Vous êtes sur la page 1sur 10

Research in Developmental Disabilities 30 (2009) 15021511

Contents lists available at ScienceDirect

Research in Developmental
Disabilities

Predictors of verbal working memory in children with


cerebral palsy
Marieke Peeters *, Ludo Verhoeven, Jan de Moor
Behavioural Science Institute, Radboud University Nijmegen, The Netherlands

A R T I C L E I N F O A B S T R A C T

Article history: The goal of the present study was to examine the precursors of verbal
Received 3 July 2009 working memory in 52 children with cerebral palsy with varying
Accepted 10 July 2009 degrees of speech impairments in the rst grade of special education.
Following Baddeleys model of working memory, childrens verbal
Keywords: working memory was measured by means of a forced-recognition
Cerebral palsy task. As precursors of verbal working memory, measures of
Verbal working memory intelligence, speech rate, speech intelligibility, auditory perception,
Intelligence
and phonological awareness were also administered. Correlations
Speech
were computed between all measures followed by Structural
Equation Modeling analyses with speech rate and speech intellig-
ibility being identied as a single factor speech. The results revealed
that verbal working memory was mostly predicted by intelligence,
auditory perception and speech ability. It was also found that children
with cerebral palsy with additional intellectual and speech impair-
ments were at risk for limited verbal working memory spans.
2009 Elsevier Ltd. All rights reserved.

1. Introduction

Working memory is used in situations in which small amounts of material are held actively and
then reproduced in a sequential or unchanging form. Baddeley and Hitch (1974) proposed a model of
working memory (WM) that consists of three aspects: a visuospatial sketchpad that is assumed to be
capable of temporarily maintaining and manipulating visuospatial information (Baddeley, 2000), a
central executive responsible for the coordination of the visuospatial sketchpad and the phonological
loop and their connection to the long-term memory. The phonological loop consists of a phonological
store, specialized in the storage of verbal material or recoding of nonverbal materials, and a

* Corresponding author at: Behavioural Science Institute, Radboud University Nijmegen, Department of Special Education,
P.O. Box 9104, 6500 HE Nijmegen, The Netherlands. Tel.: +31 0243612822; fax: +31 0243616211.
E-mail address: m.peeters@pwo.ru.nl (M. Peeters).

0891-4222/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ridd.2009.07.014
M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511 1503

phonological rehearsal component. In this rehearsal component, materials are represented in a


phonological code and rehearsed in order to prevent the traces from decaying (Baddeley, Gathercole, &
Papagno, 1998). This phonological or articulatory loop is considered to play an important role in
learning to read and write (Card & Dodd, 2006) as it might facilitate long-term phonological learning
(Baddeley, Papagno, & Vallar, 1988). For instance, the phonological loop plays an important role in
learning phonemegrapheme conversion rules necessary for word decoding and ultimately text-
comprehension (Baddeley, 2003).
Previous studies to the precursors of verbal WM have indicated four skills that play an important
role in its development. First of all, previous studies indicate a strong relationship between general
intelligence and verbal WM (cf. Colom, Rebollo, Palacios, Juan-Espinosa, & Kyllonen, 2004). In a study
of Henry and MacLean (2002), it was shown that children with intellectual disabilities scored below a
mental-age matched control group on different memory tasks.
Secondly, articulatory coding is considered to play a role in the size of memory spans. Studies
suggested that the memory span is equivalent to the number of items that can be pronounced in about
2 s (White, Craft, Hale, & Park, 1994). Both the word-length effect, i.e., the memory span for long words
is less than for short words as it takes more time to pronounce long words (Bishop, 1988), and the
phonological similarity effect, i.e., immediate serial recall is impaired when items are similar in sounds
or articulatory characteristics, is abolished by normal subjects when a visual presentation is used and
subjects are required to articulate concurrently (Bishop & Robson, 1989). These results gave rise to the
view that articulatory coding, i.e., coding of information as a set of motor commands to articulators
(Bishop, 1985, p. 245) is involved in verbal WM. However, it is not just the motor ability to articulate
but also speech output that is associated with verbal WM capacity. Previous studies have shown that
there is a relationship between articulation speed, or speech rate, and the memory span. For example,
Kail (1997) conducted a study focusing on the memory span of 120, 6- to 10-year-old children and
found their memory span, i.e., the longest sequence of words that can be correctly repeated in a serial
order, was related to their speech rate. The faster the children could articulate, the more words they
could rehearse within a xed time, and therefore the longer their memory spans. Other studies have
also indicated a signicant relationship between speech rate and memory span (Hulme, Thomson,
Muir, & Lawrence, 1984; Roodenrys, Hulme, & Brown, 1993; White et al., 1994). Another aspect of
speech that has shown relationships with verbal WM is articulation quality. Rapala and Brady (1990)
studied the relationship between the quality of articulation, measured with tongue-twister word pairs
such as [si] and [shi], and verbal STM. They found that the better the children could articulate, the
longer their memory spans. Locke and Kutz (1975) showed that children with a speech disorder were
worse in indicating the right series of pictures of spoken words compared to a control group. They
proposed that the inability to articulate could inuence the efciency of storing words in the
phonological loop.
A third precursor of verbal WM is auditory perception. Studies have indicated that it is necessary
for children to be able to hear and process incoming speech sounds correctly in order to have access to
and retrieve the phonological representations from the long-term memory (Foy & Mann, 2001) and
prevent the memory traces from decaying (cf. Brady, Shrankweiler, & Mann, 1983; Rapala & Brady,
1990).
Finally, studies have shown a relationship between verbal WM and phonological awareness.
Alloway et al. (2005) have shown that phonological skills, such as phonological awareness, are related
to memory span.
Taking into account the importance of these precursors for verbal WM, the question can be asked
how the verbal WM skills of children with cerebral palsy (CP) are developed and what role these
precursors play in the verbal WM spans of these children. According to Bax et al. (2005, p. 572),
cerebral palsy can be described as a group of disorders of the development of movement and posture,
causing activity limitation that are attributed to non-progressive disturbances that occurred in the
developing fetal or infant brain. The motor disorders of CP are often accompanied by disturbances of
sensation, cognition, communication, perception, and/or behaviour, and/or by seizure disorders. In
addition, children with CP often have dysarthria, a speech production problem that may be a direct
result of the motor impairment (Pirila et al., 2007). Kent (2000, p. 399) described dysarthrias as
speech disorders that result from neurologic impairments associated with weakness, slowness, or
1504 M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511

incoordination of the musculature used to produce speech. The disturbed neuromotor control of the
speech mechanisms may in a minority of the children result in such severe speech impairments that
these children are unable to speak and are therefore called nonspeaking children. This may have
consequences for the development of the verbal WM skills of these children.
From previous studies, it is known that children with CP lag behind in their verbal WM skills
(Dahlgren Sandberg, 2002; Peeters, Verhoeven, van Balkom, & de Moor, 2008) what might have
consequences for their reading development as working memory skills are important for the
consolidation of phonemegrapheme conversion rules and the development of a large vocabulary
(Baddeley & Hitch, 1974; Baddeley et al., 1998). However, till now it is not clear if and how the
additional speech and intellectual impairments play a role in the development of verbal WM of
children with CP. Previous studies towards the precursors of verbal WM in children with CP are scarce.
Experimental studies that have been conducted are mainly focused on the role of speech abilities in
verbal WM. For instance, it is speculated that if articulatory coding is necessary for subvocal rehearsal
then subjects who have never been able to articulate words, such as children with anarthria, would
show no or a reduced effect of word-length and phonological similarity (cf. Bishop & Robson, 1989).
Bishop and Robson (1989) studied the memory spans of 12 adolescents with dysarthria and 12
adolescent with anarthria with 24 adolescents matched for chronological age and nonverbal ability.
Both the control and the speech-impaired group showed evidence of phonological similarity and
word-length effects. In a similar vein, Foley and Pollatsek (1999) compared six adults with dysarthria
and six adults with anarthria on memory processes. Phonological similarity effects were signicantly
greater for the individuals with dysarthria. However, also two adults with anarthria demonstrated
signicantly phonological similarity effects. Similarly, word-length effects were apparent in the group
of adults with dysarthria, while also one individual with anarthria showed signicant word-length
effects.
The results of previous studies make it clear that articulatory coding is not essential for subvocal
rehearsal (Bishop & Robson, 1989; Vallar & Cappa, 1987). Therefore, researchers have suggested that
processes such as rehearsal involve a more abstract central phonological code (Bishop, 1985; Bishop,
1988; Bishop & Robson, 1989). Regarding the role of speech rate in verbal WM, Raine, Hulme,
Chadderton, and Bailey (1991) investigated the verbal STM spans of speech-disordered children and
suggested that speech rate may be a causal determinant of verbal STM capacity. The faster the children
could articulate, the more words could be rehearsed and the less will be faced in the phonological loop.
Conclusively, the role of diverse precursors in the verbal WM capacity of children is far from
conclusive from previous studies. Bearing in mind the often interrelated additional impairments of
children with CP, it is necessary to study all the precursors simultaneously in one-design. The present
study is a nationwide study of the verbal STM skills of children with CP in Grade 1. In the present study,
an attempt has been made to overcome previous limitations by investigating the verbal WM skills and
its precursors such as intelligence, speech, auditory perception, and phonological awareness by a large
group of children with CP with varying intellectual and speech abilities. The main research question of
this study was: To what extent can the variation in verbal WM in Grade 1 be explained from its
precursors?
It is expected that, rst of all, intelligence will play an important role in the memory spans of
children with CP as working memory is a cognitive process. Taken into account the role of IQ for the
capacity of the verbal memory spans, it is expected that other precursors would still be related to
verbal WM. First of all, speech abilities will play a major role in the verbal WM spans of these children
when controlling for intellectual skills. In addition, a major role is ascribed to auditory perception in
the process of remembering words.

2. Method

All 32 schools for children with physical and multiple disabilities in the Netherlands were asked to
participate and obtain written consent from the parents of children with CP who t the following
inclusion criteria: Dutch had to be their native language, the intellectual level must range from a mild
intellectual disability to average intelligence or above, hearing and vision must be within the normal
range, with the ability to respond intentionally, either through speaking or by means of alternative
M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511 1505

communication (e.g., looking, pointing or gesturing), and they must be 5-years-old at the beginning of
the longitudinal study. Although the present study is conducted in Grade 1, the children of the present
study were all participating in a larger longitudinal study to the emergent literacy development of
children with cerebral palsy (cf. Peeters, Verhoeven, de Moor, & van Balkom, 2009). Fifty-two children
with CP participated in the present study; 33 boys (63.5%) and 19 (36.5%) girls. Fifty children (96.1%)
had spastic CP and 2 children had ataxia (3.8%). Of the children with spastic CP, 13 children (26.0%) had
quadriplegia, 22 children (44%) had diplegia, 8 children (16%) had hemiplegia, 5 children (10%) had a
combination of spastic-ataxic CP, 1 child (2%) had spastic-hypotonia CP, and 1 child (2%) had spastic-
dyskinetic CP. Nine of the 52 children (17.3%) had seizures. The speech-language therapists of the
children reported that 24 children (46.2%) had no speech difculties, 10 children (19.2%) had mild
dysarthria, 6 children (11.5%) had moderate dysarthria, 9 children (17.3%) had severe dysarthria and
were unable to speak, 1 child (1.9%) had moderate dyspraxia, and 2 children (3.8%) had a combination
of severe dysarthria and dyspraxia. The average age of the children in Grade 1 was 84 months
(SD = 6.1). Twelve children (23.1%) use some sort of Augmentative and Alternative Communication
(AAC) to communicate. All children with CP attended special schools for children with physical and
multiple disabilities across the Netherlands. The average score of the ne motor function as measured
by the Dutch version of the Manual Ability Classication System (Eliasson et al., 2006), which ranges
from one (low impairment) to ve (high impairment), was 2.52 (SD = .94). The average gross motor
skills as measured by the Dutch version of the Gross Motor Function Classication System (Palisano
et al., 2000), which also ranges from one to ve, was 2.71 (SD = 1.26). All parents of the children in the
present study had given written consent for their children to participate.

2.1. Materials

2.1.1. Speech intelligibility


In order to assess childrens articulation quality or speech intelligibility, the standardized Word
Articulation task of the SLI Screening test was administrated (Verhoeven, 2006). In the Word
Articulation task the child was asked to repeat real words. Words were presented one-by-one by a
computer with recorded voice, whereby the task started with words containing only one syllable
and increased to words containing up to ve syllables. When a child made ve successive errors,
the task was ended. Eighty-four percent of the children with CP were able to do this task since they
had some level of understandable speech; the other children who were given a score of zero. The
maximum score was 40. The test manual reported a good internal consistency with a Cronbachs
alpha of .94.

2.1.2. Speech rate


To measure the speed of articulation, a task was constructed in which children had to pronounce a
given word within a given time frame. The children saw a picture of a familiar word that consisted of
three syllables and they had to pronounce that word as frequently as possible within 10 s, after the test
leader said start and pushed the stopwatch. During those 10 s the children could look at the picture, in
order to relive their memory as much as possible. All words consisted of three syllables. The practice
item was a picture of a telephone [te-le-foon]. The test items were pictures of peanut butter [pin-da-
kaas], sugar bowl [sui-ker-pot], and farm [boer-de-rij]. The amount of fully pronounced words of the
three items was counted up to create a total score. The internal consistency of this task was very high
with a Cronbachs alpha of .96.

2.1.3. Auditory perception


To assess auditory perception abilities, the Auditory Discrimination Task of the standardized Dutch
Language Prociency Test was administrated (Verhoeven & Vermeer, 2001). In this task, the child was
presented with minimally differing word pairs and had to indicate whether the words in a pair
sounded alike. Response adaptations for children with speech difculties consisted of nodding or
pointing to left or right to indicate if the words sounded the same or different. All items were tested
and the maximum score was 50. The task was highly reliable with a Cronbachs alpha of .97
(Verhoeven & Vermeer, 1999).
1506 M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511

2.1.4. Phonological awareness


To test phonological awareness skills, the rst-phoneme recognition task was used (de Jong, van
Otterloo, & Regtvoort, 2006). This task consists of 10 items with CVC words. Each item consists of ve
pictures, one stimulus picture and four response pictures. During the exercise items, the test assistant
pointed at the stimulus picture and named that picture (e.g., roof). Subsequently, the test assistant
explained that the stimulus word could be split up in two parts, the rst-phoneme of that word versus
the rest of the word (e.g., rrrrr-oof). The test assistant subsequently named all four response pictures
with explicitly emphasis on the rst-phoneme of the words. The child had to point at one of the four
response pictures that started with the same rst-phoneme as the stimulus word (i.e., r). During the
test items, the test assistant named the stimulus picture and the response pictures without explicitly
emphasizing the rst-phoneme of the word. The internal consistency of this task in this study was
good with a Cronbachs alpha of .83.

2.1.5. Intelligence
Nonverbal reasoning was measured with the Raven Coloured Progressive Matrices (Raven, 1956).
This task was used as it correlated highly with general intelligence (Duncan et al., 2000). The task
measures nonverbal reasoning with a minimal interference of language and is a commonly used
instrument to assess intelligence, or general reasoning ability, in the nonspeaking population (Pueyo,
Junque, Vendrell, Narberhaus, & Segarra, 2008). Children were asked to point, aided or unaided, to one
of the six pictures that completed the presented gure. The task consisted of 36 items. Raw score were
converted to standard scores ranging from .5 to 9.5 using Dutch norms (van Bon, 1986).

2.1.6. Verbal working memory (WM)


To be able to assess the verbal WM abilities of both speaking and nonspeaking children, a forced-
choice recognition task was developed based on a serial-recognition experiment of Gathercole,
Pickering, Hall, and Peaker (2001) that did not require physical or speech production abilities. In the
newly constructed task, the child heard a string of words and after 2 s the child heard another string of
words (e.g., [boat], [knife], [cap], versus [boat], [window], [cap]). The child had to decide whether the
two successive strings of words were identical or not. The task consisted of strings of words that
increased in length, starting with a length of one word and increasing to a length of eight words. For
this task, a set of 10 highly frequent monosyllabic consonantvowelconsonant (CVC) words were
used which occurred in a list of words used in the context of kindergarten education (Schaerlaekens
et al., 1999) and differed phonologically and semantically as much as possible from each other. There
were a total 48 items; 6 items of each string length. If the child had only three or less items of a string
length correct, the task was ended. The internal consistency of this task in this study was very high
with a Cronbachs alpha of .95.

2.2. Procedure

All tasks were administrated at the end of Grade 1. All children were individually tested in a quiet
room in their schools by a trained test assistant; an assistant teacher was present as well. Prior to each
task, there was a training phase, to make sure the children understood the test. The test order was the
same for all children and the tests were divided over two or three sessions. Response adaptations for
children with speech difculties were nodding or pointing, aided or unaided. This study was proposed
to the Dutch Association of Rehabilitation Physicians [Vereniging van Revalidatieartsen] and the Dutch
Association of Schools for Children with Physical Disabilities [Vereniging van Mytyl- en Tyltylscholen]
who gave their full ofcial cooperation. This study did not need permission of the medical ethical
committee because only behavioral data and existing medical information was collected.

2.3. Statistical methods

Statistical analyses were performed to provide answers to the research questions. First, descriptive
statistics (means, standard deviations, skewness, kurtosis, minimal and maximal value) were
computed for intelligence, speech intelligibility, speech rate, auditory perception, phonological
M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511 1507

awareness, and verbal WM. To examine the variation in verbal WM and to inspect the data for possible
colinearity, Pearson correlations were computed between all measures. Subsequently, a Multiple
Regression Analysis was conducted by means by using AMOS 5 for Structural Equation Modeling
(SEM) (Raykov & Marcoulides, 2006). SEM not only affords to account for measurement errors while
examining the inuence of the predictor variables on verbal WM (Baron & Kenny, 1986; Wagner,
Torgesen, & Rashotte, 1994) but also to look at the independent contribution made by each predictor,
and the interrelationships among the predictor variables (Arbuckle & Wothke, 1999). The Goodness of
Fit of the estimated model was assessed by seven t indices: x2, with degrees of freedom and p-value,
Goodness of Fit Index (GFI), Adjusted Goodness of Fit Index (AGFI), Normed Fit Index (NFI),
Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized
Root Mean square Residual (SRMR). A model ts well if the ratio of the Chi-square value to the degrees
of freedom is smaller than two (Ullman, 2001), GFI, NFI, and CFI are between .90 and 1.00, and close to
1.00 (Bentler, 1990; Hu & Bentler, 1999; Jaccard & Wan, 1996), and AGFI is greater than .85 (Kline,
1998). In addition, Hu and Bentler (1999) advised a cutoff value of .06 for RMSEA and .08 for SRMR.

3. Results

3.1. Descriptive statistics of verbal WM and its predictor variables

Table 1 shows the descriptive statistics of verbal WM and its predictor variables. As can be seen, the
skewness and kurtosis values show a good normal distribution of each variable.
Table 2 shows the correlations between all variables. As is shown in Table 2, verbal WM and its
predictor variables are highly signicant correlated. Table 2 shows that all predictor variables were
signicant correlated with verbal WM and therefore should all be included into the Multiple
Regression Analysis to uncover what predictor variable or variables could best predict verbal WM
skills in Grade 1. However, the predictor variables show also highly signicant intercorrelations.
Pearson correlation shows that speech intelligibility and speech rate were highly correlated (r = .86,
p < .001). As these two variables are highly correlated and tap into the same pool of variance, putting
these variables together in a regression analysis results in an unstable regression model (see for details
Belsey, Kuh, & Welsch, 1980). Therefore, it was decided not to enter these variables separately into the
model to predict verbal WM but to examine if these variables could be entered into the model by
means of a latent factor called speech.

Table 1
Descriptive statistics of verbal working memory and its predictor variables.

M SD Skewness Kurtosis Min Max

Intelligence 2.29 1.68 .95 .77 .5 7.2


Speech intelligibility 21.38 15.14 .33 1.58 0 40
Speech rate 20.52 14.00 .27 1.20 0 44
Auditory perception 37.69 10.79 .99 .26 13 50
Phonological awareness 5.73 3.32 .17 1.17 0 10
Verbal WM 21.94 9.33 .17 .51 3 42

Table 2
Pearson correlations between all predictor variables and verbal WM.

1 2 3 4 5 6

1 IQ
2 Speech intelligibility .18
3 Speech rate .24 .86***
4 Auditory perception .38** .45** .51***
5 Phonological awareness .33* .61*** .65*** .57***
6 Verbal WM .48*** .57*** .63*** .73*** .63***
*
p< .05.
**
p< .01.
***
p< .001.
1508 M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511

Fig. 1. A SEM analysis showing the inuence of speech, auditory perception, phonological awareness, and intelligence on verbal
WM. The numbers indicate standardized beta coefcients.

3.2. Regression analysis

In order to predict verbal WM skills by its predictor variables, a SEM analysis was conducted with
one latent variable, the factor called speech consisting of speech rate and speech intelligibility, and
three observed predictor variables, i.e., intelligence, auditory perception, and phonological awareness.
The correlations between the variables are presented in Appendix A. First of all, the latent variable
speech turned out to be a good factor of speech intelligibility and speech rate as it explained 91.8% of
the variation in speech rate and 81.0% of the variation in speech intelligibility. The beta coefcients
were .90 (p < .001) for speech intelligibility and .96 (p < .001) for speech rate.
The results show that the SEM model has a good t, x2(3) = .603, p = .896, GFI = 1.00, AGFI = .97,
NFI = 1.00, CFI = 1.00, RMSEA = .00, and SRMR = .01, indicating that the data ts very well with the tested
model. Fig. 1 shows the model. As can be seen in Fig. 1, three out of four predictor variables turned out
to be signicantly related to verbal WM when taken into account the other predictor variables,
indicating that these variables were all explaining unique variance in verbal WM skills. The best
predictor of verbal WM turned out to be auditory perception (b = .43, p < 0.001), followed by the latent
factor speech (b = .29, p < .05), intelligence (b = .21, < .05), and nally, phonological awareness (b = .12,
p > .05). In this model, 68.2% of the variance in verbal WM is explained by its predictor variables.

4. Discussion and conclusions

From the present study several conclusions can be drawn. First of all, correlation analyses have
indicated a highly signicant correlation between speech intelligibility and speech rate. Considering
the fact that a lot of children with cerebral palsy have dysarthrias coming from neurologic impairment,
it is no surprise that speech intelligibility and speech rate are strongly related as the speech
impairments of these children have a physical or motor origin. When children have problems with the
oral motor coordination to produce speech this will negatively affect their speech intelligibility as well
as the speed at which they are able to articulate.
Regarding the prediction of verbal WM, auditory perception turned out to be the best predictor
when taken into account other predictor variables. The fact that besides intelligence, auditory
perception still predicts performance of the verbal WM capacity indicates that the ability to process
incoming auditory speech signals and having access to the phonological representations of the words
in long-term memory is highly predictive of the verbal WM span. The better children with CP are in
processing these incoming speech signals and match them with existing phonological representations
in the brain, the better these signals will be rehearsed and ultimately remembered. These results are in
accordance with previous studies of Brady et al. (1983) and Rapala and Brady (1990) conducted with
normally developing children.
M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511 1509

The fact that the latent factor speech turned out to be a predictor of verbal WM, indicates that
speaking is not an all-or-nothing phenomenon as previous research has suggested. For example,
previous experimental studies indicate that the ability to speak and articulatory coding is not
necessary for subvocal rehearsal. The fact that there is a linear relationship between speaking and
verbal WM skills indicates that the better the speech abilities are of the children with CP, the better
they are in remembering words. Tentatively, it can be concluded that the ability to speak facilitates
childrens ability to remember words.
Of course some limitations apply to the present study. One limitation deals with the nature of the
verbal WM task. As a lot of children with CP were having speech impairments, it was not possible to
develop a task that requires a verbal response as most verbal WM tasks do. Therefore, we decided to
construct a forced-choice task where the child had to indicate, by means of looking, talking or pointing,
if two strings of words were identical or not. This made the task a bit different from the well known
digit span tasks. Besides, the child had to do more then just remembering the strings of words as he or
she had to compare the strings in order to decide if they were identical or not. Another possibility
could have been to construct a task in which children had to point to the pictures they just heard, but
taken into account the physical impairments of these children, using this task had more to do with
motor speed than verbal memory.
The results of the present study have clinical implications. To begin with, the fact that auditory
perception is strongly related to verbal WM capacity indicates that children with low levels of auditory
perception skills are at risk for memory problems. Therefore, early intervention is needed in order to
prevent memory difculties and ultimately reading impairments. A possibility of stimulating the
auditory perception skills of children is using education software such as LinguaBytes in which auditory
discrimination skills are trained (Hengeveld et al., 2008; Voort et al., 2008). Another way of improving
auditory perception skills is slowing down or stretching the entire speech signal of words (Segers &
Verhoeven, 2005) a method used in the remediation program called Fast ForWord (Tallal et al., 1996). An
intervention based on this principle could improve the auditory perception skills of these children.
The fact that speech abilities are strongly predictive of verbal memory skills shows that it is
necessary for children with speech impairments to improve their speech abilities in order to prevent
them for memory problems and ultimately reading difculties. One way of improving the speech
abilities for nonspeaking children is by providing them with AAC devices with a speech-output
component (cf. Millar, Light, & Schlosser, 2006). Speech-language therapists as well as parents play an
important role in improving the speech abilities of these children. Although their physical
impairments limit them for being able to speak, the more frequently they hear the sounds the
better they will be able to use these sounds in speech output or subvocal rehearsal.

Acknowledgements

This study is funded by the European Union, Interreg 4-BMG-V-1 = 31 and Viataal (St.
Michielsgestel, The Netherlands). Special appreciation is expressed to the children who participated,
and the clinical experts of the participating schools.

Appendix A

Input correlation matrix for Structural Equation Modeling analyses.

1 2 3 4 5 6 7

1 Phonological awareness 1.00


2 Auditory perception .57 1.00
3 Intelligence .33 .38 1.00
4 Speech .67 .52 .24 1.00
5 Verbal WM .63 .73 .48 .65 1.00
6 Speech rate .65 .50 .23 .96 .63 1.00
7 Speech intelligibility .60 .47 .22 .90 .58 .86 1.00
1510 M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511

References

Alloway, T. P., Gathercole, S. E., Adams, A., Willis, C., Eaglen, R., & Lamont, E. (2005). Working memory and phonological awareness as
predictors of progress towards early learning goals at school entry. British Journal of Developmental Psychology, 23, 417426.
Arbuckle, J. L., & Wothke, W. (1999). Amos 4.0 users guide. Chicago: SPSS/SmallWaters.
Baddeley, A. D. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4, 417423.
Baddeley, A. D. (2003). Working memory and language: An overview. Journal of Communication Disorders, 361, 189208.
Baddeley, A., Gathercole, S. E., & Papagno, C. (1998). The phonological loop as a language learning device. Psychological Review, 105,
158173.
Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. A. Bower (Ed.), The psychology of learning and motivation, Vol. 8: Advances
in research and theory. Academic Press.
Baddeley, A. D., Papagno, C., & Vallar, G. (1988). When long term learning depends on short-term storage. Journal of Memory and
Language, 27, 586595.
Baron, R. M., & Kenny, D. A. (1986). The moderatormediator variable distinction in social psychological research: Conceptual,
strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182.
Bax, M., Goldstein, M., Rosenbaum, P., Leviton, A., Paneth, N., Dan, B., et al. (2005). Proposed denition and classication of cerebral
palsy. Developmental Medicine and Child Neurology, 47, 571576.
Belsey, D. A., Kuh, E., & Welsch, R. E. (1980). Regression diagnostics: Identifying inuential data and sources of collinearity. New York:
Wiley.
Bentler, P. M. (1990). Comparative t indexes in structural models. Psychological Bulletin, 170, 238246.
Bishop, D. V. (1985). Spelling ability in congenital dysarthria: Evidence against articulatory coding in translating between phonemes
and graphemes. Cognitive Neuropsychology, 2, 229251.
Bishop, D. V. (1988). Language development in children with abnormal structure or function of the speech apparatus. In D. V. Bishop
& K. Mogford (Eds.), Language development in exceptional circumstances. New York: Churchill Livingstone.
Bishop, D. V., & Robson, J. (1989). Unimpaired short-term memory and rhyme judgement in congenitally speechless individuals:
Implications for the notion of articulatory coding. Quarterly Journal of Experimental Psychology: Human Experimental Psychology,
41, 123140.
Brady, S., Shrankweiler, D., & Mann, V. (1983). Speech perception and memory coding in relation to reading ability. Journal of
Experimental Child Psychology, 35, 345367.
Card, R., & Dodd, B. (2006). The phonological awareness abilities of children with cerebral palsy who do not speak. Augmentative and
Alternative Communication, 22, 149159.
Colom, R., Rebollo, R., Palacios, A., Juan-Espinosa, M., & Kyllonen, P. C. (2004). Working memory is (almost) perfectly predicted by g.
Intelligence, 32, 277296.
Dahlgren Sandberg, A. (2002). Phonological recoding problems in children with severe congenital speech impairments: The
importance of productive speech. In E. Witruk, A. D. Friederici, & T. Lachmann (Eds.), Basic functions of language, reading and
reading disability (pp. 315327). Dordrecht, the Netherlands: Kluwer Academic Publishers.
de Jong, P. F., van Otterloo, S. G., & Regtvoort, A.G.F.M. (2006). Fonologische bewustzijn taken voor kleuters: Rijm en beginfoneem
[Phonological awareness tasks for kindergarten children: Rhyme and rst-phoneme]. The Netherlands: University of Amsterdam.
Unpublished Manuscript.
Duncan, J., Seitz, R. J., Kolodny, J., Bor, D., Herzog, H., Ahmed, A., et al. (2000). A neural basis for general intelligence. Science, 289, 457
460.
Eliasson, A. C., Krumlinde-Sundholm, L., Rosblad, B., Beckung, E., Arner, M., Ohrvall, A. M., et al. (2006). The Manual Ability
Classication System (MACS) for children with cerebral palsy: Scale development and evidence of validity and reliability.
Developmental Medicine and Child Neurology, 48, 549554.
Foley, B. E., & Pollatsek, A. (1999). Phonological processing and reading abilities in adolescents and adults with severe congenital
speech impairments. Augmentative and Alternative Communication, 15, 156173.
Foy, J. G., & Mann, V. (2001). Does strength of phonological representations predict phonological awareness in preschool children?
Applied Psycholinguistics, 22, 301325.
Gathercole, S. E., Pickering, S. J., Hall, M., & Peaker, S. M. (2001). Dissociable lexical and phonological inuences on serial decoding and
serial recall. Quarterly Journal of Experimental Psychology: Human Experimental Psychology, 54A, 130.
Hengeveld, B. J., Hummels, C. C. M., Overbeeke, C. J., Voort, R., van Balkom, H., & de Moor, J. (2008). Let me actuate you. Second
International Conference on Tangible and Embedded Interaction, Bonn, Germany, 1820 February 2008: TEI08 (pp. 159166). New
York, NY, USA: ACM.
Henry, L. A., & MacLean, M. (2002). Working memory performance in children with and without intellectual disabilities. American
Journal of Mental Retardation, 6, 421432.
Hu, L. T., & Bentler, P. M. (1999). Cut-off criteria for t indexes in covariance structure analysis: Conventional criteria versus new
alternatives. Structural Equation Modeling, 6, 155.
Hulme, C., Thomson, N., Muir, C., & Lawrence, A. (1984). Speech rate and the development of short-term memory span. Journal of
experimental Child Psychology, 38, 241253.
Jaccard, J., & Wan, C. K. (1996). Lisrel approaches to interaction effects in multiple regression. Thousand Oaks: Sage Publications Inc.
Kail, R. (1997). Phonological skill and articulation time independently contribute to the development of memory span. Journal of
Experimental Child Psychology, 67, 5768.
Kent, R. D. (2000). Research on speech motor control and its disorders: A review and prospective. Journal of Communication Disorders,
33, 391428.
Kline, R. B. (1998). Principles and practice of structural equation modeling. New York: Guilford Press.
Locke, J. L., & Kutz, K. J. (1975). Memory for speech and speech for memory. Journal of Speech and Hearing Research, 18,
176191.
Millar, D., Light, J., & Schlosser, R. (2006). The impact of augmentative and alternative communication intervention on the speech
production of individuals with developmental disabilities: A research review. Journal of Speech Language Hearing Research, 49,
248264.
M. Peeters et al. / Research in Developmental Disabilities 30 (2009) 15021511 1511

Palisano, R. J., Hanna, S. E., Rosenbaum, P. L., Russel, D. J., Walter, S. D., Wood, E. P., et al. (2000). Validation of a model of gross motor
function for children with cerebral palsy. Physical Therapy, 80, 974985.
Peeters, M., Verhoeven, L., de Moor, J., & van Balkom, H. (2009). Importance of speech production for phonological awareness and
word decoding: The case of children with cerebral palsy. Research in Developmental Disabilities, 30, 712726.
Peeters, M., Verhoeven, L., van Balkom, H., & de Moor, J. (2008). Foundations of phonological awareness in children with cerebral
palsy: The impact of intellectual disability. Journal of Intellectual Disability Research, 52, 6878.
Pirila, S., van der Meere, J., Pentikainen, T., Ruusu-Niemi, P., Korpela, R., Kilpinen, J., et al. (2007). Language and motor speech skills in
children with cerebral palsy. Journal of Communication Disorders, 40, 116128.
Pueyo, R., Junque, C., Vendrell, P., Narberhaus, A., & Segarra, D. (2008). Ravens coloured progressive matrices as a measure of
cognitive functioning in cerebral palsy. Journal of Intellectual Disability Research, 52, 437445.
Raine, A., Hulme, C., Chadderton, H., & Bailey, P. (1991). Verbal short-term memory span in speech-disordered children: Implications
for articulatory coding in short-term memory. Child Development, 62, 415423.
Rapala, M. M., & Brady, S. (1990). Reading ability and short-term memory: The role of phonological processing. Reading and Writing:
An Interdisciplinary Journal, 2, 125.
Raven, J. C. (1956). Guide to using the coloured progressive matrices. London: HK Lewis.
Raykov, T., & Marcoulides, G. A. (2006). A rst course in structural equation modeling (2nd ed.). Mahwah, NJ: Psychology Press.
Roodenrys, S., Hulme, C., & Brown, G. (1993). The development of short-term memory span: Separable effects of speech rate and
long-term memory. Journal of Experimental Child Psychology, 56, 431442.
Schaerlaekens, A., Kohnstamm, D., Lejaegere, M., de Vries, A., Peeters, L., & Zink, I. (1999). Streefwoordenslijst voor 6-jarigen (3rd ed.).
Lisse: Swets & Zeitlinger.
Segers, E., & Verhoeven, L. (2005). Effects of lengthening the speech signal on auditory word discrimination in kindergartners with
SLI. Journal of Communication Disorders, 38, 499514.
Tallal, P., Miller, S. L., Bedi, G., Myma, G., Wang, X., Nagarajan, S. S., et al. (1996). Language comprehension in language-learning
impaired children improved with acoustically modied speech. Science, 271, 8184.
Ullman, J. B. (2001). Structural equation modeling. In B. G. Tabachnick & L. S. Fidell (Eds.), Using multivariate statistics (4th ed., pp.
653771). New York: Allyn & Bacon.
Vallar, G., & Cappa, S. F. (1987). Articulation and verbal short-term memory: Evidence from anarthria. Cognitive Neuropsychology, 4,
5578.
van Bon, W. H. J. (1986). Ravens coloured progressive matrices: Manual of Dutch norms. Lisse: Swets & Zeitlinger.
Verhoeven, L. (2006). ESM-Screeningsinstrument [SLI-Screeningtest]. Arnhem: CITO.
Verhoeven, L., & Vermeer, A. (1999). Cognitive factors in early rst and second language prociency. In C. Perdue & R. Rast (Eds.), Actes
du 8e colloque EUROSLA-conference Paris (Vol. 1: La personne bilingue, pp. 97110). Paris: Universite Paris.
Verhoeven, L., & Vermeer, A. (2001). Taaltoets Alle Kinderen (TAK) [Language Prociency Test for All Children]. Arnhem: CITO.
Voort, R., van Balkom, H., Hengeveld, B., Hummels, C., de Moor, J., & Overbeeke, K. (2008, August). LinguaBytes: An interactive, tangible
play and learning system to enhance early literacy and language development of toddlers with multiple disabilities. Paper presented at
the International Society of Augmentative and Alternative Communication (ISAAC) Conference. Montreal, Canada.
Wagner, R. K., Torgesen, J. K., & Rashotte, C. A. (1994). Development of reading-related phonological processing abilities: New
evidence of bidirectional causality from a latent variable longitudinal study. Developmental Psychology, 30, 7387.
White, D. A., Craft, C., Hale, S., & Park, T. S. (1994). Working and articulation rate in children with diplegic cerebral palsy.
Neuropsychology, 8, 180186.