paper nonword repetition in children and adults: … nonword repetition in children and adults:...

PAPER

Nonword repetition in children and adults: effects on movementcoordination

Jayanthi Sasisekaran,1 Anne Smith,2 Neeraja Sadagopan3 and ChristineWeber-Fox2

1. Department of Speech-Language-Hearing Sciences, University of Minnesota, USA2. Department of Speech, Language, Hearing Sciences, Purdue University, USA3. Department of Speech, Language, and Hearing Sciences, University of Colorado, Boulder, USA

Abstract

Hearing and repeating novel phonetic sequences, or novel nonwords, is a task that taps many levels of processing, includingauditory decoding, phonological processing, working memory, speech motor planning and execution. Investigations of nonwordrepetition abilities have been framed within models of psycholinguistic processing, while the motor aspects, which also are criticalfor task performance, have been largely ignored. We focused our investigation on both the behavioral and speech motorperformance characteristics of this task as performed in a learning paradigm by 9- and 10-year-old children and young adults.Behavioral (percent correct productions) and kinematic (movement duration, lip aperture variability – an index of theconsistency of inter-articulator coordination on repeated trials) measures were obtained in order to investigate the short-term(Day 1, first five vs. next five trials) and longer-term (Day 1 vs. Day 2, first five vs. next five trials) changes associated withpractice within and between sessions. Overall, as expected, young adults showed higher levels of behavioral accuracy and greaterlevels of coordinative consistency than the children. Both groups, however, showed a learning effect, such that in general, laterDay 1 trials and Day 2 trials were shorter in duration and more consistent in coordination patterns than Day 1 early trials.Phonemic complexity of the nonwords had a profound effect on both the behavioral and speech motor aspects of performance.The children showed marked learning effects on all nonwords that they could produce accurately, while adults’ performanceimproved only when challenged by the more complex nonword stimuli in the set. The findings point to a critical role for speechmotor processes within models of nonword repetition and suggest that young adults, similar to children, show short- and longer-term improvements in coordinative consistency with repeated production of complex nonwords. There is also a cleardevelopmental change in nonword production performance, such that less complex novel sequences elicit changes in speech motorperformance in children, but not in adults.

Introduction

The ability to hear and repeat novel phonetic sequencesis crucial to word learning throughout the lifetime. Thisseemingly simple human capacity which begins ininfancy is the cornerstone of language acquisition andinvolves complex multistage processes (Gathercole, 2006;Gupta, 2006). Mediated initially by phonologicallearning, nonword repetition eventually results inlexicalization of the sounds constituting new words. Anumber of models have been proposed to account for thedifferent stages involved in the repetition of novelsequences (Gathercole, 2006; Gupta & MacWhinney,1997; Jones, Gobet & Pine, 2007). Such models haveattributed relevant roles to different mechanisms,including phonological storage, phonological short-termmemory and its interaction with longer-term memory.Among these models, the theoretical framework proposedby Gathercole (2006) has received considerable

experimental attention. Gathercole’s framework involvesvarious stages, including auditory processing, phono-logical analysis, phonological storage and retrieval,speech motor planning and execution.

The effects of communication differences and dis-orders on the ability to repeat nonwords have been testedextensively (Edwards, Beckman & Munson, 2004;Gathercole & Baddeley, 1990; Papagno & Vallar, 1995).It is now well established that poor nonword repetitionperformance is a behavioral marker of specific languageimpairment (Bishop, North & Donlan, 1996; Dollaghan& Campbell, 1998; Michas & Henry, 1994). Suchfindings highlight the diagnostic relevance ofperformance in nonword repetition to the underlyinglanguage and speech motor processes.

The results from studies of nonword repetition havebeen interpreted within the scope of underlying languageprocesses (for a review, see Gathercole, 2006). Measuredprimarily as the percent of phonemes produced correctly,

Address for correspondence: Jayanthi Sasisekaran, Department of Speech-Language-Hearing Sciences, University of Minnesota, 164 Pillsbury Drive,SE, Minneapolis, MN 55414, USA; e-mail: [email protected]

� 2009 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

Developmental Science 13:3 (2010), pp 521–532 DOI: 10.1111/j.1467-7687.2009.00911.x

nonword repetition performance is influenced by factorsthat have an effect on initial phonological learning aswell as later lexicalization of the learned phonologicalsequences. For instance, the ability to repeat nonwordshas been associated with digit span in both children andadults (Baddeley & Wilson, 1993; Butterworth,Campbell & Howard, 1986; Gathercole, Briscoe, Thorn& Tiffany, 2008; Gupta, 2003). Nonword repetition hasalso been associated with vocabulary size and strength ofexisting long-term representations, although this effect isevident in older children and adults only in situationswhere lexically supported learning, based on existingrepresentations, is possible (e.g. Gathercole, Hitch,Service & Martin, 1997; Gupta, 2003). For example,younger children experienced more difficulties inrepeating nonwords that did not resemble wordscompared to word-like nonwords, and such differencesdiminished with age (Munson, 2001). Nonwordrepetition is also influenced by the frequency ofphonemic sequences in a language measured as high vs.low phonotactic probability (e.g. Gathercole, Frankish,Pickering & Peaker, 1999; Storkel, Armbruster & Hogan,2006; Storkel, 2001).

Thus, findings from earlier studies of nonwordrepetition have been interpreted within the context ofstrength of existing phonological representations andshort-term phonological storage capacity. However,another critical process underlying nonword repetitionabilities involves output implementation mediated by thespeech motor system. But, the role of speech motoroutput processes underlying nonword repetition abilitiesand the changes at the motor level with the acquisition ofnovel phonetic strings have not been investigatedextensively. It could be argued that some of the variablesknown to influence nonword repetition accuracy, such asnonword length, word likeness, and phonemic complexity,also operate on speech motor output processes such asspeech kinematics. In fact, measures of speech kinematicswould likely provide a much more sensitive index of theunderlying processes mediating formulation andimplementation (e.g. Goffman & Smith, 1999). Inaddition, evidence from speech disordered populationsindicates that the contributions of speech motorprocesses to nonword repetition abilities are critical for acomplete understanding of the necessary underlyingfunctions. For example, poor performance in non-word repetition has been reported in individualsdiagnosed with stuttering (Hakim & Ratner, 2004;Anderson, Wagovich & Hall, 2006) and in articulationdisorders (Yoss & Darley, 1974), where a primarycomponent of the disorder involves the speech motorsystem. In summary, the speech motor processesmediating the implementation of nonword repetitionshave received little experimental attention, and it is theseproduction processes that are the focus of the presentinvestigation.

The nonword repetition task engages the speech motorsystem and offers an opportunity to investigate the

effects of successive repetition of novel material onspeech movement coordination through the process ofmotor learning. Schmidt (1988) defined motor learningas a set of internal processes associated with practice orexperience leading to relatively permanent changes in thecapability for responding. It results in increasingmovement accuracy and speed with increasingcoordination and decreasing variability with practice.Both transient and persistent changes in the neuralsubstrates supporting motor learning are thought toresult in short- and longer-term changes in movementcoordination (Monfils, Plautz & Kleim, 2005). Thus,movement coordination is achieved through aneurophysiologically mediated motor learning process(Harris & Wolpert, 1998; Smith & Zelaznik, 2004). Atheoretical notion that has received considerableattention in the limb motor literature is the neuromotornoise hypothesis. Accordingly, changes with motorlearning are the consequences of reducing variability inneural command signals (Kleim, Bruneau, Calder,Pocock, VandenBerg, MacDonald, Monfils, Sutherland& Nader, 2003; Kleim, Hogg, VandenBerg, Cooper,Bruneau & Remple, 2004; Newell, Liu & Mayer-Kress,2003). Both children and adults exhibit higher levels ofneural noise during the acquisition of new movementsequences, which is evident as higher movementvariability (Green, Moore & Reilly, 2002; Smith &Zelaznik, 2004; Takahashi, Nemet, Rose-Gottron,Larson, Cooper & Reinkensmeyer, 2002; Van Galen,Portier, Smitsengelsman & Schomaker, 1993; Walsh,Smith & Weber-Fox, 2006; Yan, Thomas, Stelmach &Thomas, 2000). With maturation and practice, increasingneuronal synchronization results in reduced neuromotornoise levels, thereby facilitating motor coordinationthrough the formation of optimal movement synergies.For instance, Smith and Zelaznik (2004) and Walsh andSmith (2002) demonstrated that children continue toshow increasingly consistent articulatory coordinationover the adolescent years, with adult levels of speech rateand speech coordination consistency being achieved onlyafter 16 years of age.

Experimental evidence for changes in speechcoordination with nonword repetition has been reported(Schultz, Stein & Micallef, 2001; Walsh et al., 2006). Thesefindings offer preliminary support for the role of speechmotor processes within the theoretical frameworks ofnonword learning that have been proposed, althoughfurther experimentation is required to understand how thedifferent task-relevant variables, such as phonemiccomplexity, influence speech coordination. For example,Walsh et al. (2006) investigated short-term plasticity(within a single experimental session) of the speechmotor system of 20 9- to 10-year-old children and 20young adults in a nonword repetition task. They employeda kinematic measure of consistency of upper lip, lower lip,and jaw coordination, called the lip aperture variability (LAVAR) index. The aim of the study was to investigatechanges in movement coordination with repeated

522 Jayanthi Sasisekaran et al.

� 2009 Blackwell Publishing Ltd.

nonword production. Participants were presented withblocks of five pseudo-randomized nonwords, and changesin LA VAR from the first five to the last five correctlyproduced trials of each nonword across the blocks wereanalyzed. Their findings supportedwithin-session changesin speech coordinative patterns with nonword repetition inchildren.

The findings from studies investigating limbmovement coordination have supported the occurrenceof both short- (within session) and longer-term(between sessions) changes in coordination consistencywith skill acquisition in both children and adults. Suchchanges, seen as reductions in movement variability anddurations due to practice and other associatedprocesses, such as sleep (e.g. Walker, Brakefield,Morgan, Hobson & Stickgold, 2002; Walker,Stickgold, Alsop, Gaab & Schlaug, 2005), are thoughtto reflect transient and persistent changes in corticaland subcortical motor maps (Kleim et al., 2004;McNamara, Tegenthoff, Dinse, Buchel, Binkofski &Ragert, 2007; Monfils et al., 2005). Interestingly, theadult participants in the Walsh et al. (2006) study failedto show a learning effect on movement coordination,while the 9- to 10-year-olds showed significantreductions in lip aperture variability with increasingpractice within a single session. A closer examination ofthe nonwords used by Walsh et al. (2006) revealed thatwhile these may have challenged the speech motorsystem in children, perhaps the adults were performingat ceiling, thereby failing to show changes in movementcoordination with repeated nonword production. Itremains to be seen whether adults will show a motorlearning effect, similar to that seen in children, ifthe nonwords are sufficiently complex to challenge themature motor and phonemic systems. Furthermore, theextent to which the short-term changes in speechmovement coordination for nonword repetitionobserved in children within a session is preserved thefollowing day remains to be tested.

Thus, the primary goal of this investigation was toelucidate the role of speech motor output processeswithin existing theoretical frameworks proposed toaccount for nonword repetition. More specifically, thepresent study was designed to extend earlier findings ofchanges in speech movement coordination with nonwordrepetition in both the developing motor systems in 9- to10-year-olds (e.g. Walsh et al., 2006) and the maturesystems in young adults. Both behavioral (percent correctproductions) and kinematic (lip aperture variability,movement duration) measures were used to evaluateperformance in the nonword repetition task. Wehypothesized that: (a) Based on the neuromotor noisehypothesis, children will exhibit higher error rates andmovement variability compared to adults, and bothchildren and adults will show higher variability in oralmotor coordinative consistency on Day 1 compared toDay 2; (b) Young adults will show changes in movementcoordination within a session when the phonemic

complexity of the nonwords is higher than those usedby Walsh et al.; (c) The short-term changes in accuracyand movement variability observed within a session willpersist overnight. In other words, changes in speechcoordinative consistency observed on Day 1 will persistin both children and adults such that performance on thefirst five trials of Day 2 will approximate those of the lastfive trials of Day 1 while being less than the variability ofthe first five trials on Day 1; (d) Finally, we hypothesizedthat the short- and longer-term changes associated withspeech motor learning of nonwords result in progressivemodification and organization of movement synergiesleading to more consistent and accurate performance inboth children and adults.

Method

Participants

Nineteen children (10 males) aged 9–10 years (Mage = 9.75, SD = .40), and an equal number of adults (10males) in the age range of 18 to 25 years (M age = 22.3,SD = 2.0) participated in the study. Participants completedtwo experimental sessions over consecutive days (Day 1 andDay 2). We chose children in the 9- to 10-year age rangebased on past reports that children at this age showsensitivity to potential interactions between the developingspeech motor and language systems, and because they areable to perform the nonword repetition task (Smith &Zelaznik, 2004; Walsh et al., 2006). All participants werenative speakers of North American English. Participantswere recruited based on their responses to a screening formthat was used to rule out a positive history of language,hearing, and ⁄ or neurological deficits, and current usageof drugs likely to affect the outcome of the experiment(e.g. drugs for ADHD and anti-anxiety drugs). Childrenwere screened for language deficits using the ClinicalEvaluation of Language Fundamentals-3 (CELF-3; Semel,Wiig & Secord, 1995). Normal articulatory structures andmovements in both the age groupswere confirmed using theOral Speech Mechanism Screening Evaluation-Revised(OSMSE-R; St Louis & Ruscello, 1987). Finally,participants passed a hearing screening test performed at.5, 1, 2, 4, and 8 KHz (20 dB) in both ears.

Nonword repetition skills and short-term memoryskills for participants in both groups were determinedusing the Nonword Repetition test (Dollaghan &Campbell, 1998) and the forward and backward Digitspan tests (Wechsler’s Memory scale; Wechsler, 1997).The Nonword Repetition test was administered to allparticipants as a baseline measure of the ability toperceive and repeat nonwords. The nonwords spoken bya native English speaker were pre-recorded andpresented over loudspeakers, and participants wererequired to repeat each nonword. The nonwords variedin length (1–4 syllables) and consisted of tense vowelsand consonants acquired early in development while

Nonword repetition in children and adults 523


excluding the late-eight consonants and consonantclusters (Dollaghan & Campbell, 1998). Performance inthis task was analyzed by the first author offline bycounting the percent consonant and vowel errors foreach nonword length. Children and adults performednear ceiling for the 1-, 2-, and 3-syllable nonwords.For the 4-syllable category, adults (M = 91.6%,SE = 1.4) performed better than children (M = 85.6%,SE = 1.4), t = 2.85, p = .007. Reliability coding wasperformed on participant responses by a trainedresearch assistant and inter-judge reliability computedusing Pearson product-moment correlation was .75(p < .01). In the digit span subtests, children had anaverage span of 8.3 (SD = 1.9) and 6.2 (SD = 3.1) in theforward and the backward digit span subtests,respectively, while the adults’ scores were 11.3(SD = 2.1) and 8.3 (SD = 2.5). While scores on thesetests were not used as screening criteria for subjectinclusion, the scores approached the range reported inearlier studies (e.g. Gathercole, Willis, Emslie &Baddeley, 1994; Gupta, 2003).

Nonword stimuli for the kinematic study

Four nonwords and a control word on which we did notexpect motor learning to occur, the compound word,

, were the stimuli. The first syllable ⁄ mæb ⁄and the last phoneme ⁄ b ⁄ were identical for all nonwords.This strategy was used to enable selection of consistentstart and end points for oral movement data extractionon the basis of lower lip peak opening velocities duringthe articulatory trajectory analysis. For similar reasons,the composition of the compound word, ‘potato chip’( [IPA notation]), allowed consistentdemarcation of the start and end points. Of thefour nonwords, two – ‘mabfaisheib’ ( ;3-syllable) and ‘mabsheitaidoib’ ( ;4-syllable) – used by Walsh et al. (2006), consisted ofsyllables containing consonants acquired by 4 years ofage. To increase nonword complexity in the presentexperiment, two additional 4-syllable nonwords wereadded to this list. The third nonword, ‘mabspoukweefleib’( ) included three consonantclusters, ⁄ sp ⁄ ⁄ kw ⁄ ⁄ fl ⁄ (with varying combinations ofvowels and diphthongs), acquired between 4 and 6 yearsof age, and for this reason is argued to be phonemicallymore complex than nonwords 1 and 2. Finally, the fourthnonword, ‘mabskrisploistroob’ ( ),included three consonant clusters, ⁄ skr ⁄ ⁄ spl ⁄ ⁄ str ⁄ (withvarying combinations of vowels and diphthongs),acquired between 7 and 11 years of age and thus isphonemically more complex than the other nonwords.We also used the Vitevich and Luce (2004) method toobtain a global measure of nonword complexity that wasbased on the sum of phonotactic probilities of allbiphones (all possible two-sound combinations) in eachnonword. These authors proposed using the sum ofbiphone probabilities to find the median value for this

measure from the set of potential stimuli. Items greaterthan the median value are operationally defined asnonwords with high phonotactic probability, and itemsless than the median value are defined as nonwords withlow phonotactic probability. The lower the sum ofbiphone probability of a nonword as compared to themedian, the lower is the phonotactic probability ofthe biphones constituting the nonword and higher thenonword complexity. The sum of biphone probabilitieswere .0172 for , the only 3-syllable nonword,.014 for , .018 for ,and .027 for , with a median biphoneprobability of .018 for the three 4-syllable nonwords. Thecompound word, potato chip, was included to comparenovel nonword performance to that of a familiar wordthat is comparable in phonemic composition to thenonwords but that also has a known lexicalrepresentation. The stimuli, produced by a nativefemale American English speaker, were recorded anddigitized using PRAAT software. For all the nonwords,the primary stress was on the first syllable with a weaksecondary stress on the third syllable. This particularstress pattern was chosen to be consistent with thepredominant stress pattern for a majority of Englishwords. We hypothesized that this stress pattern wouldreduce the number of production errors related to stressplacement. The duration of each stimulus model was:

– 1.41 ms, – 1.44 ms,– 2.00 ms, –

2.35 ms, and potato chip – .83 ms.

Apparatus

Participants were seated in front of an Optotrak 3020camera (Northern Digital, Waterloo, Ontario), acommercial system that allows tracking of movementsin 3D with an accuracy of less than 0.1 mm. Eight small(7 mm) infra-red light emitting diodes (IREDs) wereattached to track articulatory movements of the upperlip, lower lip, and jaw. Four of the IREDs were mountedon a set of goggles that participants wore during theexperimental session. One IRED was placed in the centerof the forehead. Together, these five IREDs were used tocalculate the 3D head coordinate system (Smith,Johnson, McGillem & Goffman, 2000), which allowedhead movement artifact to be eliminated. To track themotion of the lips, one IRED was placed at the center ofthe vermilion border of the upper lip and one at thecenter of the vermilion border of the lower lip (thismarker represents combined actions of the lower lip andjaw). The IRED motions were sampled at a rate of250 Hz. A condenser microphone placed 8 cm awayfrom the participant’s mouth was used to record thespeech signal. This acoustic signal was digitized on anA ⁄ D channel of the Optotrak system and was thussynchronized with the movement data. The acousticsignal was digitized at the rate of 16,000 Hz and low-passfiltered with a cut-off frequency of 7500 Hz.



Procedure

On Day 1, participants were instructed that they wouldhear novel words that do not exist in English. Theywere asked to listen carefully and repeat each novelword as best they could. Following instructions, eachnonword was presented via loudspeakers. Participants’productions during practice were monitored andcorrected for pronunciation and stress errors by theexperimenter. A majority of participants were able torepeat the nonwords with the primary stress on thefirst syllable. Practice was terminated when a criterionscore of two correct productions was achieved for eachtarget item. A minimum of two trials and a maximumof five trials were allowed for each participant. Thedata indicated that for a majority of the stimuli,participants were able to achieve a minimum of twoconsecutive, correct productions within five practicetrials. Following the practice trials, each nonword waspresented once more, and participants were required toproduce them in the carrier phrase ‘Say _______again’. Participants then completed the experimentaltask, producing the target items embedded in thecarrier phrase. The carrier phrase was used to controlarticulatory positions at the beginning and ending ofthe target nonwords.

The experimental session lasted approximately 60 minon Day 1 and 30 min on Day 2. Both sessions weresimilar except for the practice trials, which were excludedon Day 2. During each session, the nonwords werepresented auditorily in 13 pseudo-randomized blocks.Each block consisted of five trials, with the fournonwords and ‘potato chip’ occurring once in randomorder. During the experiment, the participant wasinstructed to repeat the target nonword played by theexperimenter in the carrier phrase. Each target stimuluswas separated by a 2-sec gap. Participants’ responseswere recorded by the experimenter during each session todetermine if the subject had produced a minimum of 10correct trials for each stimulus, because this number ofrepetitions is needed for the trajectory analysis. Failure toobtain 10 correct productions within the 13 blocksresulted in the presentation of two or three additionalblocks. The experiment was then terminated regardlessof whether 10 correct productions had been obtained.Correct responses were defined as those trials which werefree of disfluencies or articulation errors. The first andthird authors ran the experiment and had clear visibilityof the face and articulators of participants at all times.Both consonant and vowel errors were considered, butonline scoring was ‘all or none’. Only trials agreed uponby both experimenters as correct productions wereincluded in the offline LA VAR analysis. Even withinthe correctly produced tokens, repetitions with inter-syllable pause (> 3 seconds, evident as flat lines of theLA VAR trajectories) and stress errors were kept to aminimum based on instructions and correction of errorsduring initial practice.

Data analysis

The acoustic and kinematic data from each session wereloaded into MATLAB digital signal processing softwarefor analysis (for details on the analysis, see Smith et al.,2000; Smith & Zelaznik, 2004). Following low passfiltering of the upper lip and lower lip displacementsignals, the velocity signal was computed from the lowerlip displacement using the 3-point difference method.Then, the displacement trajectories associated withcorrect productions of the stimuli were obtained byidentifying each stimulus from the 13–15 blocks andsegmenting the corresponding displacement files. Thestart and end points of the upper lip and lower lip datathus segmented were determined using the lower lipvelocity signal. The start point for each segment was thepoint of maximum opening velocity for ⁄ m ⁄ (in the firstsyllable ⁄ mæb ⁄ ) and the end point was the point ofmaximum opening velocity of the lower lip openinggesture for ⁄ b ⁄ (in the last syllable). For potato chip, thestart and end points were based on the maximum openingvelocity of the lower lip for the initial and final ⁄ p ⁄. Thesepeak velocity regions were easily located by theexperimenter, and a MATLAB algorithm selects thepeak opening velocity within a user-defined window. Eachdata segment selected in this manner was then reassessedfor fluency and accuracy of extraction by listening to theassociated audio record. If there were more than 10correct productions of a target nonword, the trajectoriesassociated with the first 10 were used in the data analysis.

Following extraction of the upper lip and lower lipmovement trajectories for each set of 10 productions ofevery target nonword, a multi-step analysis wasperformed using a custom MATLAB program. Thisanalysis resulted in computation of the lip aperturevariability index (LA VAR; Smith & Zelaznik, 2004),which is a measure of the trial-to-trial spatial andtemporal variability associated with the lip aperture (LA)trajectory. First, a sample by sample subtraction of thelower lip displacement signal from the upper lipdisplacement signal was performed to obtain thedifference between the upper and lower lip IREDmarkers as a function of time. Second, the 10 LA trajec-tories for each nonword were amplitude normalized,which involved subtracting the mean and dividing by thestandard deviation. Third, using interpolation, the LAtrajectories were time normalized to a fixed record lengthof 1000 points. Fourth, standard deviation values werecalculated at 50 point intervals (2% intervals in relativetime) of the normalized waveform. The cumulative sumof the 50 standard deviations was computed to obtain anLA VAR value, a cumulative spatial and temporalcoordinative index. A higher LA VAR index suggests agreater degree of trial to trial variability in inter-articulatory coordination for nonword production, andvice versa. In addition to LA VAR, the movementduration in real time was computed as the total durationof the LA signal for each target production.



Statistical analyses

In order to investigate any changes in movement coordi-nation with nonword repetition in children and adults,three repeated measures ANOVAs were conducted on thebehavioral (percent correct responses) and kinematic (dura-tion, LAVAR) data. For the behavioral analysis, Age andSex were the between-subjects factors, while Complexity

and Day (1 vs. 2) were the within-subjects factors. For the kinematic analysis, two sepa-rate repeated measures ANOVAs were run, one onmovement duration and one on LA VAR. For theseanalyses, Age and Sex were the between-subjects factorswhile Complexity

, Day (1 vs. 2), and Trials (first five vs.later five) were the within-subjects factors. Fischer’sleast significant difference post-hoc test was used toidentify sources of the significant main and interactioneffects. For adults who produced 10 correct trials of

, the LA VAR scores were analyzedusing paired-sample t-tests. The potato chip productionswere not included in the ANOVAs; rather, separate stati-stical tests were run to ascertain group differences onkinematic data from this word. The magnitude of learning(in percent) for each nonword across both days was alsoestimated ([Day 1, LAVAR First 5 trials] – [Day 2, LAVARLater 5 trials] ⁄ [Day 1, LAVAR First 5 trials]). This valuereflects the increase in coordinative consistency from theearliest trials on Day 1 to the final trials on Day 2. The per-cent values obtained from this calculation were comparedin children and adults using a repeated measures ANOVAwith Age as the between-subjects factor and Complexity

as the within-subjects factor. Due to the limited numberof correct productions of in children,the magnitude of learning for this nonword was measuredonly in adults.

Kinematic pattern analysis

In order to test the hypothesis that speakers are makingsystematic adjustments in oral coordination to achieve afinal pattern in the nonword repetition task, a correlationanalysis (Smith, Goffman, Zelaznik, Ying & McGillem,1995) was performed on the time and amplitudenormalized lip aperture trajectory for each subject as afunction of day. An average lip aperture template wascomputed for trials 9 and 10, and a zero-lag cross-correlation was computed for the lip aperture trajectoriesof each trial (1–8) with this template (see Figure 5).The resulting correlations from the nonwords wereanalyzed in a repeated measures ANOVA with Age andSex as the between-subjects factors and Complexity

,Day, and Trial (1–8) as the within-subjects factors.Correlation data from adults forwere analyzed in a separate ANOVA as a majority of the

children were unable to produce 10 correct repetitions ofthis nonword.

Results

For all statistical analyses reported below, no effects ofSex were found; therefore all results are reportedcollapsed across Sex.

Behavioral data

Percent correct productions

Figure 1 shows the Means and SEs for the percentcorrect productions on Days 1 and 2 for children andadults. The ANOVA indicated significant main effectsof Age, F(1, 35) = 13.0, p = .001, g2 = .27, Complexity,F(3, 111) = 31.2, p = .0001, g2 = .47, and Day, F(1, 35) =46.1, p = .0001, g2 = .56. A significant Complexity · Ageinteraction was also obtained, F(3, 111) = 7.1, p = .0001,g2 = .16. Differences in performance were seen be-tween children and adults for the more complex non-words, , with theadults, as expected, performing better at both thesecomplexity levels as compared to the children. An effectof learning was also evident from the Complexity · Dayinteraction, F(3, 111) = 3.1, p = .03, g2 = .08. Bothadults and children showed an improvement in perfor-mance on the second day, which was significant for

. Atrend for similar improvement on Day 2 was seen only inadults for the most complex nonword, .

A separate repeated measures ANOVA was performedon arcsin transformed values of the percent of correctproductions in children and adults. This analysis wasperformed because the range of the percentage correctresponses for the four nonwords varied between 0 and100 (Bartlett, 1947; Studebaker, 1985). The results of this

Figure 1 Mean percent correct productions and SE barsfor each age group are plotted as a function of nonword andday.



analysis were similar to the ANOVA results reportedabove.

Kinematic data

Duration

The overall movement duration associated with eachtarget word was measured as the time between theinitial and final lip opening movements, and indicatesthe speaking rate for each movement sequence. Forpotato chip, age had a significant effect on movemen-tduration, with children showing longer movementdurations, F(1, 32) = 15.2, p = .0001. No othersignificant main or interaction effects were obtained.

In Figure 2, the duration of the nonwords is plottedfor both groups across Days 1 and 2. The most difficultnonword, , was excluded from theANOVA as only two children on Day 1 and five on Day 2could produce this nonword. Age had a significant effecton movement duration, F(1, 20) = 20.09, p = .001,g2 = .49, with adults showing shorter durations. Noother main effects were obtained. A facilitatory effect ofnonword repetition on movement duration was alsoobserved in a Complexity · Day interaction, F(2, 46) =3.9, p = .03, g2 = .14. Although participants showedshorter durations for all the nonwords on the second day,the post-hoc test revealed this reduction to be significantonly for . The four-way interaction ofComplexity · Day · Trial · Age approached signifi-cance, F(2, 46) = 2.8, p = .07, g2 = .11. In adults, for

(note that most children were unableto produce this nonword), paired-sample t-tests revealeda reduction in movement duration with practice, that is,the production duration of the first five trials wassignificantly longer than the later five trials, t(12) = 1.9,p = .03, on Day 1. Retention of learning was observedas a reduction in movement durations of the first, t(12) =2.2, p = .02, and later five trials on Day 2, t(12) =

2.7, p = .009, as compared to the first five trials onDay 1.

Lip aperture variability

For the compound word, potato chip, age had a significanteffect on the LAVAR scores with children showing higherLAVAR scores, and variable LA trajectory patterns, F(1,31) = 19.2, p = .0001. No other significant main orinteraction effects were obtained. Figure 3 shows sets ofthe amplitude- and time-normalized lip aperturetrajectories and the average LA VAR scores for

on Day 1 in a 10-year-old and a youngadult. The 10-year-old shows more variable lip aperturetrajectory patterns for the 10 trials, which is reflected in thehigher LAVAR score. In contrast, the adult speaker showsa relatively consistent inter-articulatory coordination pat-tern. In Figure 4, Means and SEs of LA VAR scores forthe four target nonwords are plotted for both days. Again,because most children could not produce the most diffi-cult nonword, their data are not included in Figure 4. Sta-tistical analysis

confirmed that the children showedmore variability than adults in the inter-articulatorytrajectory pattern associated with repeated productionsof each nonword, F(1, 20) = 20.7, p = .001, g2 = .47. Aneffect of Complexity indicated that participants in bothage groups used more consistent inter-articulatory co-ordination patterns over multiple repetitions of the leastcomplex nonword, , F(3, 60) = 12.8, p =.0001, g2 = .39. The LA VAR scores were lower on thesecond day in both children and adults, F(1, 20) = 20.01,p = .0001, g2 = .50.

As illustrated in Figure 4, children and adultsdiffered in the pattern of change in LA VAR withnonword complexity across both days. A significantComplexity · Day · Trial · Age interaction, F(2,48) = 3.4, p = .04, g2 = .12, revealed effects ofnonword repetition on movement coordination inboth children and adults. Children showed areduction in the LA VAR scores for all threenonwords. Adults, on the other hand, failed to showthis effect for the least complex nonword, .With increasing phonemic complexity, both childrenand adults showed a reduction in LA VAR that wasmost evident between Days 1 and 2, rather than withina session, i.e. first five vs. later five.

For adults who produced 10 correct trials of, the LA VAR scores were analyzed

using paired-sample t-tests. A reduction in LA VAR wasseen from the first five to the last five trials on Day 1, thusshowing a clear effect of nonword repetition on movementcoordination with practice, t(16) = 3.05, p = .004. Signi-ficant reductions in the LAVAR scores were also observedon the first, t(16) = 2.6, p = .009, and later five trials onDay 2, t(16) = 4.7, p = .0001, as compared to the first fivetrials on Day 1, thereby showing a speech motorconsolidation effect. Finally, the analysis of magnitude

Figure 2 Mean duration and SE of target nonwords on Days 1and 2 are plotted for adults and children.



of learning revealed that the extent of speech motorlearning in children and adults was comparable for

,F(1, 24) = 1.46, p = .23 (see Table 1).

Kinematic pattern analysis

Figure 5 illustrates the zero-lag cross-correlationpattern analysis that was computed for trials 1–8with the average template of trials 9 and 10 for oneadult subject for Day 1. Correlation values weresquared before entering them into the ANOVA. InFigure 6, the mean correlation (squared) values for

are plotted forchildren and adults for Day 1 and Day 2. Overall, thisanalysis did not reaveal a main effect of Age; both

children and adults showed highly positive correlations.A significant main effect of Trial, F(7, 133) = 6.3, p =.0001, revealed that the trajectories of the later trials (6–8) correlated more highly with the final templatecompared to the earlier trials (1–3). An interactioneffect of Complexity · Day was also obtained, F(2, 38) =17.2, p = .0001. As illustrated in Figure 6, for both

, higher correla-tions were observed between the individual trial lipaperture trajectories and the final template on the secondday as compared to the first day. In adults, for

, a significant main effect of Trial,F(7, 77) = 2.4, p = .02, was obtained.

Discussion

In the present study, we examined transient, short-term(within-session) and persistent, longer-term (between-sessions) changes in movement coordination with therepetition of phonemically simple and complex nonwords.Past studies of nonword repetition typically have reportedthe percent of phonemes produced correctly, and theories ofnonword repetition based largely on such reports haveattributed a role to existing phonological representationsand storage capacity to performance in this task whilefocusing less on the speech motor output aspects (e.g.Gathercole, 2006). Furthermore, earlier studies of changesin movement coordination with nonword repetition havefailed to show effects in adults, while demonstrating signi-ficant gains in oro-motor coordination consistency andmovement speed within an experimental session in children(Walsh et al., 2006). Our findings demonstrate thatnonword repetition involves incremental changes at bothbehavioral and kinematic levels of performance in both

Figure 3 Comparison of data from a 10-year-old and a young adult. In the left panel,amplitude- and time-normalized articulatorytrajectories for the first five trials of nonword‘mabspoukweefleib’ are plotted and thetrajectories for the later five trials are plottedin the right panel.

Figure 4 Mean and SE bars for lip aperture variability for thefirst five (F, filled) and later five (L, open) trials on Days 1 and 2are plotted for children and adults.



children and adults. The findings suggested that whenhumans learn new words, even with only five to 10 practicetrials, short-term increases in movement coordinationoccur. Furthermore, the changes that occur within a30-min session persist overnight to result in furtherconsolidation of the observed changes in movementcoordination. The correlation pattern analysis revealedthat these changes from early to later trials were systematicin both children and adults, suggesting adjustments inthe motor commands toward a more optimal coordinativestrategy. The short- and longer-term changes in move-ment coordination associated with nonword repetition arespeculated to result eventually in the acquisition of optimalmovement trajectories.

Effect of phonemic complexity of nonwords onmovement coordination

Behavioral data revealed that both children and adultsshowed a learning effect seen as an increase in thepercent of correctly produced nonwords on Day 2,although as expected, such differences were significantin adults only for the more complex nonwords. Thisfinding corroborates earlier reports of changes in wordlearning with increasing exposure to novel words usingother tasks, such as phoneme monitoring, that alsofacilitate phonological learning (e.g. Gaskell & Dumay,2001; Leach & Samuel, 2007). A majority of thechildren were unable to achieve 10 correct productionsof the most complex nonword, , onboth days. Thus, behavioral performance demonstrateda clear effect of phonemic complexity on nonwordproduction. Improvements in behavioral performancewere paralleled by similar changes at the kinematic

level in both age groups. In children, speech motorcoordination for all nonwords improved within a shortexperimental session involving approximately 10repetitions. This finding, previously reported by Walsh etal. (2006), offers evidence for short-term changes incoordination consistency with nonword repetition inchildren. For the first time, we also demonstrate that ifthe nonwords are sufficiently challenging in phonemiccomplexity, young adults show reductions in movementvariability with repeated nonword production similar tothat seen in children.

The nonword complexity in the present study was variedbased on syllable number, age of acquisition of theconsonants, and the extent of left-branching length ofthe consonant clusters (Sevald, Dell & Cole, 1995).Behavioral studies have demonstrated that phonemiccomplexity of nonwords varied based on these factorsinfluenced speech production speed and accuracy inchildren as well as adults (e.g. Gathercole & Baddeley,1990; Gathercole et al., 1997; Storkel, 2001). Kinematicstudies of nonword repetition have, however, failed todemonstrate an effect of these factors on movementcoordination in adults (e.g. Walsh et al., 2006). Thesignificant Complexity · Day · Trial · Age effectdemonstrates that a learning effect is indeed evident inadults for multisyllabic nonwords containing consonantclusters that are acquired at a later age. For the leastcomplex nonword, , behavioral performancewas at ceiling and differences in movement coordinationwere not observed between Days 1 and 2 in adults. For thesame nonword in children, movement variability was lessfor the later five trials on Day 1 compared to the firstfive trials (Fischer LSD, p = .048). This finding, alsoreported by Walsh et al. (2006), suggests the absence of

Table 1 Magnitude of learning (%) in children and adults

Age group mabfaisheib mabsheitaidoib mabspoukweefleib mabskrisploistroob

Young adults M )6.5 18.4 16.8 16.8SE 9.8 5.0 5.4 4.3

9- to 10-year-olds M 4.8 15.7 15.0 –SE 8.4 7.8 7.5 –

Figure 5 An illustration of the zero-lagcross-correlation (XCORR) pattern analysis.Plot A – average template from trials 9 and10. Plots B to E – LA trajectories for trials 1,4, 7, and 8. The dotted line plotted with theLA trajectories for each trial is the averagetemplate for trials 9 and 10. This subjectshows a dramatic increase in the XCORRfrom trials 1 to 8.



an effect of nonword repetition on movement coordi-nation in adults for relatively simple nonwords. However,with increase in phonemic complexity, for instance in

, initial performance was not at ceilingand improvements in oral motor coordination consis-tency and reduction in lip aperture variability wereobserved between the first and later five trials on Day 1(Fischer LSD, p = .02) and between the first five trials ofDay 1 and the first (Fischer LSD, p = .01) and later fivetrials on Day 2 (Fischer LSD, p = .01). With furtherincrease in phonemic complexity, that is, for

, significant reduction was observed inmovement variability only between the first five trials onDay 1 vs. the last five trials on Day 2 (Fischer LSD,p = .031). This finding suggests that adults show speechmotor learning for phonemically complex nonwords forwhich they do not have previously established optimalcoordinative synergies.

Taken together, the behavioral and kinematic dataoffer preliminary evidence that even within nonwordsof equal length in syllables

changes in move-ment coordination with nonword repetition varies basedon syllable-internal, phonemic composition. One futureline of research is necessary to determine the effect ofnonword complexity, indexed by frequency of phonemictransitions (i.e. phonotactic probability) within syllables,on movement coordination. Such an investigation will

enable direct testing of speech production models thatposit the existence of a mental syllabary to account forproduction differences between high and low frequencysyllables in a language (e.g. Levelt, Roelofs & Meyer,1999). Furthermore, some of the observed behavioraland kinematic changes may also be related to improvedphonological representation of the nonwords in workingmemory with repeated practice. While our study was notdesigned specifically to test the independent contri-butions of cognitive and linguistic factors, futurestudies are required to isolate the effects of suchvariables on movement coordination.

Short- and longer-term changes in movementcoordination with nonword repetition

Effects of nonword repetition on movement coordinationwere observed in both children and adults on both Days1 and 2. Overall, children exhibited longer movementdurations and more variability in the lip aperturetrajectories. This finding is consistent with previouslyreported age-related differences in speech (Green et al.,2002; Smith & Zelaznik, 2004; Walsh & Smith, 2002) andlimb (Takahashi et al., 2002; Van Galen et al., 1993; Yanet al., 2000) motor performances. We hypothesize thatthe higher variability in children is associated with higherlevels of general neuromotor noise, which results insuboptimal movement trajectories (Harris & Wolpert,1998). We found additional support for higher levels ofgeneral neuromotor noise in children from the Age effectfor the compound word, potato chip. In the absence ofa learning effect, that is, a reduction in lip aperturevariability from Day 1 to Day 2, higher movement varia-bility associated with this familiar target word suggestsfurther that the speech motor systems in children arephysiologically constrained by inherently high neuro-motor noise levels, even for the production of familiarsyllable sequences that do not require practice. Interest-ingly, despite the higher variability in children, a com-parable magnitude of learning was observed in bothchildren and adults on the nonwords. This suggests thatalthough performance efficiency is limited by neuro-motor noise, children show adaptive changes with learn-ing to the same extent as do adults – a finding thatalludes to the potential role of external factors, such asextended practice and external feedback, in facili-tating learning in children (e.g. Miall & Wolpert, 1996;Newell et al., 2003). For instance, with extended prac-tice would children show a learning effect, similar tothat seen in adults, for the most complex nonword,

With testing the next day, the gains in consistency andspeed of production were maintained in the first fivetrials. In adults, improvements in oral motorcoordination consistency on Day 1 were retained in theearly trials on Day 2, although not for all nonwords.Such changes are thought to be the result of bothpractice and sleep-related consolidation effects (Walker

Figure 6 Illustrated are the mean squared cross-correlationvalues for the early trials (1–3) and later trials (6–8) with thefinal template (average of trials 9 and 10) for ‘mabfaisheib’ and‘mabsheitaidoib’.



et al., 2002; Walker et al., 2005). As noted earlier, thisconsolidation effect was also accompanied by an increasein the percent of correct nonwords produced on Day 2.We speculate that short- and longer-term increases incoordinative consistency with nonword repetition reflectdecreases in general background noise and decreasedvariability in the motor commands required to generatethe speech movement sequences. Evidence for the lattermechanism is provided by the correlation analysis, whichindicates systematic changes in movement trajectoriestowards an optimal coordinative pattern in both childrenand adults. From the limited evidence in the motorcontrol literature, we speculate that the reduction inthe variability of motor commands for the speechmovement sequences would reflect reorganization ofcortical microcircuitry through changes in synapticefficiency ⁄ weights (Monfils et al., 2005; Newell et al.,2003; Takahashi et al., 2002).

Conclusions

For the first time, we demonstrate that the maturemotor systems of adults show neural plasticity, similar tothat seen in children, with the repetition of phone-mically complex nonwords. The findings also suggestthat the phonemic composition of the nonwordsdetermines the extent of the effect of repetition onmovement coordination as well as the extent ofconsolidation of learning, thereby supporting a role forspeech motor output processes within models ofnonword repetition. Our study demonstrates that theshort-term changes observed within a single sessionpersist the next day. We interpret this to suggest thatthe changes that occurred in the short experimentalsession were associated with underlying changes in theneural circuitry that controls speech movements. Onepossibility could be that the short-term changes do nottranslate to persistent, stored changes in patterns, inwhich case, the lip aperture variability would have beenidentical for Days 1 and 2. Thus we can hypothesize thatthe short-term changes do reflect changes in synapticdensities and weights. With consolidation, our patterncorrelation analysis indicated that changes in motorcommands to muscles occur in a systematic way to reachan optimal coordinative pattern in both children andadults. Future lines of research are proposed to furtherunderstand the nature of speech motor learning and thechanges associated with the learning of novel phoneticstrings in children and adults.

Acknowledgements

This study was funded by an NIH grant (# DC00559)awarded to the second and fourth authors. We expressgratitude to our participants and acknowledge BarbaraBrown, Janna Berlin, and Bridget Walsh for technicalsupport and data collection.

References

Anderson, J.D., Wagovich, S.A., & Hall, N.E. (2006).Nonword repetition skills in young children who do andnot stutter. Journal of Fluency Disorders, 31 (3), 177– 199.

Baddeley, A.D., & Wilson, B.A. (1993). A developmental def-icit in short-term memory: implications for language andreading. Memory, 1, 65–78.

Bartlett, M.S. (1947). The use of transformations. Biometrics, 3,39–52.

Bishop, D.V.M., North, T., & Donlan, C. (1996). Nonwordrepetition as a behavioural marker for inherited languageimpairment: evidence from a twin study. Journal of ChildPsychology and Psychiatry, 37, 391–403.

Butterworth, B., Campbell, R., & Howard, D. (1986). The usesof short-term memory: a case study. Quarterly Journal ofExperimental Psychology, 38, 705–737.

Dollaghan, C., & Campbell, T.F. (1998). Nonword repetitionand child language impairment. Journal of Speech, Language,and Hearing Research, 41, 1136–1146.

Edwards, J., Beckman, M.E., & Munson, B. (2004). Theinteraction between vocabulary size and phonotactic proba-bility effects on children’s production accuracy and fluency innonword repetition. Journal of Speech, Language, and Hear-ing Research, 47 (2), 421–436.

Gaskell, G.M., & Dumay, N. (2001). Lexical competitionand the acquisition of novel words. Cognition, 89, 105–132.

Gathercole, S.E. (2006). Complexities and constraints in non-word repetition and word learning. Applied Psycholinguistics,27 (4), 599–613.

Gathercole, S.E., & Baddeley, A.D. (1990). The role of pho-nological memory in vocabulary acquisition: a study ofyoung children learning new names. British Journal of Psy-chology, 81, 439–454.

Gathercole, S.E., Briscoe, J., Thorn, A., & Tiffany, C. (2008).Deficits in verbal long-term memory and learning in childrenwith poor phonological short-term memory skills. QuarterlyJournal of Experimental Psychology, 61 (3), 474–490.

Gathercole, S.E., Frankish, C., Pickering, S.J., & Peaker, S.(1999). Phonotactic influences on short-term memory.Journal of Experimental Psychology: Learning, Memory, andCognition, 25, 84–95.

Gathercole, S.E., Hitch, G.J., Service, E., & Martin, A.J.(1997). Short-term memory and new word learning in chil-dren. Developmental Psychology, 33, 966–979.

Gathercole, S.E., Willis, C., Emslie, H., & Baddeley, A.D.(1994). The Children’s Test of Nonword Repetition: a test ofphonological working memory. Memory, 2, 103–127.

Goffman, L., & Smith, A. (1999). Development and phoneticdifferentiation of speech movement patterns. Journal ofExperimental Psychology: Human Perception and Perfor-mance, 25 (3), 649–660.

Green, J.R., Moore, C.A., & Reilly, K.J. (2002). The sequentialdevelopment of jaw and lip control for speech. Journal ofSpeech, Language, and Hearing Research, 45 (1), 66–79.

Gupta, P. (2003). Examining the relationship between wordlearning, nonword repetition, and immediate serial recall inadults. Quarterly Journal of Experimental Psychology, 56A,1213–1236.

Gupta, P. (2006). Nonword repetition, phonological storage,and multiple determination. Applied Psycholinguistics, 27 (4),564–568.



Gupta, P., & MacWhinney, B. (1997). Vocabulary acquisitionand verbal short-term memory: computational and neuralbases. Brain and Language, 59 (2), 267–333.

Hakim, H.B., & Ratner, N.B. (2004). Nonword repetitionabilities of children who stutter: an exploratory study. Journalof Fluency Disorders, 29 (3), 179–199.

Harris, C.M., & Wolpert, D.M. (1998). Signal-dependent noisedetermines motor planning. Nature, 394, 780–784.

Jones, G., Gobet, G., & Pine, J.M. (2007). Linking workingmemory and long-term memory: a computational model ofthe learning of new words. Developmental Science, 10 (6),853–873.

Kleim, J.A., Bruneau, R., Calder, K., Pocock, D., VandenBerg,P.M., MacDonald, E., Monfils, M.H., Sutherland, R.J., &Nader, K. (2003). Functional organization of adult motorcortex is dependent upon continued protein synthesis. Neu-ron, 40 (1), 167–176.

Kleim, J.A., Hogg, T.M., VandenBerg, P.M., Cooper, N.R.,Bruneau, R., & Remple, M. (2004). Cortical synaptogenesisand motor map reorganization occur during late, but notearly, phase of motor skill learning. Journal of Neuroscience,24 (3), 628–633.

Leach, L., & Samuel, A.G. (2007). Lexical configuration andlexical engagement: when adults learn new words. CognitivePsychology, 55, 306–353.

Levelt, W.J.M., Roelofs, A., & Meyer, A.S. (1999). A theory oflexical access in speech production. Behavioral and BrainSciences, 22 (1), 1–37.

McNamara, A., Tegenthoff, M., Dinse, H., Buchel, C., Bin-kofski, F., & Ragert, P. (2007). Increased functional con-nectivity is crucial for learning novel muscle synergies.NeuroImage, 35 (3), 1211–1218.

Miall, R.C., & Wolpert, D.M. (1996). Forward models forphysiological motor control. Neural Networks, 9 (8), 1265–1279.

Michas, I.C., & Henry, L.A. (1994). The link between phono-logical memory and vocabulary acquisition. British Journal ofDevelopmental Psychology, 12, 147–164.

Monfils, M.H., Plautz, E.J., & Kleim, J.A. (2005). In searchof the motor engram: motor map plasticity as a mecha-nism for encoding motor experience. Neuroscientist, 11 (5),471–483.

Munson, B. (2001). Phonological pattern frequency and speechproduction in adults and children. Journal of Speech,Language, and Hearing Research, 44 (4), 778–792.

Newell, K.M., Liu, Y.T., & Mayer-Kress, G. (2003). Adynamical systems interpretation of epigenetic landscapes forinfant motor development. Infant Behavior and Development,26 (4), 449–472.

Papagno, C., & Vallar, G. (1995). Verbal short-term memoryand vocabulary learning in polyglots. Quarterly Journal ofExperimental Psychology, 48A, 98–107.

Schmidt, R.A. (1988). Motor control and learning: A behavioralemphasis (2nd edn.). Champaign, IL: Human Kinetics.

Schulz, G.M., Stein, L., & Micallef, R. (2001). Speech motorlearning: preliminary data. Clinical Linguistics and Phonetics,15 (1), 157–161.

Semel, S., Wiig, E.H., & Secord, W.A. (1995). Clinical Evalu-ation of Language Fundamentals-3. San Antonio, TX: Har-court Brace & Company.

Sevald, C.A., Dell, G.S., & Cole, J. (1995). Syllable structure inspeech production: are syllables chunks or schemas? Journalof Memory and Language, 34, 807–820.

Smith, A., Goffman, L., Zelaznik, H.N., Ying, G.S., &McGillem, C. (1995). Spatiotemporal stability and patterningof speech movement sequences. Experimental Brain Research,104 (3), 493–501.

Smith, A., Johnson, M., McGillem, C., & Goffman, L. (2000).On the assessment of stability and patterning of speechmovements. Journal of Speech, Language, and Hearing Re-search, 43 (1), 277–286.

Smith, A., & Zelaznik, H.N. (2004). Development of func-tional synergies for speech motor coordination in child-hood and adolescence. Developmental Psychobiology, 45 (1),22–33.

St Louis, K.O., & Ruscello, D.M. (1987). Oral SpeechMechanism Screening Evaluation-Revised. Austin, TX:Pro-Ed.

Storkel, H.L. (2001). Learning new words: phonotactic prob-ability in language development. Journal of Speech, Lan-guage, and Hearing Research, 44 (6), 1321–1337.

Storkel, H.L., Armbruster, J., & Hogan, T.P. (2006). Differ-entiating phonotactic probability and neighborhood densityin adult word learning. Journal of Speech, Language, andHearing Research, 49 (6), 1175–1192.

Studebaker, G.A. (1985). A ‘rationalized’ arcsin transform.Journal of Speech and Hearing Research, 28, 455–462.

Takahashi, C.D., Nemet, D., Rose-Gottron, C.M., Larson,J.K., Cooper, D.M., & Reinkensmeyer, D.J. (2002). Neuro-motor noise limits motor performance, but not motoradaptation, in children. Journal of Neurophysiology, 90, 703–711.

Van Galen, G.P., Portier, S.J., Smitsengelsman, B.C.M., &

Schomaker, L.R.B. (1993). Neuromotor noise and poorhandwriting in children. Acta Psychologica, 82 (1–3), 161–178.

Vitevich, M.S., & Luce, P.A. (2004). A Web-based interface tocalculate phonotactic probability for words and nonwords inEnglish. Behavior Research Methods, Instruments, & Com-puters, 36 (3), 481–487.

Walker, M.P., Brakefield, T., Morgan, A., Hobson, J.A., &Stickgold, R. (2002). Practice with sleep makes perfect: sleep-dependent motor skill learning. Neuron, 35 (1), 205–211.

Walker, M.P., Stickgold, R., Alsop, D., Gaab, N., & Schlaug,G. (2005). Sleep-dependent motor memory plasticity in thehuman brain. Neuroscience, 133 (4), 911–917.

Walsh, B., & Smith, A. (2002). Articulatory movements inadolescents: evidence for protracted development of speechmotor control processes. Journal of Speech, Language, andHearing Research, 45 (6), 1119–1133.

Walsh, B., Smith, A., & Weber-Fox, C. (2006). Short-termplasticity in children’s speech motor systems. DevelopmentalPsychobiology, 48 (8), 660–674.

Wechsler, D. (1997). Wechsler Memory Scale-Third Edition.San Antonio, TX: Psychological Corporation.

Yan, J.H., Thomas, J.R., Stelmach, G.E., & Thomas, K.T.

(2000). Developmental features of rapid aiming arm move-ments across the lifespan. Journal of Motor Behavior, 32,121–140.

Yoss, K.A., & Darley, F.L. (1974). Developmental apraxia ofspeech in children with defective articulation. Journal ofSpeech and Hearing Research, 17, 399–416.

Received: 18 July 2008Accepted: 5 April 2009



paper nonword repetition in children and adults: … nonword repetition in children and adults:...

Documents