Thus, it appears that dyslexics are in particulardelayed in the triggering of involuntary attention followingstimulus change, possible the result of weakened attentioncapture (Hari & Renvall, 2001), or in the alternationbetween different cognitive processes related to stimulusrelevance (Breznitz & Meyler, 2003). Note, however, thatabove findings were obtained in adults or children olderthan the subjects in the present study, and by averagingERPs over series, rather than based upon single trialtechniques.

The present paper is an attempt to investigate the role ofunderlying learning mechanisms in relation to the develop-ment of dyslexia by assessing the neural efficacy of thebrain in genetically at-risk children who are still in thepre-reading phase. In this group of children the presenceof deficient underlying learning mechanisms at the age offive years may signify the onset of difficulties in the acqui-sition of reading and spelling. Given that a substantialnumber of genetically at-risk children will not developreading problems (Elbro & Petersen, 2004; Hindsonet al., 2005; Snowling et al., 2003), it is of great interest,theoretically and practically, to establish whether the at-risk children, who are facing reading difficulties, can be dis-tinguished from the at-risk children who manage to learnto read and spell fluently. Therefore, we will study the effi-ciency of the processing of novel information in five-year-old children at-risk and not-at-risk, before they have started

formal reading-instruction. On the basis of their literacyskills two years later, children are either assigned to a groupof poor-reading children or to one of the two groups ofnormal-reading children. This way, both the at-risk factorand the reading and spelling ability can be taken intoaccount in explaining possible differences between the threegroups of children. A similar approach, that is, the record-ing of ERPs in young children before the emergence of read-ing disabilities has been used previously in normal andat-risk populations. For instance, Molfese (2000) reportedthat auditory ERPs recorded at birth discriminated with81.3% accuracy among groups of eight-year-old normal,poor and dyslexic readers. Lyytinen et al. (2005) reportedpredictive correlations between ERPs to sounds at sixmonths in at-risk children and their later literacy skills justbefore school entry.

While employing a passive non-attending set-up, in thepresent study short-term habituation and speed of process-ing to visual stimuli over a series of subsequent trials willserve as the qualitative and temporal measures of learningas indexed by ERP components N1 and P3. The combina-tion with a single-trial technique assures recording oftrial-to-trial variation in each subject (Molenaar, 1994).Without disputing the importance of auditory-phonologi-cal processing ability in reading acquisition, we will focuson underlying learning within the visual modality, to stressthat if the learning disorder in dyslexic children implicatesnot only a deficit in phonological processing, but also inautomatisation, it is less specific than commonly assumed.In that case learning deficiencies need not necessarily berestricted to the auditory modality.

We presume that the ability to process new informationmore efficiently and intensively will result in learning at ahigher rate. Thus, with respect to the effects of stimulusrepetition on amplitude of the ERP components we expecta normal pattern of habituation, that is, a sharp decline ofamplitude over the first trials, in particular for the N1 com-ponent, to be shown by the children identified as normalreaders. In contrast, poor-reading children are expectedto be less susceptible to habituation, based on the assump-tion that they are ‘slower’ learners of unfamiliar stimuluscontent. Since children’s processing is assessed with a seriesof identical stimuli that bear no task relevance, longerlatencies in the poor readers may be expected on the veryfirst trial(s) only, due to a delayed attentional shift. Wedo not expect substantial group differences over the subse-quent trials in timing of early processing and stimulus eval-uation, once non-task relevant attention is engaged.

2. Methods

2.1. Subjects

Twenty-seven children from families with a history ofdyslexia (mean age 5 years; 6 months, range 59–73 months,14 boys, 13 girls) and 16 control children (mean age 5years, 4 months, range 58–69 months; 9 boys, 7 girls)


participated in the study. All children were recruited in thebeginning of the second year of kindergarten, but some ofthem attended kindergarten for an extra year. In the Neth-erlands, the primary educational system includes kinder-garten that spans two years. The majority of Dutchchildren enter school at age four. Formal schooling startsin Grade 1 at the age of six. The children were pupils of ele-mentary schools in the area of Amsterdam, except for someof the at-risk children who visited an elementary schooloutside this area. All families participated voluntarily inresponse to handouts distributed by the kindergartenteachers. On enrollment, parents were requested to fillout an information form, including data on parental educa-tion level, and the medical history of the child. None of thechildren were showing serious emotional disturbances orattentional problems, according to their parents. All chil-dren had normal hearing, but some of the control childrenwere wearing glasses, mainly for correcting problems withnear-sightedness.

Selection of the subjects at-risk or as controls was basedon self-report of the parents confirmed by a questionnaireand the parent’s performance on a screening test. This testincluded two standardized measures of speeded decodingof single words and pseudowords, the One-Minute-Test(OMT) (Brus & Voeten, 1999) and the Klepel (van denBos, lutje Spelberg, Scheepstra, & de Vries, 1999), as wellas the Wechsler Adult Intelligence Scale (WAIS) subtestSimilarities (Wechsler, 1997) measuring concept formation.For inclusion in the at-risk group the reading performanceof (at least one of) the parents had to fall below or at the10th percentile on either one of the reading tests or belowor at the 20th percentile on both reading tests. Also leadingto inclusion was a discrepancy score which represented adifference of at least 60 percentiles between a high scoreon the WAIS subtest Similarities and the score on eitherone of the reading tests (Koster et al., 2005; Kuijperset al., 2003). For inclusion in the control group reading per-formance of both parents had to be on or above the 40thpercentile on both reading tests.

Only children for whom ERP and behavioural data wereavailable were included in the data analysis. Consequently,three at-risk children who dropped out prior to the acqui-sition of literacy data were excluded (attrition at-riskgroup = 11%). Furthermore, artefact rejection resulted inan insufficient number of trials in the habituation seriesfor two control children and they were also excluded fromfurther data analyses (attrition control group = 13%). For24 at-risk children (13 boys and 11 girls) and 14 not-at-riskcontrol children (8 boys and 6 girls) habituation data mea-sured at the beginning of the second year of kindergartenand literacy data measured two years later were available.

Preceding single trial data analysis, the remaining chil-dren in the two samples were regrouped taking intoaccount both the risk factor and children’s reading andspelling ability at the end of Grade 2. The five standardizedmeasures we used to determine literacy proficiency werethree decoding speed tests consisting of mono- and polysyl-

labic words (Verhoeven, 1995), a decoding speed test con-sisting of mono- and polysyllabic pseudowords (van denBos et al., 1999), and a spelling test (van den Bosch, Gilli-jns, Krom, & Moelands, 1993) consisting of mono- andpolysyllabic words with orthographic patterns of variouscomplexities. Reclassification of the children resulted intothree groups consisting of poor-reading at-risk and controlchildren, normal-reading at-risk children, and normal-reading control children. The group of poor readers(n = 11, mean age 5 years; 6 months, range 59–72 months,6 boys, 5 girls) consisted of 9 children at-risk and 2 controlchildren with scores below the 25th percentile on at leasttwo of the word decoding tests, and scores below the25th percentile on either pseudoword decoding or wordspelling. The normal-reading at-risk children (n = 15, meanage 5 years; 5 months, range 59–71 months, 9 boys, 6 girls);and the normal-reading control children (n = 12, mean age5 years, 4 months, range 58–69 months; 6 boys, 6 girls) con-sisted of the remaining at-risk and control children, respec-tively, with scores above the 25th percentile on all readingmeasures.

With respect to differences in parental performance onthe screening tests, parents of the normal-reading controlchildren performed better on the reading measures, butnot on verbal competence as compared to the two poor-reading and normal-reading at-risk groups. No differenceswere found between these latter groups on any of the selec-tion measures. According to the selection criteria, 36% ofthe parents in the poor-reading group (including those thatwere selected as control children) and 33% of the parents inthe normal-reading at-risk group scored below or equal tothe 10th percentile on both reading measures; 18%, respec-tively 27% of the parents scored below or equal to the 20thpercentile on both measures and/or below or equal to the10th percentile on one of the measures. Twenty-seven per-cent of the poor-reading group and 40% of the normal-reading at-risk group were included on the basis of adiscrepancy-score. Dyslexic problems in relatives of theaffected parents were reported in 64% and 60% of thepoor-reading group and the normal-reading at-risk group,respectively. The three groups did not significantly differ ingender or age, nor in cognitive ability, as assessed with twostandardized measures, a test of receptive vocabulary(Verhoeven & Vermeer, 2001) and the Coloured Progres-sive Matrices (Raven, Court, & Raven, 1984), measuringnon-verbal intelligence. Halfway through the second yearof kindergarten, all subjects scored above the 10th percen-tile of receptive vocabulary, and one year later all subjects(except for one poor-reading child) performed within orabove the normal range (between the 25th and 75th percen-tiles) on non-verbal intelligence. Since the child with abelow average score on non-verbal intelligence scored inthe average range in the receptive vocabulary measure,we refrained from excluding it from the analyses. Withrespect to reading and spelling performance, regardless ofrisk factor, normal-reading children performed better thanpoor-reading children on all literacy measures. On none of

Table 1Mean scores (M) and standard deviations (SD) for poor-reading children, normal-reading at-risk (AR) children and normal-reading control (C) childrenon Grade 2 measures

Measures Poor n = 11 Normal AR n = 15 Normal C n = 12 F (2,35)

M SD M SD M SD

Speeded decodingWordlist 1 (max = 150) 41.0 12.9 76.9 10.0 78.9 13.9 35.78*** 1 < 2 = 3Wordlist 2 (max = 150) 26.7 10.3 71.2 12.7 74.2 14.1 52.41*** 1 < 2 = 3Wordlist 3 (max = 120) 16.1 9.0 54.9 13.5 54.7 14.9 35.50*** 1 < 2 = 3Speeded decoding pseudowords (max = 116) 15.5 4.6 41.7 10.3 49.3 10.3 44.26*** 1 < 2 = 3Spelling words (max = 38) 21.4 7.0 32.3 5.3 34.0 6.2 14.68*** 1 < 2 = 3

*** p < .001.


the measures differences were found between the normal-reading at-risk and normal-reading control group. Thedescriptive statistics for the literacy measures appear inTable 1.

2.2. Apparatus

A 32-channel electrocap was used employing sinteredAg/AgCL electrodes in a 10–20 system montage includinglocations (Oz, O1/2, Pz, P3/4, P7/8, CPz, CP3/4, TP7/8,Cz, C3/4, T7/8, FCz, FC3/4, FT7/8, Fz, F3/4, F7/8, andFp1/2). Additional electrodes were used for recording thevertical electro-oculogram (VEOG; above and below theleft eye) and the horizontal electro-oculogram (HEOG; atthe outer canthus of each eye). Linked ear electrodes servedas the inactive reference. Electrocortical activity wasrecorded with 500 Hz per channel and filter settings of0.1–70 Hz (Synamps model 5083, Neuroscan Inc.). Imped-ances were kept below 20 kX (Ferree, Luu, Russell, &Tucker, 2001).

2.3. Stimuli

Stimuli were two abstract, black and white patterns,with equal luminance (same number of black and whitepixels), and an information value of 26 bits for the stan-dard and 60 bits for the deviant, as has been determinedemploying a procedure developed by Attneave (1954),and similar to the ones used by Kenemans et al. (1989).They were presented in the centre of a computer screen,with a presentation time of 1000 ms. To facilitate propercentral fixation, a fixation cross was presented in theinter-stimulus interval. The screen was positioned in frontof the subject at a distance of 100 cm from the subject’seyes. Stimulus presentation and sampling procedures werecontrolled by a standard PC. A total of 211 stimuli werepresented in a single block, in which the first 14 presenta-tions were identical standards, followed by a deviant onthe 15th trial. On the remaining 196 trials, (13) additionaldeviants were pseudo-randomly presented with a 6.6%probability. Following this procedure, the first 14 ‘habitu-ation’ trials could be assessed separately from following tri-als, and here only data pertaining to this first series will be

presented. A variable ISI was used (to accommodate addi-tional conditions, not presented here with increasing taskrequirements) with a mean of 1500 ms and a range between1000 and 2000 ms.

2.4. Procedure

The ERP recording session took about 1–112

h. An adult,in most cases one of the parents, accompanied the child.Upon arrival, the child was put at ease and familiarizedwith the procedure and the methods employed which start-ed only when the child was ready to cooperate. To distractattention during preparation of the electrocap the child wasgiven a picture book. Recording took place in an adjacentroom with dimmed lights and the door closed. Subjectsreceived task instructions as soon as they were seated in achair that was positioned in front of the computer screen.They were told that pictures would be presented, but thatthey merely had to watch them by looking at the middleof the screen, and that they did not have to pay particularattention to them. The children were instructed to maintainfixation on a small black cross at the centre of the screenappearing between presentations, to sit as still as possibleand to prevent eye blinking during presentation of the stim-ulus as much as possible. During the experiment an assis-tant sat with the child to assure that the child remainedfixating the screen centrally. Parents were not allowed tobe present in the recording room.

2.5. EEG processing and scoring

2.5.1. ERPs

The EEG was bandpass filtered (.5–20 Hz, 48 dB/octave) and eye blinks were corrected according to a spatialfilter transform based on a linear derivation approach (seeUser Guide: Vol. II; Neuroscan Inc.). Subsequently, arte-facts exceeding 100 lV in any channel were automaticallyrejected and epochs (starting 100 ms before stimulus onsetand lasting 1024 ms) were obtained for the first 14 stan-dards (referred to as habituation trials). Individual habitu-ation trials were subsequently inspected for remainingartefacts and occasional missing trials were replaced byaveraging preceding and following trials, to ensure 14 trials


for each subject. Grand average ERPs (across the habitua-tion series) yielded two distinct Global Field Power peaks(microstates; Lehmann), one for the N1 (130–260 ms)and one for P3 (450–850) with peaks at Cz and Pz, respec-tively. Traditionally, N1 and P3 are quantified at theseelectrode locations, so we decided to restrict the furtheranalysis to these sites. Subsequently, single-trial ERPs weredetermined by a dynamic factor analysis technique(Molenaar, 1994). Next, for each resulting single trialERP, N1 amplitude and latency were scored at Cz as thelargest negative going peak within a latency windowbetween 130 and 260 ms, relative to the largest positivegoing peak within 100 ms preceding the N1. Similarly, P3amplitude and latency were scored at Pz as the largestpositive going peak within a latency window between 450and 850 ms, relative to the largest negative going peak pre-ceding it.

2.6. Statistical analysis

The Linear Mixed Model procedure as implemented inthe statistical software package SPSS was used to testwhether the groups differed over time for the habituationstimuli. The Linear Mixed Model (see Verbeke & Mole-nberghs, 2000) is essentially similar to a standard RepeatedMeasures ANOVA but offers more flexibility in modelspecification. Dummy parameters were used to representthe groups, where one dummy parameter is sufficient torepresent two groups, and the two dummy parametersare sufficient to represent three groups. A different param-eterization of the dummy parameters was used to test spe-cific hypotheses on group differences, similar to the posthoc group comparisons of standard ANOVA. ERP com-ponents N1 and P3 were analysed in two linear mixedmodel designs (fixed effects only) by means of maximum

Fig. 1. Mean single-trial ERPs of N1 at Cz for poor readers (solid), normal at-rand 13.

likelihood (ML) estimation. In the first design the depen-dent variables were measured at 14 occasions and threegroups of readers poor-reading children (Poor), normal-reading at-risk children: (Normal-AR) and normal-readingcontrol children (Normal-C). In each of the three groupsintercepts and linear slopes were estimated for amplitudeand latency. The intercepts were defined as the status atthe first trial, and the slopes represent the linear changeacross trials. For testing the expectations regarding groupdifferences in the intercepts of latency, and in the interceptsand slopes of amplitude, the first design contains the fol-lowing set of dummies: Normal-C versus Poor, Normal-C versus Normal-AR. The second design resembles the firstin all major aspects with the exception of the includedgroups (poor-reading and normal-reading at-risk) repre-sented by the dummy: Poor versus Normal-AR. Maineffects and interactions were tested by means of the Likeli-hood Ratio (LR) test (i.e., a v2-difference test), the dummyparameters are tested by means of the t-test (e.g., estimate/standard error). We started with a full model containing allmain and interaction effects, and subsequently constrainedthe model further. For each ERP component the estimatesof the most parsimonious model will be presented. Then,effects for group are reported on intercept (for amplitudeand latency), and lastly, effects for group are reported onslope (for amplitude only). Following the presentation ofthe first design the same procedure was used for the second.

3. Results

Mean single-trial ERPs at Cz (where N1 was scored)and Pz (where P3 was scored) are shown in Figs. 1 and2, for Trials 1, 2, 5, 8, 10 and 13. Mean intercepts and linearslopes of N1 and P3 latency and amplitude, and mean lin-ear slopes of N1 and P3 amplitude are listed in Table 2.

isk readers (dashed) and normal control readers (dotted): trials 1, 2, 5, 8, 10

Fig. 2. Mean single-trial ERPs of P3 at Pz for poor readers (solid), normal at-risk readers (dashed) and normal control readers (dotted): trials 1, 2, 5, 8, 10and 13.

Table 2N1 and P3 intercepts (for latency and amplitude) and slopes (for amplitude), per group

N1 P3

Poor Normal AR Normal C Poor Normal AR Normal C

Interc. Slope Interc. Slope Interc. Slope Interc. Slope Interc. Slope Interc. Slope

Latency (ms) 183.11a 194.58 186.99a 617.51b 639.42b 573.98Amplitude (lV) �7.89a �0.65 �9.74a �0.11 �16.37 0.43 14.42b 0.33c 13.13b 0.06c 13.46b �0.07c

Note: Values that share the same subscript do not differ.

Fig. 3. Intercepts and slopes of N1 amplitude for poor readers (solid),normal at-risk readers (dashed) and normal control readers (dotted) fortrials 1 to 14.


3.1. Amplitude: Intercept and slope

N1 Starting with the full model, within the first designthere were both significant main and interaction effectsfor the intercept and the slope for N1 amplitude. The fullmodel fitted the data significantly better than a restrictedmodel without interaction effects for the slope (v2(2) =15.92, p = .000). Given the full model, the dummyparameters representing the main effect for the interceptshowed a significantly less negative intercept for Poor(t(248.61) = 3.92, p = .000), and Normal-AR (t(248.61) =3.31, p = .001) as compared to Normal-C. With regard tolinear slope, the dummy parameters showed that Poor(t(261.36) = �4.05, p = .000), and Normal-AR (t(261.36) =�2.17, p = .031) differed significantly from the Normal-C.As shown in Fig. 3, in Normal-C N1 amplitude had a morenegative initial level and gradually decreased in negativityover trials, whereas in Poor amplitude increased stronglyover trials. In Normal-AR N1 amplitude neither decreasednor increased in negativity. The second design showed alsosignificant main and interaction effects for the intercept andthe slope and the full model fitted the data significantlybetter than a restricted model without the interaction effectfor slope (v2(1) = 7.20, p = .007). The dummy parametersrepresenting the effects for group and slope showed that

the groups did not differ significantly for the intercept(t(148.09) = �1.53, p = .128), but that Normal-AR showeda less negative slope for N1 amplitude than Poor(t(189.15) = 2.73, p = .007).

P3 Within the first design, no main and interactioneffects for the intercept or for the slope were found forP3 amplitude. The full model and the most parsimoniousmodel appeared not to differ significantly (v2(5) = 2.86,

Fig. 4. Intercepts and slopes of P3 amplitude for poor readers (solid),normal at-risk readers (dashed) and normal control readers (dotted) fortrials 1 to 14.

Fig. 6. Intercepts of P3 latency for poor readers (solid), normal at-riskreaders (dashed) and normal control readers (dotted) for trials 1 to 14.


p = .72), indicating that there were no differences betweenthe three groups for the intercept and the slope, seeFig. 4. Within the second design there was no significantdifference between the full and the most parsimoniousmodel either (v2(3) = 2.24, p = .52). Poor and Normal-AR did also not differ for the intercept and the slope ofP3 amplitude.

3.2. Latency: Intercept

N1 Starting with the full model within the first design,no main and interaction effects for slope were found forN1 latency. That is, removing these effects from the modeldid not lead to a significant deterioration in model fit(v2(3) = 2.68, p = .44). Given this restricted model, themain effect for group on intercept was significantly differentfrom zero (v2(2) = 12.03, p = .002). The dummy parame-ters representing this main effect showed a significantlyhigher N1 intercept for Normal-AR as compared toNormal-C (t(528.02) = 2.33, p = .020), as can be seen inFig. 5. No significant difference was found betweenNormal-C and Poor. Within the second design no mainand interaction effects for slope were found either, andthese terms could be removed from the model(v2(2) = 2.68, p = .26). Given this restricted model, grouphad a significant effect on the intercept (v2(1) = 12.21,

Fig. 5. Intercepts of N1 latency for poor readers (solid), normal at-riskreaders (dashed) and normal control readers (dotted) for trials 1 to 14.

p = .001). The dummy parameter representing this maineffect showed a significantly higher intercept for Normal-AR as compared to Poor (t(355.58) = 3.64, p = .000).

P3 within the first design no main and interaction effectsfor slope were found for P3 latency (v2(3) = 1.41, p = .70).Given the restricted model, group had a significant effecton the intercept (v2(2) = 32.80, p = .000). The dummyparameters representing this main effect showed a signifi-cantly higher intercept for Poor (t(527.11) = 3.57,p = .000), and Normal-AR (t(527.11) = 5.79, p = .000) ascompared to Normal-C, see Fig. 6. Within the seconddesign no main and interaction effects for the slope andthe intercept were found (v2(3) = 3.87, p = .28). Given thismost parsimonious model, there was no effect for group onthe intercept. Poor and Normal-AR did not differ signifi-cantly for the intercept of P3 latency.

4. Discussion

4.1. Effects of stimulus repetition

The normal-reading control children showed a N1amplitude decrease over trials, whereas such habituationwas not present in the normal-reading at-risk childrenand the poor-reading children. In contrast, the P3 compo-nent appeared not to be sensitive to any form of habitua-tion, nor in the normal-reading control children andneither in the other groups. Interestingly, for N1 amplitudethe effects of stimulus repetition appeared to turn out quitedifferent for the three groups. The assumption that the nor-mal-reading at-risk children would show habituation to thesame degree as the normal-reading control group was notsupported by our data. However, the poor-reading childrendiffered in an even more striking manner from the group ofnormal-reading control children, because they displayed anincrease of N1 amplitude over trials. Given that N1 ampli-tude may also reflect endogenous processing (Taylor,Chevalier, & Lobaugh, 2003) and is susceptible to attentionlevel (Naatanen & Picton, 1987; Vogel & Luck, 2000), theabnormal strong covert orienting in the poor-readingchildren towards non-task relevant stimuli could reflect a


subjective interpretation of task relevance, despite the useof the standard instruction to merely watch the stimuli.Task uncertainty may have let the poor-reading childrengradually develop some kind of expectancy about the stim-uli to come (see also discussion below about P3 amplitude).It is unclear, however, why this effect was absent in the twoother groups.

Another remarkable group difference concerned the sizeof the intercept, an indication of the amplitude on the veryfirst trial. In contrast to the normal-reading controls, thepoor-reading children as well as the normal-reading at-riskchildren showed a much higher, that is, less negative initialamplitude. Following the instruction to direct the attentionto the screen, the poor-reading and normal-reading at-risksubjects may have been slower in shifting their attentionalfocus (Dykman, Ackerman, Holcomb, & Boudreau, 1983;Hari & Renvall, 2001) to the first uncued stimulus in theseries. Only recently, poor control of spatial attention shift-ing was also reported in a group of dyslexic students(Wijers, Been, & Romkes, 2005). This slower directing ofattention towards the stimulus location may be related totheir lower levels of orienting. In combination with theensuing perception of a fast sequence of stimuli, this mayhave caused a subjective experience of increased stimulusintensity in the poor-reading children. This concept wasraised before by Castren et al. (2003) to explain anenhanced N1 to auditory standards in a oddball paradigmand the absence of habituation to a train of stimuli in asmall group of children with fragile X syndrome. Theseauthors suggested that both findings could be caused byexcessive responses in neurons leading to changes in alert-ness as well as in general attentiveness and heightened stateof arousal. This conjecture may be further corroborated byevidence that increasing stimulus intensity has the effect ofaugmenting P1N1 amplitude slopes at midline sites, andthat this pattern of sensory modulation can be seen inthe majority of children (Carrillo de la Pena, Holguin, Cor-ral, & Cadaveira, 1999). Thus, the possibility that an incli-nation to respond with increased excitability to repeatedstimuli played a part in early processing differences betweenthe groups of poor- and normal-reading children cannot beruled out. However, in the last section we will discuss theincrement of N1 amplitude over trials in poor-reading chil-dren in relation to their inability to acquire reading andspelling fluency.

The present absence of a P3 amplitude effect suggestseither that we cannot always assume rapid decrement uponstimulus repetition for this component in young children,or that we did not control adequately for additional situa-tional or cognitive interference. The trial-to-trial variationin neural responding might reflect a varying expectancyabout the stimuli to come (Sambeth et al., 2004). Quite afew of the present subjects in fact inquired if or when other‘pictures’ would be presented. Then, with every new presen-tation a kind of mismatch may have occurred betweenanticipation and external event, which in turn may havegiven subjective relevance to the standard stimulus. Also,

it may be assumed that ERPs to stimuli presented at vari-able ISIs, as in the present study (mandatory for the taskrelevant blocks, not presented here), do less readily habitu-ate due to increased uncertainty about temporal probabil-ity (Graham, 1997; Kenemans, 1990; Mallardi, 1979;Papanicolaou, Loring, & Eisenberg, 1985). Moreover,amplitude decrement to visual as opposed to auditory stim-uli has been reported to be slower, unless many successivetrial blocks are presented, because visual processingrequires a greater utilization of attentional resources(Geisler & Polich, 1994; Ravden & Polich, 1998). Neverthe-less, the application of a single-trial technique enabled us toidentify N1 habituation in the normal-reading controls,thus revealing groups differences in trial-dependent effectsof stimulus repetition, which otherwise would have goneunnoticed in traditional grand averages, as nicely demon-strated by the fact that overall group means to all (197)standards presented were comparable.

4.2. Effects of speed of processing

The latency of the P3 peak amplitude (as an index of thechildren’s stimulus evaluation time (Polich, Ladish, &Burns, 1990)) appeared to be longer in the poor-readingchildren in comparison to the normal-reading controls. Inparticular the normal-reading at-risk children, however,displayed longer latencies for the N1 and the P3. For bothcomponents, latencies are age related with values decreas-ing with age (Fuchigami et al., 1993; Polich et al., 1990;Taylor et al., 2003), but the developmental changes thatare observed for P3 latencies are also related to changesin memory capacity (Polich et al., 1990). Although it seemsnot very likely that age related maturational processes wereinvolved in the longer latencies of the poor-reading andnormal-reading at-risk children, being in the same agerange does not necessarily mean that their neuronal matu-ration was similar. Regarding differences in memory func-tion, recently a relation between risk of dyslexia andpoorer, albeit verbal, memory skills at five years of agewas reported in an (follow-up) ERP study with newbornsat genetic risk of dyslexia (Guttorm et al., 2005).

Task difficulty and/or subjectively perceived complexityare known to influence processing times (Breznitz & Mey-ler, 2003). So, the prolonged processing of non-targetsunder passive conditions, might either imply that stimulusprocessing per se proved to be more taxing for the poor-reading and the normal-reading at-risk subjects, due to alagging neuronal maturation or a reduced memory capaci-ty, or that the new situation instigated a more prolongedextracting of stimulus information to promote processingand analysing. The normal-reading at-risk children’s delaysin both the earlier and later stage of processing show thattheir passively processing of incoming stimuli as well astheir more controlled stimulus processing were affected.The involuntary tendency towards a lower speed of pro-cessing in this group might be seen as an adaptation, thatis, as a kind of trade-off for their relatively (i.e., compared


to normal-reading controls) reduced early neuronal activityto ensure a better encoding of stimulus occurrence and con-tent. Given that the poor-reading children were onlydelayed in the later stage of processing, speed of processingamong poor-reading children seems in particular be sloweddown in evaluating information, but not in the modelling(of occurrence) of novel information. Thus, unlike the nor-mal-reading at-risk children, poor-reading children did notadapt their early timing to their (initial) reduced brainactivity. So, speed of processing might signal a deficientunderlying factor affecting learning in children at-risk fordyslexia.

4.3. Early visual processing in the pre-reading phase andreading proficiency

The apparent increase in brain activity over 14 trialsshown by the poor-reading children in response to a repeat-edly presented stimulus may indicate a less efficient way ofprocessing new information. The poor-reading childrenneeded much more exposures to reach the (initial) levelof neuronal responding of the normal-reading control chil-dren. By contrast, in combination with their slower pro-cessing times the normal-reading at-risk children seemedto have mobilized a sufficient amount of early brain activityto deal with the stimulus presentation. Dissimilar to thepoor-reading children their initial small negative responseswere not followed by an increase over trials, although nodecrement occurred either. To clarify why adjustment ofinitial reduced brain activity in response to newly present-ed, unfamiliar but neutral information may account for theproblematic development of automatised reading in dyslex-ic children, we first need to address some theoretical con-siderations about automaticity acquisition.

In general, processing may occur without subjectiveawareness of the stimulus that produced it. However, oftenpre-attentive processing is not sufficient for the perceptualorganization of (visual) input. Attentive processing is need-ed to integrate the various features belonging to one object,otherwise object representation would be incoherent (Spe-kreijse, 2000). According to Logan, Taylor, and Etherton(1999) attentive processing is a necessary condition forautomaticity, that is, ‘what is learned during automatisa-tion depends on what is attended to and how attention isdeployed’ (p. 179). It has been suggested that the earlyN1 component signals a pre-attentive comparison stageof the OR, providing the informational basis for attention-al selection (Naatanen, 1990). Thus, the entrance of visualstimuli into subjective awareness might not only be theintentionally result of (self)-controlling, for example actingupon instructions to attend, but also the result of an invol-untary process, for example an evoked OR. Although pre-attentive processing is not to be considered a necessarycondition for automaticity (Logan, 1992), through attract-ing attention to novel or salient features it may benefit theacquisition of automaticity. In his instance theory of auto-matisation, Logan (1997) points out that in gaining auto-

maticity the number of presentations is less importantthan the number of times a stimulus is interpreted in a par-ticular way. This notion relates to one of the theories’ mainassumptions, namely that automaticity is processing basedon memory retrieval. Retrieval might be considered a racebetween different but equivalent memory traces or instancerepresentations encoded and stored separately in memory.The first trace to finish, providing either a correct or anincorrect response, directs performance. The more tracesthere are in memory the quicker a trace will be retrieved.This argument implies that the more traces signify a correctresponse, the greater the chance a subject will perform well.Thus, initial small amplitudes and relative normal process-ing times at an early stage of processing together with anincrease in negativity upon stimulus repetition, might sig-nal that upon the initial observation of (novel) informationinsufficient neuronal resources are being allocated, therebyfailing an effective attraction of attention to and encodingof distinguishing features. As a consequence, multipleobservations of the same aspects of information are neces-sary. Assuming that this inefficient dealing with novelinformation may result in storage of relative less, more dif-fuse or incorrect memory traces of form and detail, it fol-lows that memory retrieval and as a consequence,automatisation is hampered. If such initial inadequate pro-cessing of abstract visual information in poor-reading chil-dren is extended to the processing of novel orthographicinformation, it might explain why children with severe dys-lexia are able to obtain an adequate level of accuracy andeven to gain in speed on highly practised words, whereasno transfer takes place to non-practised words that bearorthographic resemblance (van der Leij & van Daal,1990). With enough practice, these children are able to relyon sufficient correct word-specific orthographic memorytraces. However, accuracy and speed problems are to beexpected with non-practised words, since in those casesdecoding will be based on retrieval and combining of sev-eral traces of sublexical orthographic information, presum-ably of lower quality. Under more complex readingconditions like oral reading, decoding of unfamiliar wordswill deteriorate even more.

Since the automatic processing of words is stronglyinfluenced by problems with phoneme awareness shownby a majority dyslexics, if not by all (Ramus et al., 2003),a deviant processing of novel (absolutely or relatively)information may not be restricted to the visual modality,operating in isolation, but may also hold for the auditory/phonological system. In this light it is very interesting thatdeviances in early phonological processing also have beenfound in infants with later reading difficulties (Espy,Molfese, Molfese, & Modglin, 2004), as well as in infantsat familial risk for dyslexia (see for a review, Lyytinenet al., 2005). Espy et al. (2004) reported variations in thedevelopment of the N1 component as the most evidentrelation between ERP responses to speech and non-speechstimuli between the ages of one and four years and subse-quent pseudoword reading proficiency at eight years, with


N1 amplitudes consistently more negative in less proficientreaders. Leppanen et al. (2002) found more positiveresponses (at about 190 ms) to sound onsets discriminatedbetween infants with and without familial risk and preli-minary analyses showed that these could be linked to betterlater phonological skills in both groups (Lyytinen et al.,2005). It is unclear, however, whether these positiveresponses reflect the same information-processing mecha-nisms as manifest in N1. Evidence for deviances in earlyphonological processing shown by adult dyslexics comesfrom two studies using magnetoencephalography (MEG).Helenius, Salmelin, Richardson, Leinonen, and Lyytinen(2002) reported an abnormally strong N1m response tothe beginning of a speech sound under actively listeningand ignored speech conditions as the most robust differencebetween dyslexic and normal-reading individuals. Strongenhancement of a pre-semantic N1m was also found in dys-lexic adults to spoken words within a sentence (Heleniuset al., 2002). However, it should be stressed that none ofthese studies directly addressed the issue of habituation.Nevertheless, together with our findings the presented evi-dence suggests that the pre-attentive level of informationprocessing, either visual or auditory, could be an importantindicator or predictor of (later) reading problems.

In summary, by studying covert processing of abstractvisual information we attempted to find evidence for dif-ferences in underlying learning mechanisms in childrengenetically at-risk for dyslexia, as possible early indica-tions of reading disability. We were able to show that ina mixed group of at-risk children particularly the oneswith a lower reading proficiency showed differences inallocation of brain resources. In combination with a dis-crepancy between relative fast early and slow later pro-cessing times this mode of early processing of novelinformation is suggestive of deficient learning at a basiclevel. That the initial processing of repeatedly presentedstimuli was similarly affected in the two poor-reading chil-dren who were not-at-risk, seems to support the associa-tion between efficiency of early processing and readingabilities. Although we realize that the implication of ourfindings is limited due to small sample sizes, we would liketo emphasize that especially the association between pas-sive responding to new information at an early stage ofprocessing and later literacy skills seems strong enoughto pursue as a line of investigation. An obvious suggestionfor future research is a replication of our habituation para-digm involving meaningless auditory as well as complexphonological stimulus processing to investigate whetherthe reported deficient early processing in poor-readingchildren are modality non-specific and/or related to prob-lems with phonological coding. Equally interesting we con-sider the question whether deficient early processing is to befound in young children only, as a possible precursor of dif-ficulties in learning to read and spell, or that older subjectsare similarly affected. If that is the case, inefficient informa-tion processing resulting in less automatised learning mightbe an underlying characteristic of dyslexia.

Acknowledgments

First of all we thank all participating children and theirparents. The study was carried out at the University ofAmsterdam as part of the Dutch intervention project ‘Ear-ly diagnosis and treatment of developmental dyslexia’ incooperation with the Radboud University of Nijmegenand the University of Groningen, as one of the three com-ponents of the national research programme ‘‘Identifyingthe core features of developmental dyslexia’’, initiatedand supported by the Netherlands Organization for Scien-tific Research (NWO). Finally, we thank Peter Molenaarand Bert van Beek (Department of Psychology, Universityof Amsterdam) for sharing the dynamic factor analysisbased single-trial technique, and support with its imple-mentation in the present study.

References

Adams, M. J. (1990). Beginning to read thinking and learning about print.Cambridge, Mass: MIT Press.

Attneave, F. (1954). Some informational aspects of visual perception.Psychological Review, 61(3), 183–193.

Bernal, J., Harmony, T., Rodriguez, M., Reyes, A., Yanez, G., Fernandez,T., et al. (2000). Auditory event-related potentials in poor readers.International Journal of Psychophysiology, 36(1), 11–23.

Breznitz, Z., & Meyler, A. (2003). Speed of lower-level auditory and visualprocessing as a basic factor in dyslexia: Electrophysiological evidence.Brain and Language, 85(2), 166–184.

Brus, B. T., & Voeten, M. J. M. (1999). Een-Minuut-Test [One-Minute-

Test]. Lisse, The Netherlands: Swets & Zeitlinger.Carrillo de la Pena, M., Holguin, S., Corral, M., & Cadaveira, F. (1999).

The effects of stimulus intensity and age on visual-evoked potentials(VEPs) in normal children. Psychophysiology, 36(6), 693–698.

Castren, M., Paakkonen, A., Tarkka, I. M., Ryynanen, M., & Partanen, J.(2003). Augmentation of auditory N1 in children with fragile Xsyndrome. Brain Topography, 15(3), 165–171.

Cohen, J., & Polich, J. (1997). On the number of trials needed for P300.International Journal of Psychophysiology, 25(3), 249–255.

Dool, C. B., Stelmack, R. M., & Rourke, B. P. (1993). Event-relatedpotentials in children with learning disabilities. Journal of Clinical

Child Psychology, 22(3), 387–398.Dykman, R. A., Ackerman, P. T., Holcomb, P. J., & Boudreau, A. (1983).

Physiological manifestations of learning disability. Journal of Learning

Disabilities, 16(1), 46–53.Elbro, C., & Petersen, D. K. (2004). Long-term effects of phoneme

awareness and letter sound training: an intervention study withchildren at risk for dyslexia. Journal of Educational Psychology, 96(4),660–670.

Espy, K. A., Molfese, D. L., Molfese, V. J., & Modglin, A. (2004).Development of auditory event-related potentials in young childrenand relations to word-level reading abilities at age 8 years. Annals of

Dyslexia, 54(1), 9–38.Fawcett, A. J., Chattopadhyay, A. K., Kandler, R. H., Jarratt, J. A.,

Nicolson, R. I., & Proctor, M. (1993). Event-related potentials anddyslexia. Annals of the New York Academy of Sciences, 682,342–345.

Ferree, T. C., Luu, P., Russell, G. S., & Tucker, D. M. (2001). Scalpelectrode impedance, infection risk, and EEG data quality. Clinical

Neurophysiology, 112(3), 536–544.Fuchigami, T., Okubo, O., Fujita, Y., Okuni, M., Noguchi, Y., &

Yamada, T. (1993). Auditory event-related potentials and reactiontime in children: evaluation of cognitive development. Developmental

Medicine and Child Neurology, 35(3), 230–237.


Geisler, M. W., & Polich, J. (1994). P300 habituation from visual stimuli?Physiology and Behavior 56(3), 511–516.

Gilger, J. W., Hanebuth, E., Smith, S. D., & Pennington, B. F. (1996).Differential risk for developmental reading disorders in the offspring ofcompensated versus noncompensated parents. Reading and Writing,

8(5), 407–417.Goswami, U. (2003). Why theories about developmental dyslexia require

developmental designs. Trends in Cognitive Sciences, 7(12), 534–540.Graham, F. K. (1997). Afterword: Pre-attentive processing and passive

and active attention. In P. J. Lang, R. F. Simons, & M. T. Balaban(Eds.), Attention and orienting: Sensory and motivational processes

(pp. 417–452). Mahwah, NJ, US: Lawrence Erlbaum Associates.Guttorm, T. K., Leppanen, P. H., Poikkeus, A. M., Eklund, K. M.,

Lyytinen, P., & Lyytinen, H. (2005). Brain event-related potentials(ERPs) measured at birth predict later language development inchildren with and without familial risk for dyslexia. Cortex, 41(3),291–303.

Hari, R., & Renvall, H. (2001). Impaired processing of rapid stimulussequences in dyslexia. Trends in Cognitive Sciences, 5(12), 525–532.

Helenius, P., Salmelin, R., Richardson, U., Leinonen, S., & Lyytinen, H.(2002). Abnormal auditory cortical activation in dyslexia 100 msecafter speech onset. Journal of Cognitive Neuroscience, 14(4), 603–617.

Helenius, P., Salmelin, R., Service, E., Connolly, J. F., Leinonen, S., &Lyytinen, H. (2002). Cortical activation during spoken-word segmen-tation in nonreading-impaired and dyslexic adults. Journal of Neuro-

science, 22(7), 2936–2944.Helenius, P., Tarkiainen, A., Cornelissen, P., Hansen, P. C., & Salmelin,

R. (1999). Dissociation of normal feature analysis and deficientprocessing of letter-strings in dyslexic adults. Cerebral Cortex, 9(5),476–483.

Hindson, B., Byrne, B., Fielding Barnsley, R., Newman, C., Hine, D. W.,& Shankweiler, D. (2005). Assessment and early instruction ofpreschool children at risk for reading disability. Journal of Educational

Psychology, 97(4), 687–704.Hirano, C., Russell, A. T., Ornitz, E. M., & Liu, M. (1996). Habituation

of P300 and reflex motor (startle blink) responses to repetitive startlingstimuli in children. International Journal of Psychophysiology, 22(1-2),97–109.

John, E. R., Prichep, L. S., Ahn, H., Kaye, H., Brown, D., Easton, P.,et al. (1989). Neurometric evaluation of brain function in normal and

learning disabled children. Ann Arbor, MI, US: The University ofMichigan Press.

Kenemans, J. L. (1990). Event-related brain potential manifestations of

orienting. Utrecht: Rijksuniversiteit.Kenemans, J. L., Verbaten, M. N., Roelofs, J. W., & Slangen, J. L.

(1989). ‘‘Initial-’’ and ‘‘change-orienting reactions: an analysis basedon visual single-trial event-related potentials. Biological Psychology,

28, 199–266.Koster, C., Been, P. H., Krikhaar, E. M., Zwarts, F., Diepstra, H. D., &

Van Leeuwen, T. H. (2005). Differences at 17 months: Productivelanguage patterns in infants at familial risk for dyslexia and typicallydeveloping infants. Journal of Speech, Language, and Hearing

Research, 48(2), 426–438.Kuijpers, C., van der Leij, A., Been, P., van Leeuwen, T., ter Keurs, M.,

Schreuder, R., et al. (2003). Leesproblemen in het voortgezet ond-erwijs en de volwassenheid. [Reading problems at the secondary schoollevel and in adulthood]. Pedagogische Studien, 80(4), 272–287.

Leppanen, P. H. T., Richardson, U., Pihko, E., Eklund, K. M., Guttorm,T. K., Aro, M., et al. (2002). Brain responses to changes in speechsound durations differ between infants with and without familial riskfor dyslexia. Developmental Neuropsychology, 22(1), 407–422.

Lew, G. S., & Polich, J. (1993). P300, habituation, and response mode.Physiology and Behavior, 53(1), 111–117.

Logan, G. D. (1992). Attention and preattention in theories of automa-ticity. American Journal of Psychology, 105, 317–339.

Logan, G. D. (1997). Automaticity and reading: perspectives from theinstance theory of automatization. Reading & Writing Quarterly, 13(2),123–146.

Logan, G. D., Taylor, S. E., & Etherton, J. L. (1999). Attention andautomaticity: toward a theoretical integration. Psychological Research,

62(165-181).Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of

dyslexia. Annals of Dyslexia, 53, 1–14.Lyytinen, H., Guttorm, T. K., Huttunen, T., Hamalainen, J., Leppanen,

P. H. T., & Vesterinen, M. (2005). Psychophysiology of developmentaldyslexia: A review of findings including studies of children at risk fordyslexia. Journal of Neurolinguistics, 18(2), 167–195.

Mallardi, V. (1979). The interstimulus interval effect on the amplitude ofthe evoked acoustic response and habituation phenomenon. Acta

Neurologica, 34(3), 188–203.Maurer, U., Bucher, K., Brem, S., & Brandeis, D. (2003). Altered

responses to tone and phoneme mismatch in kindergartners at familialdyslexia risk. NeuroReport, 14(17), 2245–2250.

Molenaar, P. C. M. (1994). Dynamic factor analysis of psychophysiolog-ical signals. In J. R. Jennings & P. K. Ackles (Eds.), Advances in

psychophysiology: A research annual. Advances in Psychophysiology

(Vol. 5, pp. 229–302). London: Jessica Kingsley Publishers.Molfese, D. L. (2000). Predicting dyslexia at 8 years of age using neonatal

brain responses. Brain and Language, 72(3), 238–245.Naatanen, R. (1990). The role of attention in auditory information

processing as revealed by event-related potentials and other brainmeasures of cognitive function. Behavioral and Brain Sciences, 13(2),201–288.

Naatanen, R., & Picton, T. W. (1987). The N1 wave of the human electricand magnetic response to sound: A review and an analysis of thecomponent structure. Psychophysiology, 24(4), 375–425.

Nicolson, R. I., & Fawcett, A. J. (2000). Long-term learning in dyslexicchildren. European Journal of Cognitive Psychology, 12(3), 357–393.

Nicolson, R. I., Fawcett, A. J., & Dean, P. (2001). Developmentaldyslexia: the cerebellar deficit hypothesis. Trends in Neurosciences,

24(9), 508–511.Papanicolaou, A. C., Loring, D. W., & Eisenberg, H. M. (1985). Evoked

potential attenuation in the two-tone paradigm. International Journal

of Psychophysiology, 3(1), 23–28.Paulesu, E., Demonet, J. F., Fazio, F., McCrory, E., Chanoine, V.,

Brunswick, N., et al. (2001). Dyslexia: Cultural diversity and biolog-ical unity. Science, 291(5511), 2165–2167.

Pennington, B. F. (1990). The genetics of dyslexia. Journal of Child

Psychology and Psychiatry and Allied Disciplines, 31(2), 193–201.Pennington, B. F., & Lefly, D. L. (2001). Early reading development in

children at family risk for dyslexia. Child Development, 72(3), 816–833.Polich, J., & Kok, A. (1995). Cognitive and biological determinants of

P300: An integrative review. Biological Psychology, 41(2), 103–146.Polich, J., Ladish, C., & Burns, T. (1990). Normal variation of P300 in

children: age, memory span, and head size. International Journal of

Psychophysiology, 9(3), 237–248.Ramus, F., Rosen, S., Dakin, S. C., Day, B. L., Castellote, J. M., White,

S., et al. (2003). Theories of developmental dyslexia: insights from amultiple case study of dyslexic adults. Brain : A Journal of Neurology,

126(4), 841–865.Ravden, D., & Polich, J. (1998). Habituation of P300 from visual stimuli.

International Journal of Psychophysiology, 30(3), 359–365.Raven, J. C., Court, J. H., & Raven, J. (1984). Coloured Progressive

Matrices. Sets A, Ab, B. Manual for Raven’s Progressive Matrices and

Vocabulary Scales. London: Lewis.Riggins, B. R., & Polich, J. (2002). Habituation of P3a and P3b from

visual stimuli. Korean Journal of Thinking and Problem Solving, 12(1),71–81.

Rumsey, J. M., Andreason, P., Zametkin, A. J., & Aquino, T. (1992).Failure to activate the left temporoparietal cortex in dyslexia: Anoxygen 15 positron emission tomographic study. Archives of Neurol-

ogy, 49(5), 527–534.Rumsey, J. M., Horwitz, B., Donohue, B. C., Nace, K., Maisog, J. M., &

Andreason, P. (1997). Phonological and orthographic components ofword recognition: A PET-rCBF study. Brain: A Journal of Neurology,

120(5), 739–759.


Sambeth, A., Maes, J. H., Quian Quiroga, R., Van Rijn, C. M., & Coenen,A. M. (2004). Enhanced re-habituation of the orienting response of thehuman event-related potential. Neuroscience Letters, 356(2), 103–106.

Seymour, P. H. K. (1986). Cognitive analysis of dyslexia. London:Routledge & Kegan Paul.

Shaywitz, B. A., Shaywitz, S. E., Pugh, K. R., Mencl, W. E., Fulbright, R.K., Skudlarksi, P., et al. (2002). Disruption of posterior brain systemsfor reading in children with developmental dyslexia. Biological

Psychiatry, 52(2), 101–110.Silani, G., Frith, U., Demonet, J. F., Fazio, F., Perani, D., Price, C., et al.

(2005). Brain abnormalities underlying altered activation in dyslexia: avoxel based morphometry study. Brain, 128, 2453–2461.

Silva Pereyra, J., Rivera Gaxiola, M., Fernandez, T., Diaz Comas, L.,Harmony, T., Fernandez Bouzas, A., et al. (2003). Are poor readerssemantically challenged? An event-related brain potential assessment.International Journal of Psychophysiology, 49(3), 187–199.

Simos, P. G., Breier, J. I., Fletcher, J. M., Bergman, E., & Papanicolaou,A. C. (2000). Cerebral mechanisms involved in word reading indyslexic children: A magnetic source imaging approach. Cerebral

Cortex, 10(8), 809–816.Snowling, M. J., Gallagher, A., & Frith, U. (2003). Family risk of dyslexia

is continuous: Individual differences in the precursors of reading skill.Child-Development, 74(2), 358–373.

Sokolov, E. N., Spinks, J. A., Naatanen, R., & Lyytinen, H. (2002). The

orienting response in information processing. Mahwah, NJ, US:Lawrence Erlbaum Associates.

Spekreijse, H. (2000). Pre-attentive and attentive mechanisms in vision.Perceptual organization and dysfunction. Vision Research, 40, 1179–1182.

Spinks, J. A., & Siddle, D. (1983). The functional signifcance of theorienting response. In D. Siddle (Ed.), Orienting and habituation:

perspectives in human research (pp. 237–314). Chichester: Wiley.Stelmack, R. M., Rourke, B. P., & van der Vlugt, H. (1995). Intelligence,

learning disabilities, and event-related potentials. Developmental

Neuropsychology, 11(4), 445–465.Stephenson, D., & Siddle, D. (1983). Theories of habituation. In D. Siddle

(Ed.), Orienting and habituation perspectives in human research

(pp. 183–236). Chichester: Wiley.Taylor, M. J., Chevalier, H., & Lobaugh, N. J. (2003). Discrimination of

single features and conjunctions by children. International Journal of

Psychophysiology, 51(1), 85–95.Taylor, M. J., & Keenan, N. K. (1990). Event-related potentials to visual

and language stimuli in normal and dyslexic children. Psychophysiol-

ogy, 27, 318–327.Thaler, V., Ebner, E. M., Wimmer, H., & Landerl, K. (2004). Training

reading fluency in dysfluent readers with high reading accuracy: wordspecific effects but low transfer to untrained words. Annals of Dyslexia,

54(1), 89–113.

van den Bos, K. P., lutje Spelberg, H. C., Scheepstra, A. J. M., & de Vries,J. R. (1999). De Klepel: vorm A en B: een test voor de leesvaardigheid

van pseudowoorden: verantwoording, handleiding, diagnostiek en behan-

deling [Manual pseudoword test ’The Clapper’]. Lisse, the Netherlands:Swets & Zeitlinger.

van den Bosch, L., Gillijns, P., Krom, R., & Moelands, F. (1993). Schaal

vorderingen in spellingvaardigheid 1: Handleiding [Spelling test: Man-

ual]. Arnhem, the Netherlands: Cito.van der Leij, A., & van Daal, V. (1990). Repeated reading and severe

reading disability. In H. Mandl, E. De Corte, N. Bennett, & H. F.Friedrich (Eds.). Learning and instruction: European research in an

international context, Vol. 2.1: Social and cognitive aspects of learning

and instruction; Vol. 2.2: Analysis of complex skills and complex

knowledge domains (Vol. 2.2, pp. 235–251). Elmsford, NY, US:Pergamon Press.

van der Leij, A., & van Daal, V. H. P. (1999). Automatisation aspects ofdyslexia: speed limitations in word identification, sensitivity toincreasing task demands and orthographic compensation. Journal of

Learning Disabilities, 32(5), 417–428.van Leeuwen, T., Been, P., Kuijpers, C., Zwarts, F., Maassen, B., & van

der Leij, A. (2006). Mismatch response is absent in 2-month-oldinfants at risk for dyslexia. NeuroReport, 17(4), 351–355.

Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M.(2004). Specific reading disability (dyslexia): what have we learned inthe past four decades? Journal of Child Psychology and Psychiatry and

Allied Disciplines, 45(1), 2–40.Verbaten, M. N., Roelofs, J. W., Sjouw, W., & Slangen, J. L. (1986).

Habituation of early and late visual ERP components and theorienting reaction: the effect of stimulus information. International

Journal of Psychophysiology, 3(4), 287–298.Verbeke, G., & Molenberghs, G. (2000). Linear mixed models for

longitudinal data. New York: Springer.Verhoeven, L. (1995). Drie-Minuten-Toets: Handleiding [Three-Minutes-

Test: Manual]. Arnhem, The Netherlands: Cito.Verhoeven, L., & Vermeer, A. (2001). Taaltoets alle kinderen (TAK)

[Language test for all children]. Arnhem, the Netherlands: Cito.Vogel, E. K., & Luck, S. J. (2000). The visual N1 component as an index

of a discrimination process. Psychophysiology, 37(2), 190–203.Wechsler, D. (1997). Wechsler Adult Intelligence Scale (Third ed.). San

Antonio, Tx: Psychological Corporation.Wijers, A. A., Been, P. H., & Romkes, K. S. (2005). Dyslexics show a

deviant lateralization of attentional control: a brain potential study.Neuroscience Letters, 374(2), 87–91.

Woestenburg, J. C., Verbaten, M. N., Van Hees, H. H., & Slangen, J. L.(1983). Single trial ERP estimation in the frequency domain usingorthogonal polynomial trend analysis (OPTA): Estimation of individ-ual habituation. Biological Psychology, 17(2–3), 173–191.

Documents

Efficiency of visual information processing in children at-risk for dyslexia: Habituation of single-trial ERPs